Drop MP3Codec/libmad from paplayer and dvdplayer in favor of FFmpeg #4499

Merged
merged 10 commits into from May 1, 2014

Projects

None yet

10 participants

@herrnst
Member
herrnst commented Mar 31, 2014

Follow-up to #4386 and initially started by @fritsch - this drops libmad and MP3Codec stuff from paplayer and dvdplayer and rather utilises ffmpeg. While this removes a rather dated dependency (libmad was last updated 2004), the mpg123 decoder inside ffmpeg improves playback of MP3 audio files not being 100% compliant and might play back (partially) distorted when using libmad.

Linux should be fine. OSX (Xcode 5.1) 32bit build tested and working. Win32 project changes need review and testing. Other platforms need to be tested, too.

Probably only for G+1.

Credits belong to @fritsch who started the work :)

@t-nelson
Contributor

Any complaints about reordering the commits so all of the build system stuff is first? That way every commit should compile, which may ease a future bisect.

@t-nelson t-nelson added the Helix label Mar 31, 2014
@jmarshallnz
Member

Personally I'd prefer that we aim for this to be moved to an add-on rather than completely removed.

Ofcourse, the switch to ffmpeg could happen sooner if there's really a bunch of non-compliant files out there (I haven't come across any, but don't use weird encoders), but I'd prefer to leave the support for libmad (ifdef'd out as required) as an incentive for bin add-ons.

IIRC @opdenkamp has some work on audio codecs, as well as ofcourse spiff.

@davilla
Contributor
davilla commented Apr 1, 2014

I'm for making a dependency die off.

@t-nelson
Contributor
t-nelson commented Apr 1, 2014

There's not really a need to keep the stuff around. We have a perfectly
good SCM which it can be recovered from later.

The only benefit I can see to keeping it would be to annoy someone into
doing work. :)

On Mon, Mar 31, 2014 at 6:04 PM, davilla notifications@github.com wrote:

I'm for making a dependency die off.

Reply to this email directly or view it on GitHubhttps://github.com/xbmc/xbmc/pull/4499#issuecomment-39157303
.

@jmarshallnz
Member

The work has already been done. So it's a matter of annoying folk until it's merged.

@wsoltys
Member
wsoltys commented Apr 1, 2014

Does ffmