-
Notifications
You must be signed in to change notification settings - Fork 80
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Compilation fails under Ubuntu 20.04 LTS #484
Labels
Comments
simoninns
added
ld-decode-tools
An issue only affecting the ld-decode-tools
deficiency
labels
Apr 24, 2020
Removed all references to OpenCV. Requesting a 3D NTSC chroma-decode now results in the code running the 2D NTSC decoder instead. All compiles under 20.04 LTS |
atsampson
added a commit
to atsampson/ld-decode
that referenced
this issue
Jul 12, 2020
After the removal of the OpenCV code in 6c7f74d663d ("Fix for issue happycube#484"), ntsc3d always chooses the 2D result, so it behaves the same as ntsc2d. This makes it always use the 3D result instead, which produces artefacts on motion, but much better quality than ntsc2d for still images.
atsampson
added a commit
to atsampson/ld-decode
that referenced
this issue
Apr 13, 2021
The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484").
atsampson
added a commit
to atsampson/ld-decode
that referenced
this issue
Apr 18, 2021
The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484").
atsampson
added a commit
that referenced
this issue
Apr 21, 2021
The OpenCV code was removed in 6c7f74d663d ("Fix for issue #484").
VideoMem
added a commit
to VideoMem/ld-decode
that referenced
this issue
Apr 30, 2021
…sync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support (oyvindln#25) * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Split up replace_spikes to avoid numba compile issues and run formatter Avoid mean of empty slice warning - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Tidy indentation in chromadecoderconfigdialog.ui. The chromaGain default is 1.0 for both standards. Tidy indentation in yiqToYUVFrame. Remove duplicate spacers in chromadecoderconfigdialog.ui. The first one produced a warning from uic because it had a duplicate ID, but neither is actually needed. Remove remaining references to opencv. The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484"). Update copyright notices in ld-chroma-decoder. This catches up on the changes for noise reduction and YCbCr output. Use QString::arg rather than QString::sprintf. sprintf is deprecated in Qt 5.15. Remove some stray spaces from warning/error messages. Use \n rather than QTextStream::endl. The latter is deprecated in Qt 5.15, and the only difference is that it doesn't force a flush (which probably isn't what we want in the middle of printing a dropout list anyway). PalColour doesn't need to be a QObject. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com>
oyvindln
pushed a commit
to oyvindln/vhs-decode
that referenced
this issue
May 1, 2021
* Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Bad signal / 'Chewed tape' sample (preliminary fix) (#24) * Added chroma trap option to cvbs-decode * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support (#25) * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Split up replace_spikes to avoid numba compile issues and run formatter Avoid mean of empty slice warning - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Tidy indentation in chromadecoderconfigdialog.ui. The chromaGain default is 1.0 for both standards. Tidy indentation in yiqToYUVFrame. Remove duplicate spacers in chromadecoderconfigdialog.ui. The first one produced a warning from uic because it had a duplicate ID, but neither is actually needed. Remove remaining references to opencv. The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484"). Update copyright notices in ld-chroma-decoder. This catches up on the changes for noise reduction and YCbCr output. Use QString::arg rather than QString::sprintf. sprintf is deprecated in Qt 5.15. Remove some stray spaces from warning/error messages. Use \n rather than QTextStream::endl. The latter is deprecated in Qt 5.15, and the only difference is that it doesn't force a flush (which probably isn't what we want in the middle of printing a dropout list anyway). PalColour doesn't need to be a QObject. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com>
oyvindln
pushed a commit
to oyvindln/vhs-decode
that referenced
this issue
May 9, 2021
* Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Resync Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Restored deleted files Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * VBI serration detection test Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * vsync_levels pre-cleanup Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * VBI vsync autolevels testing Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Temporarly disabled extra-diff code to force zero crossings, it can be enabled with the -ed switch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added demodblock() envelope (env) zeroes check, to avoid potential division by zero during the high_part boost calculation. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Bad signal / 'Chewed tape' sample (preliminary fix) (#24) * Added chroma trap option to cvbs-decode * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support (#25) * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Split up replace_spikes to avoid numba compile issues and run formatter Avoid mean of empty slice warning - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Tidy indentation in chromadecoderconfigdialog.ui. The chromaGain default is 1.0 for both standards. Tidy indentation in yiqToYUVFrame. Remove duplicate spacers in chromadecoderconfigdialog.ui. The first one produced a warning from uic because it had a duplicate ID, but neither is actually needed. Remove remaining references to opencv. The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484"). Update copyright notices in ld-chroma-decoder. This catches up on the changes for noise reduction and YCbCr output. Use QString::arg rather than QString::sprintf. sprintf is deprecated in Qt 5.15. Remove some stray spaces from warning/error messages. Use \n rather than QTextStream::endl. The latter is deprecated in Qt 5.15, and the only difference is that it doesn't force a flush (which probably isn't what we want in the middle of printing a dropout list anyway). PalColour doesn't need to be a QObject. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * fixup! - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Modified the serration search code, now the envelope uses the fifth Vsync harmonic limit instead the third harmonic. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changed min valid levels to report valid serration levels Added field_state levels. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * chroma loss sample fails Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Regenerates the sync pulses amplitude Skips 100ms Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Seems to sort a lot of edge cases, saving this Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added resync branch getpulses_override() logic Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Safe clips the bottom of the sync pulses but leaves picture area unchanged, https://youtu.be/ZlsS_a4qdI4 Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Nosync sample decodes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Serration levels detection adjustments (produces new frames on chewed sample) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Suggested oyvindln changes over replace_spikes() on diff_demod Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added ldd.core logger, changed the name of the Vsync class to VsyncSerration Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Disabled serration detection in cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Encapsulated all the sync search and the serration search into Resync Class Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merged resync branch changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Corrected the fallback sync logic when serrations not found. - Changed the log output for the serration freqs. to kHz units. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - added a note on what I mean with fallback sync logic Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Added more debugging if show_decoded is enabled on VsyncSerration() class - Modified the sync fallback logic to improve decoding of the 'Bicycles sample' field_drop_issue_starting_at_frame_18_and_some_more_ capture26739-26911_2291x625_25Hz_8bit.flac Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com>
oyvindln
pushed a commit
to oyvindln/vhs-decode
that referenced
this issue
May 25, 2021
* Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Resync Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Restored deleted files Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * VBI serration detection test Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * vsync_levels pre-cleanup Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * VBI vsync autolevels testing Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Temporarly disabled extra-diff code to force zero crossings, it can be enabled with the -ed switch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added demodblock() envelope (env) zeroes check, to avoid potential division by zero during the high_part boost calculation. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Bad signal / 'Chewed tape' sample (preliminary fix) (#24) * Added chroma trap option to cvbs-decode * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support (#25) * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Split up replace_spikes to avoid numba compile issues and run formatter Avoid mean of empty slice warning - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Tidy indentation in chromadecoderconfigdialog.ui. The chromaGain default is 1.0 for both standards. Tidy indentation in yiqToYUVFrame. Remove duplicate spacers in chromadecoderconfigdialog.ui. The first one produced a warning from uic because it had a duplicate ID, but neither is actually needed. Remove remaining references to opencv. The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484"). Update copyright notices in ld-chroma-decoder. This catches up on the changes for noise reduction and YCbCr output. Use QString::arg rather than QString::sprintf. sprintf is deprecated in Qt 5.15. Remove some stray spaces from warning/error messages. Use \n rather than QTextStream::endl. The latter is deprecated in Qt 5.15, and the only difference is that it doesn't force a flush (which probably isn't what we want in the middle of printing a dropout list anyway). PalColour doesn't need to be a QObject. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * fixup! - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Modified the serration search code, now the envelope uses the fifth Vsync harmonic limit instead the third harmonic. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changed min valid levels to report valid serration levels Added field_state levels. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * chroma loss sample fails Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Regenerates the sync pulses amplitude Skips 100ms Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Seems to sort a lot of edge cases, saving this Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added resync branch getpulses_override() logic Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Safe clips the bottom of the sync pulses but leaves picture area unchanged, https://youtu.be/ZlsS_a4qdI4 Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Nosync sample decodes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Serration levels detection adjustments (produces new frames on chewed sample) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Suggested oyvindln changes over replace_spikes() on diff_demod Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added ldd.core logger, changed the name of the Vsync class to VsyncSerration Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Disabled serration detection in cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Encapsulated all the sync search and the serration search into Resync Class Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merged resync branch changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Corrected the fallback sync logic when serrations not found. - Changed the log output for the serration freqs. to kHz units. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - added a note on what I mean with fallback sync logic Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Added more debugging if show_decoded is enabled on VsyncSerration() class - Modified the sync fallback logic to improve decoding of the 'Bicycles sample' field_drop_issue_starting_at_frame_18_and_some_more_ capture26739-26911_2291x625_25Hz_8bit.flac Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added field hash debug info for SysParams and field.data["video"]["demod_05"] (sync_reference) Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Added some checks for SysParams consistency, and found the AGC changes both the ire0 parameter and the hz_ire parameter (all other parameters seems to be constant). Added an assertion to check if the length of the data input of demodblock() corresponds to blocklen and the same lengths of the output arrays (no issues found with that) Added the field data sync reference hash to the debug log for diagnosing the demod data randomness issue. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Refactored the moving average filter for re-utilisation on other code points. Added it to the AGC measuring values to filter out noise. It seems to fix/workaround AGC issues on the samples I have. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Modified fallback sync/levels logic Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * vsync_envelope_simple() not so tested corrections Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Added experimental changes to vsync_envelope_simple(), merged oyvindln numba and np.square optimisations Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * fallback logic getLevels fixes + oyvindln process.py changes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Some tape batch tests script intro Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Tape batch tests improvements Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Tape batch tests improvements Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Passed the optional debug flag to the resync code. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Restored vsync_envelope_simple() logic, to fix bicycle sample start problem. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com>
oyvindln
pushed a commit
to oyvindln/vhs-decode
that referenced
this issue
Jul 23, 2021
…ag (#37) * Merge of vhs_decode branch * typo glitch fix * DC restore intro * DC restore, before sync restore * DC restore improved metrics * DC restore improved metrics, pre merge check * Added chroma trap option to cvbs-decode * Auto-sync level, first try * Auto-sync level, added to cvbs + wavesink * Auto-sync level (block level, not working) , added to cvbs + wavesink * Auto-sync level (block level, not working) , added to cvbs + wavesink * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal * Previous merge commit * Merge of oyvindln/vhs_decode changes * Resync * Moved experimental sync separator code to resync branch * Restored deleted files * Comment typo fix * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * VBI serration detection test * Merge of vhs_decode branch * typo glitch fix * Added chroma trap option to cvbs-decode * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal * Merge of oyvindln/vhs_decode changes * Comment typo fix * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) * vsync_levels pre-cleanup * VBI vsync autolevels testing * Temporarly disabled extra-diff code to force zero crossings, it can be enabled with the -ed switch * Added demodblock() envelope (env) zeroes check, to avoid potential division by zero during the high_part boost calculation. * - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. * Bad signal / 'Chewed tape' sample (preliminary fix) (#24) * Added chroma trap option to cvbs-decode * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Some dead code removal Merge of oyvindln/vhs_decode changes Comment typo fix Comment typo fix getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support (#25) * Merge of vhs_decode branch * typo glitch fix * DC restore intro * DC restore, before sync restore * DC restore improved metrics * DC restore improved metrics, pre merge check * Added chroma trap option to cvbs-decode * Auto-sync level, first try * Auto-sync level, added to cvbs + wavesink * Auto-sync level (block level, not working) , added to cvbs + wavesink * Auto-sync level (block level, not working) , added to cvbs + wavesink * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal * Previous merge commit * Merge of oyvindln/vhs_decode changes * Moved experimental sync separator code to resync branch * Comment typo fix * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * Merge of vhs_decode branch * typo glitch fix * Added chroma trap option to cvbs-decode * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal * Merge of oyvindln/vhs_decode changes * Comment typo fix * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Split up replace_spikes to avoid numba compile issues and run formatter Avoid mean of empty slice warning - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Tidy indentation in chromadecoderconfigdialog.ui. The chromaGain default is 1.0 for both standards. Tidy indentation in yiqToYUVFrame. Remove duplicate spacers in chromadecoderconfigdialog.ui. The first one produced a warning from uic because it had a duplicate ID, but neither is actually needed. Remove remaining references to opencv. The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484"). Update copyright notices in ld-chroma-decoder. This catches up on the changes for noise reduction and YCbCr output. Use QString::arg rather than QString::sprintf. sprintf is deprecated in Qt 5.15. Remove some stray spaces from warning/error messages. Use \n rather than QTextStream::endl. The latter is deprecated in Qt 5.15, and the only difference is that it doesn't force a flush (which probably isn't what we want in the middle of printing a dropout list anyway). PalColour doesn't need to be a QObject. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. * fixup! - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Modified the serration search code, now the envelope uses the fifth Vsync harmonic limit instead the third harmonic. * Changed min valid levels to report valid serration levels Added field_state levels. * chroma loss sample fails * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() * Regenerates the sync pulses amplitude Skips 100ms * Seems to sort a lot of edge cases, saving this * Added resync branch getpulses_override() logic * Safe clips the bottom of the sync pulses but leaves picture area unchanged, https://youtu.be/ZlsS_a4qdI4 * Nosync sample decodes * Serration levels detection adjustments (produces new frames on chewed sample) * Suggested oyvindln changes over replace_spikes() on diff_demod * Added ldd.core logger, changed the name of the Vsync class to VsyncSerration * Disabled serration detection in cvbs-decode * Encapsulated all the sync search and the serration search into Resync Class * Merged resync branch changes * - Corrected the fallback sync logic when serrations not found. - Changed the log output for the serration freqs. to kHz units. * - added a note on what I mean with fallback sync logic * - Added more debugging if show_decoded is enabled on VsyncSerration() class - Modified the sync fallback logic to improve decoding of the 'Bicycles sample' field_drop_issue_starting_at_frame_18_and_some_more_ capture26739-26911_2291x625_25Hz_8bit.flac * Added field hash debug info for SysParams and field.data["video"]["demod_05"] (sync_reference) * Added some checks for SysParams consistency, and found the AGC changes both the ire0 parameter and the hz_ire parameter (all other parameters seems to be constant). Added an assertion to check if the length of the data input of demodblock() corresponds to blocklen and the same lengths of the output arrays (no issues found with that) Added the field data sync reference hash to the debug log for diagnosing the demod data randomness issue. * Refactored the moving average filter for re-utilisation on other code points. Added it to the AGC measuring values to filter out noise. It seems to fix/workaround AGC issues on the samples I have. * Modified fallback sync/levels logic * vsync_envelope_simple() not so tested corrections * Added experimental changes to vsync_envelope_simple(), merged oyvindln numba and np.square optimisations * fallback logic getLevels fixes + oyvindln process.py changes * Some tape batch tests script intro * Tape batch tests improvements * Tape batch tests improvements * Passed the optional debug flag to the resync code. * Restored vsync_envelope_simple() logic, to fix bicycle sample start problem. * Merged from vhs_decode upstream. Minor README changes Added SVHS support to batch_tests.py * ChromaAFC introduction * ChromaAFC enable flag -cafc requires pyhtt (sudo pip3 install pyhht) It will use unwarp_hilbert(), but not tuned yet Added pyhht to the README.md * Improved Chroma AFC * Chroma AFC, now it does something * Merge of upstream and Chroma AFC changes * Chroma AFC measure smoothing changes * Serration fixes * Added --sync_clip optional flag Added fixes to vsyncserration.py Merge of upstream. * Improved -cafc * -cafc (only phase, not freq) * -cafc (only phase, not freq), enables in try_detect_track * If -cafc filters chroma after TBC + chroma related refactor * Changes IGNORED_LINES in burst phase detection * -cafc selects the nearest peak harmonic from the chroma channel as carrier * -cafc more interesting results * -cafc some corrections * -cafc searchs peak carrier Does chroma trap after TBC Cleans up the gen_chroma scripts a bit * -cafc max carried dev adjusted to +- 8% * -cafc restores nearest center carrier selection logic * -cafc fraction of fh selector * serration/AGC moving average window adjustments * serration/AGC moving average window adjustments * serration/AGC moving average window adjustments * Chroma AFC changes, serration changes, chroma trap regression. * Sync levels adjustments * Sync logic fixes and cleanups * Chewed tape case review * Infinite loop fix * Cleanup * Readme fixes * Readme fixes * IGNORED_LINES parameter reversion Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com>
oyvindln
pushed a commit
to oyvindln/vhs-decode
that referenced
this issue
Feb 20, 2023
* Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Resync Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Restored deleted files Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * VBI serration detection test Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * vsync_levels pre-cleanup Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * VBI vsync autolevels testing Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Temporarly disabled extra-diff code to force zero crossings, it can be enabled with the -ed switch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added demodblock() envelope (env) zeroes check, to avoid potential division by zero during the high_part boost calculation. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Bad signal / 'Chewed tape' sample (preliminary fix) (#24) * Added chroma trap option to cvbs-decode * Comment typo fix * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support (#25) * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore intro Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore, before sync restore Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * DC restore improved metrics, pre merge check Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, first try Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level, added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Auto-sync level (block level, not working) , added to cvbs + wavesink Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Previous merge commit Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Moved experimental sync separator code to resync branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of vhs_decode branch Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * typo glitch fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added chroma trap option to cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge branch 'vhs_decode' of github.com:oyvindln/ld-decode into vhs_decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> # Conflicts: # vhsdecode/process.py (solved) * Some dead code removal Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merge of oyvindln/vhs_decode changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Comment typo fix Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * getpulses_override() keep the last valid blacklevel, synclevel, and vsync_locs for use in the case where they cannot be measured on the current field. Fixes some bad signal cases. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added -palm -pm command line switches and RFParams_MPAL_VHS on formats.py Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * cvbs-decode / vhs-decode common options deduplication + preliminary PAL-M support Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changes cleanup (vhs-decode/cvbs-decode command line deduplication + preliminary PAL-M support) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function as shown before. - Added a command line option to optionally disable diff demod - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Split up replace_spikes to avoid numba compile issues and run formatter Avoid mean of empty slice warning - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Tidy indentation in chromadecoderconfigdialog.ui. The chromaGain default is 1.0 for both standards. Tidy indentation in yiqToYUVFrame. Remove duplicate spacers in chromadecoderconfigdialog.ui. The first one produced a warning from uic because it had a duplicate ID, but neither is actually needed. Remove remaining references to opencv. The OpenCV code was removed in 6c7f74d663d ("Fix for issue happycube#484"). Update copyright notices in ld-chroma-decoder. This catches up on the changes for noise reduction and YCbCr output. Use QString::arg rather than QString::sprintf. sprintf is deprecated in Qt 5.15. Remove some stray spaces from warning/error messages. Use \n rather than QTextStream::endl. The latter is deprecated in Qt 5.15, and the only difference is that it doesn't force a flush (which probably isn't what we want in the middle of printing a dropout list anyway). PalColour doesn't need to be a QObject. - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * fixup! - Changed the replace_spikes function to use np.where instead np.argwhere. - Added a command line option to optionally disable diff demod -nodd - Added a command line option to optionally disable DC offset blanking level compensation -nodc - Added a warning print when the RF goes weak and prevented env to divide by zero on the high boost - Improved the field state sync/blank levels, instead of keeping the last valid measured values, it stacks them and does a moving average filter. It seems to fix the numpy error sample, and the 'chewed tape' sample. Both decodes completely. Modified the serration search code, now the envelope uses the fifth Vsync harmonic limit instead the third harmonic. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Changed min valid levels to report valid serration levels Added field_state levels. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * chroma loss sample fails Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * RuntimeWarning: Div by zero prevented at lddecode/utils.calczc_do() Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Regenerates the sync pulses amplitude Skips 100ms Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Seems to sort a lot of edge cases, saving this Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added resync branch getpulses_override() logic Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Safe clips the bottom of the sync pulses but leaves picture area unchanged, https://youtu.be/ZlsS_a4qdI4 Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Nosync sample decodes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Serration levels detection adjustments (produces new frames on chewed sample) Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Suggested oyvindln changes over replace_spikes() on diff_demod Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added ldd.core logger, changed the name of the Vsync class to VsyncSerration Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Disabled serration detection in cvbs-decode Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Encapsulated all the sync search and the serration search into Resync Class Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Merged resync branch changes Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Corrected the fallback sync logic when serrations not found. - Changed the log output for the serration freqs. to kHz units. Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - added a note on what I mean with fallback sync logic Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * - Added more debugging if show_decoded is enabled on VsyncSerration() class - Modified the sync fallback logic to improve decoding of the 'Bicycles sample' field_drop_issue_starting_at_frame_18_and_some_more_ capture26739-26911_2291x625_25Hz_8bit.flac Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> * Added field hash debug info for SysParams and field.data["video"]["demod_05"] (sync_reference) Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Added some checks for SysParams consistency, and found the AGC changes both the ire0 parameter and the hz_ire parameter (all other parameters seems to be constant). Added an assertion to check if the length of the data input of demodblock() corresponds to blocklen and the same lengths of the output arrays (no issues found with that) Added the field data sync reference hash to the debug log for diagnosing the demod data randomness issue. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Refactored the moving average filter for re-utilisation on other code points. Added it to the AGC measuring values to filter out noise. It seems to fix/workaround AGC issues on the samples I have. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Modified fallback sync/levels logic Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * vsync_envelope_simple() not so tested corrections Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Added experimental changes to vsync_envelope_simple(), merged oyvindln numba and np.square optimisations Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * fallback logic getLevels fixes + oyvindln process.py changes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Some tape batch tests script intro Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Tape batch tests improvements Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Tape batch tests improvements Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Passed the optional debug flag to the resync code. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Restored vsync_envelope_simple() logic, to fix bicycle sample start problem. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Merged from vhs_decode upstream. Minor README changes Added SVHS support to batch_tests.py Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * ChromaAFC introduction Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * ChromaAFC enable flag -cafc requires pyhtt (sudo pip3 install pyhht) It will use unwarp_hilbert(), but not tuned yet Added pyhht to the README.md Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Improved Chroma AFC Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Chroma AFC, now it does something Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Merge of upstream and Chroma AFC changes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Chroma AFC measure smoothing changes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Serration fixes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Added --sync_clip optional flag Added fixes to vsyncserration.py Merge of upstream. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Improved -cafc Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc (only phase, not freq) Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc (only phase, not freq), enables in try_detect_track Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * If -cafc filters chroma after TBC + chroma related refactor Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Changes IGNORED_LINES in burst phase detection Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc selects the nearest peak harmonic from the chroma channel as carrier Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc more interesting results Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc some corrections Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc searchs peak carrier Does chroma trap after TBC Cleans up the gen_chroma scripts a bit Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc max carried dev adjusted to +- 8% Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc restores nearest center carrier selection logic Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * -cafc fraction of fh selector Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * serration/AGC moving average window adjustments Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * serration/AGC moving average window adjustments Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * serration/AGC moving average window adjustments Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Chroma AFC changes, serration changes, chroma trap regression. Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Sync levels adjustments Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Sync logic fixes and cleanups Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Chewed tape case review Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Infinite loop fix Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Cleanup Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Readme fixes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Readme fixes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * IGNORED_LINES parameter reversion Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Added -cshift option and updated advanced_flags.md Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Reverts some auto-format code, adds pdb-attach to pip3 install in README.md. Restores ntsc2d as default decoder on gen_chroma_vid_ntsc.sh Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Merge of upstream + cafc chroma filter sample rate fix. Removes chroma shift command line parameter Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Fix typo on gen_chroma_vid_ntsc.sh Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * First attempts at HiFi decoding Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * HiFi code cleanup Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Updates to main fork changes Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Adds hifi-decode to main branch Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Adds hifi-decode to setup.py Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> * Fixes unwanted changes on cmdcommons.py Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com> --------- Signed-off-by: Sebastian Wilwerth <sebastian.wilwerth@gmail.com> Signed-off-by: VideoMem <sebastian.wilwerth@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Testing with the new Ubuntu 20.04 LTS shows issues compiling against the OpenCV2 headers:
../ld-chroma-decoder/opticalflow.h:34:10: fatal error: opencv2/core/core.hpp: No such file or directory
34 | #include <opencv2/core/core.hpp>
| ^~~~~~~~~~~~~~~~~~~~~~~
Since the NTSC optical flow isn't functioning, the requirement to include opencv should be removed (and once reimplemented a more up-to-date header should be included instead).
The text was updated successfully, but these errors were encountered: