Psychtoolbox 3.0.19.13 release "Stupido!"

This is a bug fix release for 3.0.19.12, which reverts some broken change to SetupPsychtoolbox() for macOS.

Psychtoolbox 3.0.19.12 release "Scraps"

Psychtoolbox 3.0.19 update "Scraps" was released at 5th June 2024.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.12”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.12

Some of this Psychtoolbox release was sponsored by Mathworks under the year
2023/2024 contract..

Compatibility changes wrt. Psychtoolbox 3.0.19.11:

• None.

Highlights:

• None. Just some small fixups for improved Ubuntu 24.04-LTS and Matlab R2024a
  compatibility.

All:

Linux:

• Psychtoolbox was built and tested against Matlab R2024a.

• PsychLinuxConfiguration: Quote paths, to improve MLTBX compatibility.

• PsychLinuxConfiguration(): Add users also to 'gamemode' group.
  This provides additional permissions to the gamemoded running on behalf
  of the user, to execute additional performance optimizations as part
  of Priority(n) with n > 0. Needed on at least Ubuntu 24.04-LTS and later.

• Screen('OpenMovie'): On Linux, use pipewiresink as audio sink if
  available. Should fix the following issue on Ubuntu 24.04-LTS and later,
  and other distributions that use Pipewire as desktop sound server and
  have an outdated buggy Pulseaudio 16 installation, instead of the more
  recent Pulseaudio 17. The bug can't be fixed on Ubuntu 22.04-LTS without
  manual user workarounds. See following issue for this bug, which is
  likely not encountered often:

  #814

  On unfixed affected systems (mostly Ubuntu 22.04-LTS), if there is
  a notable time delay between Screen('OpenMovie') and start of
  video+audio playback via Screen('PlayMovie'), then the movie
  playback will freeze after displaying the 1st video frame for the
  duration of that delay between OpenMovie and PlayMovie, then
  continue. Problem didn't exist in earler Linux distributions like
  Ubuntu 20.04-LTS and is fixed by this bug fix for later distros like
  Ubuntu 24.04-LTS from April 2024.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

macOS:

• Psychtoolbox was built against Matlab R2023b and tested against
  Matlab R2024a and Octave 8.4 from HomeBrew. It also works on the
  new Octave 9.1 from HomeBrew

• SetupPsychtoolbox: Quote paths, to improve MLTBX compatibility.

• TextInOffscreenWindowTest.m: Fix it for Retina displays on macOS. Also cleanup.

• BitsPlusPlus: Handle Apple macOS Retina displays properly for imaging pipeline validation runs.

• ImagingStereoDemo: Handle Apple Retina displays properly during movie writing.

• TwoStateQuery(): Add workaround for broken Matlab R2024a on Apple macOS.

Enjoy!

Psychtoolbox 3.0.19.11 release "Eyegazer"

Psychtoolbox 3.0.19 update "Eyegazer" was released at 18th May 2024.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.11”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.11

Some of this Psychtoolbox release was sponsored by Mathworks under the year
2023/2024 contract..

Our partner SR-Research, makers of the Eyelink gaze trackers, sponsored
the major 2024 Eyelink toolbox update.

Compatibility changes wrt. Psychtoolbox 3.0.19.10:

• None expected. But possibility due to massive Eyelink toolbox changes.

Highlights:

• Big Eyelink toolbox update by SR-Research with the most massive
  improvements in over a decade.

All:

• Release of the new Eyelink toolbox. This is the biggest update of the
  Eyelink toolbox for Eyelink gaze trackers from SR-Research since over
  a decade! Most of the work was done by the team at SR-Research. Credits
  to Brian Richardson, Marcus Johnson, and Kurt Debono. Additional 88
  hours of my work time were sponsored by SR-Research, to provide code
  review, feedback and advice on improving the integration while
  maintaining backwards compatibility, and code improvements. Among
  many other improvements, the new Eyelink toolbox can integrate better
  with PsychPortAudio for playing feedback sounds during tracker calibration,
  and it can use dynamic animated calibration targets by playing back
  movies. Upload of backdrop images to the tracker host computer is possible.
  Stereoscopic/Binocular display modes of Psychtoolbox are now supported
  for use with the gazetracker. Many new functions have been added to improve
  ease and efficiency of use. A new set of demos demonstrates efficient use
  and best practices with the new toolbox. Legacy functions should be still
  supported for backwards compatibility, but many legacy demos have been
  removed. The toolbox has been carefully tested by SR-Research on all
  supported operating systems, so we hope this major update will go without
  major problems. Thanks to our partner SR-Research for this contribution
  and the team at SR-Research for good collaboration!

• CedrusResponseBox(): Disable any flow control.
  It was supposed to use hardware flow control (FlowControl=Hardware),
  but that did not work until PTB v3.0.19.6 due to an IOPort parameter
  handling bug. Since the bug is fixed, hardware flow control would be
  used. This was supposed to make Cedrus to work more robust, but it
  apparently had the opposite effect, as the following PTB forum message
  suggests:

  https://psychtoolbox.discourse.group/t/waiting-forever-for-data-transmission-with-cedrus-rb/5319

  Disable flow control explicitely, to revert to pre 3.0.19.6 behaviour,
  where Cedrus boxes worked as well as they could.

  If this fixes the bug though, that means my original opinion about
  the quality of Cedrus hardware - or the lack thereof - would be
  true, and the fragile communication was not caused by our IOPort
  driver bug, ie. by lack of hardware flow control. Let's see...

• ImagingStereoMoviePlayer.m: Also handle top-bottom stereo encoded movies.

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

• Fix most basic compatibility with Ubuntu 24.04.0-LTS. Extensive compatibility
  testing with Ubuntu 24.04-LTS is still outstanding.

• Fix Matlab R2020 - R2024a plotting on Intel graphics. After this Matlab bug
  being reported to Mathworks and going unfixed for multiple years, I decided
  to cave and add a workaround to PsychLinuxConfiguration(). It will add some
  config option during post-install to Matlab's Java JOGL OpenGL interface, to
  avoid use of OpenGL ARB contexts. This works around the Matlab bug and makes
  Matlab's graphics and plotting functions work on Intel graphics chips.

• Refine PsychCV's Apriltags implementation. Make sure that if the same marker
  is apparently detected multiple times in an image, that then the detection
  with the lowest hamming error and highest score is returned, so good detections
  can't override false positive noise. Allow limiting the number of markers used
  from a tagFamily, so one can select tag families whose total marker count would
  cause a very high memory consumption, leading to out-of-memory errors on machines
  with even 8 GB of RAM, or very long startup times, by only using a subset of
  markers. Allow to select more tuning parameters for marker detection. Suppress
  useless debug output.

• Screen('OpenVideoCapture'): Allow to disable videorate converter during recording.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

• PsychPortAudio: Use auto samplerate conversion under WASAPI shared mode
  with reqlatencylcass 0 or 1 (the default). This should allow to request
  arbitrary audio sample rates, as the operating system should perform
  proper sample rate conversion. Should automatically fix issues like the
  following:

  https://psychtoolbox.discourse.group/t/ptb-thinks-usb-sound-card-doesnt-support-44100hz/5324

  For now this is only implemented for Octave, not yet Matlab, until it can be
  tested a bit more.

macOS:

• Psychtoolbox was built and lightly tested against Matlab R2023b and
  Octave 8.4 from HomeBrew. It also works on the new Octave 9.1 from
  HomeBrew

Enjoy!

Psychtoolbox 3.0.19.10 release "Bitty Bits"

Psychtoolbox 3.0.19 Beta update "Bitty Bits" was released at 26th March 2024.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.10”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.10

Compatibility changes wrt. Psychtoolbox 3.0.19.9:

• None.

Highlights:

• None, just bug fixes and minor improvements, and Octave 9.1 on macOS support.

All:

• PropixxImageUndistortionThrowaway.m: Fix stupid bug. Don't close the onscreen
  window and then try use it again afterwards!

• Fix wrong definition of glPathGlyphIndexRangeNV().

• Sound input demos: Allow specification of number of input channels.
  For BasicSoundInputDemo and DelayedSoundFeedbackDemo, allow to specify the
  number of inputs, to also deal with sound cards which only provide one mono
  input channel. Continue to default to 2 stereo channels.

• Various internal cleanups with no immediate user facing benefit.

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

• Screen: Fix idiotic bug in Screen('GetFlipInfo') there since day 1!
  If a query for a flip info struct was made, but no new struct was available,
  the function errored out, instead of returning an empty [] return argument!

  This was not ever noticed due to the way our only test scripts for this
  PerceptualVBLSyncTestFlipInfo2 had try-catch-end statement placed just so
  that the error was swallowed and never showed up!

  Fix this properly by returning an empty [] return argument in this case,
  as expected.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

macOS:

• Psychtoolbox was built and lightly tested against Matlab R2023b and
  Octave 8.4 from HomeBrew. It also works on the new Octave 9.1 from
  HomeBrew

• Enable use with the brand-new Octave 9.1.0 from HomeBrew.

Psychtoolbox 3.0.19.9 release "Tag games"

Psychtoolbox 3.0.19 Beta update "Tag games" was released at 16th March 2024.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.9”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.9

Compatibility changes wrt. Psychtoolbox 3.0.19.8:

• None

Highlights:

• Experimental Apriltags fiducial marker tracking support on Linux.

All:

• New helper function PsychCamIntrinsicsToGLProjectionMatrix().
  Given camera intrinsic parameters from some camera calibration software,
  and the desired zNear and zFar clipping planes of the view frustum,
  compute and return a projection matrix for use as GL_PROJECTION_MATRIX
  for OpenGL rendering with a virtual pinhole camera that approximates the
  properties of the real camera.

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

• (Re-)add PsychCV() mex file for Octave and Matlab, with support for
  vision based tracking of AprilTag fiducial markers (aka tags) by use
  of the apriltag library. This PsychCV mex file will only work on
  Ubuntu 22.04-LTS and later, not on Ubuntu 20.04-LTS anymore.

Psychtoolbox 3.0.19.8 release "CamJam"

Psychtoolbox 3.0.19 Beta update "CamJam" was released at 14th March 2024.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.8”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.8

Some of this Psychtoolbox release was sponsored by Mathworks under the year
2023/2024 contract., and our partner SR-Research, makers of the Eyelink gaze trackers.

Compatibility changes wrt. Psychtoolbox 3.0.19.7:

• dontOpenExisting non-zero flag in Eyelink('OpenFile') now error aborts,
  instead of trying to implement something unsupportable.

Highlights:

• None. Mostly a small refinement release.

All:

• Eyelink('OpenFile'): Error out if optional dontOpenExisting flag is set
  to a non-zero value. Update help text. According to feedback from SR-Research
  this is simply not supportable with all current and past Eyelink runtimes and
  tracker host software. Game over for this feature, just warn users about it.

• Gamepad() / Joystick() et al.: Improve docs, help, setup instructions. Especially
  for proper choice of joystick driver on Ubuntu 20.04/22.04-LTS.

• Downloaders/Updaters etc.: Update again for final SVN death on GitHub. In combo
  with website updates, to make clear that functions DownloadPsychtoolbox and
  UpdatePsychtoolbox and DownloadAdditionsForLinux are currently dead.

• SetupPsychtoolbox(): Dead code removal, cleanup, docs updates.

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

• XOrgConfSelector: Add error handling and error resolution tips.
  There wasn't error handling in case of delete() or copyfile() failure.
  Add something, and try to give some helpful feedback and advice, e.g.,
  running PsychLinuxConfiguration().

• Allow selection of a different clock to be used as the general GetSecs
  clock for (almost) all timing and timestamping, e.g., by GetSecs, WaitSecs,
  PsychPortAudio, Screen, etc. Since the year 2007, the so called wall clock,
  or gettimeofday() clock, or Posix CLOCK_REALTIME clock was used for this,
  and it continues to be the default. But now one can use the new PsychTweak
  function PsychTweak('GetSecsClock', clockid); at the beginning of a
  script (before any other calls to Psychtoolbox mex files!) to select a clock
  with a different clockid. A clockid setting of 0 uses the default CLOCK_REALTIME
  clock, as in the past. A clockid setting of 1 uses CLOCK_MONOTONIC instead.
  CLOCK_MONOTONIC, as opposed to default CLOCK_REALTIME, does not maintain time
  in seconds since 1st January 1970 0:00:00 UTC, but has an unspecified zero
  point, but usually 0 denotes time of operating system boot. CLOCK_MONOTONIC
  is always guaranteed to be monotonically increasing, not subject to step
  adjustments by the system administrator, NTP, PTP, or other external clock
  reference sources. This prevents synchronization to other computer on the
  network/setup, but may avoid clock trouble caused by time jumping in certain
  situations. CLOCK_MONOTONIC behaves like the GetSecs clocks for Psychtoolbox
  on MS-Windows and macOS. For now the new monotonic mode is only lightly
  tested, and only available on Psychtoolbox for Intel processors, not yet for
  RaspberryPi - this will follow in a future release.

  This improvement was sponsored by a Mathworks sponsored support incident under
  our 2023/2024 contract.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

macOS:

• Psychtoolbox was built and lightly tested against Matlab R2023b and
  Octave 8.4 from HomeBrew.

• Detect futile attempt of running under native Matlab for Apple Silicon / ARM,
  ie. not under Rosetta2 emulation and bail with more enlightening messages.

Psychtoolbox 3.0.19.7 release "Christmas looks"

Psychtoolbox 3.0.19 Beta update "Christmas looks" was released at 16th December 2023.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.7”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.7

This Psychtoolbox release was sponsored to a large degree by Mathworks under the year
2023/2024 contract., and
to some degree by our partner SR-Research, makers of the Eyelink gaze trackers.

Compatibility changes wrt. Psychtoolbox 3.0.19.6:

None.

Highlights:

• Further refinement to the performance improvements for hardware accelerated
  movie playback via GStreamer and gpu hardware video decoders on Linux and
  MS-Windows, with only very minor benefit to macOS.

• Eyelink mex file improvements.

• Usability improvements for macOS.

All:

• Screen: Allow dynamic, per-movie, enable and disable of hardware accelerated video
  decoding when using GStreamer 1.18 or later, iow. on MS-Windows, macOS, RaspberryPi OS,
  and Ubuntu 22.04-LTS and later. Only on older GStreamer versions, e.g., Ubuntu 20.04,
  will a hardware decoding disable last for the remainder of a session, ie. until
  Octave/Matlab is quit, once the disable flag for hardware decoding has been given
  once in a session. Also some minor robustness improvements to movie playback if
  special flags are given, which could cause some trouble sometimes.

• Eyelink: New and improved Eyelink mex files with some contributions by SR-Research,
  specifically from Brian Richardson, Marcus Johnson, and Kurt Debono. Thanks!

  • A new subfunction 'EDKVersion' provides version information about the Eyelink
    software core library runtime in use.

  • A new subfunction 'EyelinkSendFile' to send a file to the Eyelink tracker
    computer. Using this function, an image or video can be uploaded from the
    display PC to the host Tracker PC.

  • Eyelink('OpenFile'): Another attempt to fix the 'dontOpenExisting' flag.

    'dontOpenExisting' functionality was broken from day 1, but nobody bothered
    to ever give us feedback about this, not even the person who requested that
    functionality! As a new attempt at a fix by myself, lets try the following
    untested fix:

    Assume that create_path() is a way to find out if a file already exists
    on the tracker computer, given that file_exists() actually operates on
    the stimulation computer, contrary to what anybody expected. The Eyelink
    programmers manual is very unclear about these file utility functions.

    This may be wrong as well, no feedback from SR-Research yet. Let's see...

  • Eyelink('ImageTransfer'): Restore my version of it, but also add some bug
    fixes by SR-Research, so that it hopefully now finally works as intended.

  • Fix regression for potential buffer overflow in 'EyelinkGetQueuedData', and
    generally make 'EyelinkGetQueuedData' and 'EyelinkGetFloatDataRaw' more robust
    and memory-safe.

  • Minor other improvements and cleanups to Eyelink mex.

• Simplify OS type and 64-Bit detection functions, ie. IsWin(), IsLinux(), IsOSX(),
  IsARM(), Is64Bit(), IsWindows(), IsWinVista(), IsWinMatlabR11Style().

• IsARM(): Try to detect running on 64-Bit ARM or on native Apple Silicon Matlab/Octave.

• DownloadPsychtoolbox and UpdatePsychtoolbox: Prepare for EOL at 8-Jan-2024.

  GitHub will shut down and permanently remove its Subversion frontend that day,
  so our current SVN based downloader and updater methods will fail completely.

  There may be, and already have been, temporary brownouts in the service, as a
  heads up by GitHub about what's coming.

  Inform users, and about their inconvenient alternatives, so they are less
  stumped by what is happening, assuming an ability to read. Due to the severe
  lack of financial support by the vast majority of our users, we didn't have
  and don't have the time right now to deal with this and find new solutions or
  workarounds for this problem. Finally people get what they pay for.

• Merge in some new GitHub actions, contributed by Mathworks open-source team,
  which allow auto-building of zip files of the Psychtoolbox folder and of a
  corresponding MTLBX file for each tagged release. This is preliminary and needs
  more work and refinement, but may help a bit to work around imminent failure
  of DownloadPsychtoolbox and UpdatePsychtoolbox for MS-Windows and Apple macOS
  users. Thanks to Bensingh Pancras from Mathworks for this work!

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2023b.

macOS:

• Psychtoolbox was built and lightly tested against Matlab R2023b and Octave 8.4
  from HomeBrew.

• IsARM(): Try to detect running on 64-Bit ARM or on native Apple Silicon Matlab/Octave.

• PsychtoolboxPostInstallRoutine(): Detect and reject native Apple Silicon Matlab/Octave.
  Point to Intel variants under Rosetta 2 emulation instead. Untested so far.

Psychtoolbox 3.0.19.6 release "Raspberry Cake"

Psychtoolbox 3.0.19 Beta update "Raspberry cake" was released at 12th December 2023.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.6”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.6

This Psychtoolbox release was sponsored to a large degree by Mathworks under the year 2023/2024 contract.,
and to some degree by Cambridge Research Systems.

Compatibility changes wrt. Psychtoolbox 3.0.19.5:

• Psychtoolbox is now tested and supported on Matlab R2023b, instead of R2022b.
  It is expected to continue to work on older releases than R2023b, but no longer
  developed or tested against them.

• Slight behavioral changes in IOPort BreakBehaviour settings on MS-Windows, and
  generally a few IOPort config options. Not expected to affect (m)any user scripts.

Highlights:

• This release was tested for compatibility with Matlab R2023b on Linux and Windows,
  specifically Ubuntu 22.04.3-LTS with AMD RavenRidge (AMD Ryzen 5 2400G processor
  integrated graphics) and AMD Polaris graphics, and on Microsoft Windows 10 22H2
  with AMD RavenRidge and NVidia GeForce GTX 1650 graphics, and to a lesser degree
  on Ubuntu 20.04.6-LTS and Windows 11 22H2 with Intel Kabylake (GT2) "UHD 620"
  integrated graphics. Testing with some onboard sound chips was also performed.
  The testing led to various smaller improvements, and various workarounds on
  MS-Windows for AMD OpenGL and Vulkan graphics driver bugs, and for some NVidia
  driver bugs, as well as some bugs in Matlab wrt. backwards compatibility and
  OS compatibility. Most improvements in this release are due to that.

  This was done as part of out Mathworks 2023/2024 contract for ensuring compatibility
  with recent Matlab releases, spending over 65 work hours. Due to not even remotely
  sufficient funding for this work, no significant testing beyond some very quick touch
  and go spot testing was performed for macOS.

• Substantial movie playback performance improvements in pixelFormat 11 for HDR
  and WCG movies when hardware video decoding is used on Linux and MS-Windows. A
  small fraction of this work was sponsored as part of the Mathworks contract.

• 10 bit deep color framebuffer and HDMI output for the RaspberryPi 4/400 by
  improvements to Psychtoolbox and to the Mesa 23.3.1 OpenGL graphics drivers
  for RaspberryPi's VideoCore 6+ gpu. Mostly paid for by Cambridge Research Systems.

All:

• Screen: Optimize storage format for LUMINANCE16 textures for texture normalization.
  If a 16 bpc luminance texture is turned into an offscreen window, or drawn into,
  the needed normalization / conversion to RGBA format now converts into RGBA16
  format instead of RGBA32 float format. This retains full precision of the original
  image content, but at half the memory consumption.

• Screen: In DrawTexture(s), always filter planar textures via nearest neighbour
  sampling instead of any of the other filtering/sampling modes, as any sampling
  other than nearest neighbour will cause massive artifacts.

• Screen: Improve performance of HDR movie playback by more efficient use of hardware
  video decoding. Enable the GStreamer movie playback engine in HDR mode (aka pixelFormat 11)
  to accept and handle typical semi-planar YUV formats used for HDR movies when
  gpu hardware accelerated video decoding is in use. This allows GStreamer to avoid
  cpu expensive frame format conversions from semi-planar frames provided by the
  gpu hardware decoders to planar formats previously accepted by PTB's texture
  creation code, leading to way higher performance when such hardware decoders
  are used. Software decoding still works via the previous path, whenever that is
  wanted or needed. Software decoding can also be enforced by disabling of hardware
  decoding by passing in specialFlags1 flag +4 in Screen('OpenMovie', ...). One
  limitation of the disable is that hardware decoding will stay off for the remainder
  of the whole session, ie. until Octave/Matlab is quit, once the disable flag
  for hardware decoding has been given once in a session. This could be fixed by
  requiring GStreamer 1.20+ as minimum requirement, ie. Ubuntu 22.04 and later.

  Note that I've repurposed/redefined specialFlags1 flag +4 as hw accel disable,
  from previously "draw 2D flow vectors on content for debugging/visualisation".
  This is ok, as the feature was probably never used by anybody, and also because
  this feature is deprecated and unsupported in all supported GStreamer versions
  since quite a while.

  The following new formats are supported for efficient handling:

  • Y42B = YUV planar 4:2:2 with 8 bpc.

  • NV16 = YUV semi-planar 4:2:2 with 8 bpc, aka P208.

  • P0xx formats of YUV semi-planar 4:2:0 sampling with 8, 10, 12 and 16 bpc,
    ie. NV12, P010_10LE, P012, P016. These are the relevant ones for typical
    HDR-10 hardware decoding from H265/HEVC, AV1, and from VP9.

  So far this has been tested on MS-Windows 10 with NVidia GeForce GTX1650 and
  AMD Raven Ridge / Vega as part of a Ryzen 5 2400G processor. On NVidia it shows
  a 2x speedup in decoding speed, to allow playback of 4k HDR at 60 fps with stable
  framerate and none to minimal frame dropping. This typically uses NVCODEC on
  NVidia hardware or Direct3D11/DXVA decoding on NVidia, AMD and Intel gpus.

  Tests on Ubuntu 20.04.6-LTS + Intel Kabylake + GStreamer 1.16, and on two
  Ubuntu 22.04.3-LTS machines with GStreamer 1.20.3, once with AMD Polaris11
  and once with AMD Raven Ridge, also showed 2x performance improvements.
  This uses VAAPI for hardware accelerated video decoding on AMD and Intel,
  and maybe on NVidia, but NVidia is untested so far on Linux.

  Note that while AMD and NVidia worked well in my testing, at pixelFormat 11,
  Intel Kabylake produced some visual decoding artifacts under Ubuntu 20.04, and
  massive visual artifacts in decoded video on Windows 11. May or may not be
  general Intel bugs, or may just be a problem on my specific old Intel iGPU.

  Additional successful testing was performed by a user under Windows-10 with
  a NVidia GeForce RTX 3080 with H265 and AV1 4k HDR 60 fps content.

  As a general guideline, as of November 2023, these video codec formats
  are generally hardware accelerated on Linux and Windows with typical
  recent hardware from AMD, NVidia and Intel:

  • MPEG2, VC1, H264, VP8: YUV 4:2:0, 8 bpc.

  • H265, VP9: YUV 4:2:0, 10 bpc HDR.

  • Sometimes also MPEG4, and AV1, the latter only on latest gpu's from
    NVidia, AMD and Intel, specifically NVidia Ampere+ RTX3000+, or AMD RDNA2
    Navi 2nd gen, or Intel Xe+/Arc+ or macOS 14 on Apple M3+.

  For best cross-platform, cross-vendor compatibility and with older hardware,
  the safest choices for high performance playback are probably H264 and H265
  at the moment.

  None of this applies to Apple macOS, as macOS too primitive OpenGL implementation
  currently does not allow use of pixelFormat 11 playback.

• PlayMoviesDemo.m: Optimize further for fast playing movies.
  Suppress costly text drawing for any fps >= 30. Skip wait for flip complete for
  non-HDR playback. Ths way the demo also serves as a real-world performance test.

• help GetChar: Update wrt. OS+GUI+Matlab/Octave support, internationalization.

• PerceptualVBLSyncTest[FlipInfo2]: Add optimizations, especially for RaspberryPi.
  Avoid redundant stimulus draw in non-stereo mono-mode. Do clear the backbuffer
  after each flip, as that seems to give a substantial performance boost for the
  VideoCore gpu's - See speculation in code comments - probably related to it being
  a tiled renderer, or a split gpu + display engine design at least for VideoCore 6+
  of RaspberryPi 4 and later. I haven't checked the driver source for the more likely
  reason yet.

  Another thing that can really help sync tests on the Pi is Priority(1), because
  then gamemode daemon chooses a more aggressive cpu governor. This only matters at
  high display refresh rates though, e.g., on a RPi 400 with a 1920x1080 120 Hz
  display. At somewhat lower refresh rates, e.g., 60 Hz, it isn't needed.

• IOPort: Fix 'OpenSerialPort' bugs where default settings override user settings.

  As GitHub user @qx1147 found out during a code review, there is a flaw in
  the handling of serial port options, where default serial port options may
  override user provided configuration options, so the wrong settings are
  silently applied! This only happens if the user script provides different
  settings in IOPort('OpenSerialPort', ..., settings), ie. on initial port
  open. Calls to IOPort('ConfigureSerialPort', ..., settings) are not
  affected.

  Luckily (dumb luck!), only a few options were mishandled, and most of these
  are rarely changed by user scripts, specifically:

  ProcessingMode=Cooked would get ignored, and raw mode used instead. Rarely
  used on Linux/macOS and not supported on Windows anyways. Low impact.

  BreakBehaviour=Ignore would be used instead of Flush or Zero. Not a
  problem on Windows, as that is the only supported option. The fact that
  this went unnoticed on Linux/macOS suggests not much use of Flush or Zero
  by users scripts. No use by PTB itself. Low expected impact. Note that the
  implementation of 'Flush' on the Unixes is of questionable use, as Flush
  would not only flush read/write buffers, but also send a SIGINT, which may
  end in unexpected ways on Unix, as we don't handle SIGINT specifically.

  StopBits=2 would get ignored and the default of 1 stop bits would be used.
  Unclear how many devices want 2 stop bits, but I haven't ever seen one, so
  probably low expected imp...

Psychtoolbox 3.0.19.5 release "Eye to eye"

The Psychtoolbox v3.0.19.5 update "Eye to eye" was released on 17th November 2023.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.5”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.5

This Psychtoolbox release was fully sponsored by Mathworks under the year
2023/2024 contract.

This is the last release built and tested with Matlab R2022b. Future development
and user support will immediately switch to Matlab R2023b after this announcement,
as we no longer have access to earlier Matlab releases, but Psychtoolbox is expected
to continue to work just fine on older Matlab versions.

Highlights:

• Eye gaze tracking support for OpenXR supported devices via PsychVRHMD(), ie.
  Virtual Reality (VR) / Augmented Reality (AR) / Mixed Reality (MR) Headsets
  with built in eye trackers and suitable eye tracking runtimes!

  This release should support monocular "cyclops eye" eye gaze tracking on HMDs
  with monocular eye trackers and binocular eye trackers (typically reporting a
  fused gaze vector for noise reduction, based on sensor fusion between both
  eyes tracking data) on all hardware and OpenXR runtimes which support the
  XR_EXT_eye_gaze_interaction OpenXR extension. Data reporting is very limited
  due to the limited functionality provided by that extension: Only a binary
  confidence value (eye tracked / not tracked) and a 3D eye gaze vector aka
  gaze ray aka eye position and looking direction, usable for determining the
  looking direction in a 3D VR scene and hit testing with 3D scene geometry, as
  well as remapped 2D pixel coordinates of where the user is looking inside the
  Psychtoolbox onscreen window, both for monoscopic mode and stereoscopic mode.
  Demos which demonstrate 3D gaze are VRInputStuffTest.m if the optional
  withGazeTracking flag is set to 2, or just 2D position for flag setting 1.
  2D gaze reporting is demonstrated - if the proper flags are set - by the
  following demos and tests: VRHMDemo.m - simple mono/stereo view reporting,
  GazeContingentDemo.m - monoscopic display, "foveated area" is moved via gaze,
  and VREyetrackingTest.m for extensive demo and testing.

  The API / how to use and description of features and limitations is in the
  help PsychVRHMD where one would expect VR/AR/MR/XR related functionality.

  Setup instructions can be found under help OpenXR.

  For VR HMDs with builtin eyetrackers from HTC, e.g., HTC Vive Pro Eye, on
  MS-Windows only, a second more feature rich eyetracking mechanism via use
  of HTC's proprietary Windows-only SRanipal eyetracking runtime is used. It
  additionally to the above "cyclops eye" reporting allows for binocular
  eyetracking, more precise gaze sample timestamps, and estimates reported for
  "eye openess" aka how much are the eyelids open, and pupil diameter.

  Due to lack of other equipment, time and funding, this functionality was only
  tested under Microsoft Windows 10 with a rented HTC Vive Pro Eye HMD under
  SteamVR 1.26, and to a lesser extent, the new SteamVR 2.0.10 runtime from
  Valve. The monocular "cyclops" eye gaze tracking should work on any other HMD
  with OpenXR compliant eye tracking though, also on Linux. And binocular tracking
  should work on other HTC HMDs under Windows as well. Theory and practice, we
  will see...

  Performance and precision is reasonably good, although functionality wise there
  is a lot of room for improvement. Further funding would be needed to do that
  work. This OpenXR eye gaze tracking was sponsored by Mathworks as the main
  new sponsored improvement under the 2023/2024 contract.
  As so often though, the funding was insufficient to fully cover our costs
  up to this point, so future development of this eyetracking support will be
  mostly dependent on future funding or contract work from interested labs
  or users.

  Given that the HTC Vive Pro Eye used for developing and testing this support
  is only rented, and due to the lack of funding, we will likely have to give
  the HMD back before christmas, so no further debugging / refinement or
  related user support can be provided after that point in time, unless some
  labs or users contract and fund us for further development and support.

  Therefore I urge you to test this functionality early and give timely feedback,
  before it is too late for a cheap solution. That still means of course that
  while constructive feedback is appreciated, no free help will be provided
  for setup and first steps in general. That's what our paid support is for.

All:

• OpenXR VR/AR/MR/XR eyetracking support. See above.

• PlayMoviesDemo.m: Some performance optimizations for high fps HDR movie playback.

• CreateResolutionPyramid(): Fix and deuglify MipMapDownsamplingShader.

• Minor bug fixes, documentation updates and improvements.

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2022b and Octave 6.4.

• OpenXR VR/AR/MR/XR eyetracking support. See above.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2022b and Octave 7.3.

• Additionally to OpenXR eyetracking, also SRAnipal binocular eyetracking on
  suitable HTC HMDs, as tested with the HTC Vive Pro Eye.

macOS:

• Psychtoolbox was built and lightly tested against Matlab R2022b and
  Octave 8.3 from HomeBrew. Octave 8.4 was also lightly tested and seems fine.

Enjoy!

Psychtoolbox 3.0.19.4 release "Virtuality!" SP4

Psychtoolbox 3.0.19 Beta update "Virtuality" SP4 was released at 27th October 2023.
As usual, the complete development history can be found in our GitHub repository.
The release tag is “3.0.19.4”, with the full tree and commit logs under the URL:

https://github.com/Psychtoolbox-3/Psychtoolbox-3/tree/3.0.19.4

This Psychtoolbox release was partially sponsored by Mathworks under the year
2023/2024 contract.

Compatibility changes wrt. Psychtoolbox 3.0.19.3:

• On MS-Windows, GStreamer 1.22 is now required, the latest lightly tested
  version is GStreamer 1.22.5. With the older GStreamer 1.20, text rendering
  with the high quality drawtext plugin will no longer work, and the fallback
  GDI text renderer would be used, which is officially unsupported in case of
  trouble and has lower performance/quality/features/reliability.

• On macOS if using Octave, Octave 8.3 is now recommended and tested, but older
  Octave versions back to v6.3 are expected (but not tested) to continue to work.

Highlights:

None, just a large grab bag of various minor and major fixes and improvements.

All:

• Some Unix file permission cleanup contributed by Yaroslav from NeuroDebian.

• CalibrateMonSpd(): Fix some fallout from previous fixes. Set cal.describe.dacsize
  also if g_usebitspp is already set. Reference:
  kleinerm#252

• DegreesToRetinalEccentricityMM(): Fix typo in code that prevented replacement
  of small angles by the linear approximation. Contributed by Stella Prins.

• MakeSineImage(): Allow passing of center (0 phase) position of the sinusoid.
  Contributed by David Brainard.

• ComputePhotopigmentBleaching(): Add constants from Wyszecki and Stiles.
  Contributed by David Brainard.

• New touchscreen demo: MultiTouchPinchDemo.m, to show detection and handling of
  two finger pinch gestures on touchscreens.

• Output info message with potential troubleshooting tips if drawtext init takes
  unusually long, hinting at potential fontconfig cache rebuild (problems) on
  MS-Windows. May or may not help anybody, but probably doesn't hurt. Suggested
  by GitHub user @mirh.

• DrawFormattedText(): Add new keyword 'left' to use for the 'sx' parameter. It
  will left-align drawn text to the left border of an optionally provided
  'winRect', similar to the 'right' keyword for right-alignment. Improvement
  contributed by GitHub user @SVNKoch.

• PsychPortAudio: Add potential workaround to deal with temperamental / weird
  audio sound cards. Add a new optional parameter to the tweaking command
  PsychPortAudio('EngineTunables', ...., workarounds);, which allows to
  specify a non-default (ie. non-zero) workarounds bitmask to selectively
  disable or enable workarounds. The currently defined workaround bits are the
  following, which modify how the Portaudio audio format test function
  Pa_IsFormatSupported is handled:

  +1 = Do not error abort on test failure, ie. print warnings but don't abort.
  +2 = Skip the whole test and always assume success.

  Both could help if the test reports false positives (+1 to continue), or
  if some hardware queries/operations themselves during the test trigger
  some trouble (+2 to skip the whole test).

  Also various other minor improvements to PsychPortAudio and some of the audio
  demos and tests.

• Snd(): Switch fallback method from use of sound() to use of audioplayer().
  Both modern Octave and Matlab implement sound() as a wrapper around their
  audioplayer() objects, so using audioplayer() directly gives us more control
  for a better Snd() fallback implementation. Also use the fallback method as
  new default by default. Iow. unless specified otherwise, Snd() will play
  via audioplayer(). This provides good interop with other audio clients and
  with Screen()'s GStreamer based movie playback engine.

• Beeper(): Formatting/Indentation fixes, refine soundvector calculation.

• BitsPlusIdentityClutTest: Disable encoder test if Vulkan display is used.
  Current design lead to the tests running before the Vulkan backend is fully
  in charge, so we display the PTB welcome screen or pixeltrash during the test,
  instead of the test stims, which leads to false-positive test failure. Just
  don't offer the test option under Vulkan. DatapixxGPUDitherpatternTest is an
  alternative working way to test with Vulkan at the moment.

• PsychOpenHMDVR: Use correct ipd/2 instead of ipd for warp-mesh setup in our
  OpenHMD driver.

• Minor bug fixes, documentation updates and improvements.

Linux:

• Psychtoolbox was built and lightly tested against Matlab R2022b.

• RPiGPIOMex: Various bug fixes and improvements. Also merged two alternative
  implementations of the file co-written by Steve Van Hooser and myself. These
  reimplement functionality by using the pigpio library instead of the old and
  deprecated wiringPi library. This should be more future-proof and maintainable.
  For now, the original file based on wiringPi is still used though, until we
  decide which of the two new variants is the better choice.

• Fix exception handling on Octave for RaspberryPi on RaspberryPi OS, so errors
  only abort the users script and don't terminate the whole Octave application.

Windows:

• Psychtoolbox was built and lightly tested against Matlab R2022b.

• GStreamer 1.22 is now required, GStreamer 1.20 will have limitations.

• Fix drawtext plugin again, so it no longer breaks under Matlab with
  GStreamer 1.22 on some systems. This will now require the installation
  of GStreamer 1.22, the older GStreamer 1.20 will no longer work with
  the drawtext plugin.

• Windows: Remove support for building 32-Bit mex files.
  Matlab is 64-Bit only since a long time on Windows, Octave is about to
  remove 32-Bit support as well. Windows-10 - the last MS 32-Bit operating
  system is on the way to final retirement. Ergo, no need for 32-Bit builds
  in the future anymore.

• Make Screen('Openwindow') timing startup tests and calibrations more
  robust. This by disabling processor idling during the tests, ie. ACPI
  C-State processor power management transitions out of C0 (active). Such
  transitions can induce latency / variability in code execution timing bad
  enough to affect timing tests on some setups. If and how much this helps
  in practice remains to be seen. Based on investigations / measurements
  by GitHub user @mirh, see GitHub isse #793 for reference. If this new
  optimization causes trouble or interop problems with cpu performance
  tweaking tools, e.g., the Windows tools "throttlestop", it can be disabled
  by use of the command before the first time you try to open an onscreen
  window in your experiment script:

  PsychTweak('DontDisableProcessorIdling', 1);

macOS:

• Psychtoolbox was built and lightly tested against Matlab R2022b and
  Octave 8.3 from HomeBrew.

• Screen(): Add iMac20,1 and iMac20,2 aka year 2020 iMacs to timing fixup lut,
  so the visual timing fixes also apply to these final Intel iMac models with
  AMD Navi graphics chips. As the PsychtoolboxKernelDriver does not support AMD
  Navi graphics, our visual timestamping and our diagnostic for visual timing
  problems on these machines will be more limited than on older machines, but
  the untested expectation is that this should fix timing on 2020 iMac internal
  displays.

• Screen(): Remove conserveVRAM flag kPsychDontCacheTextures. It was useless and
  even buggy since years, so let it die.

• Screen(): Add new conserveVRAM preference flag 2 == kPsychDontSwitchToOptimalVidMode.
  May or may not help fullscreen display on connected external video splitters like the
  Matrox DualHead2Go. This is based on a hunch, not on proper root causing. Cfe.
  https://psychtoolbox.discourse.group/t/open-window-cant-set-to-a-specified-rect-size-on-macos-with-matrox-dualhead2go/5061

Enjoy!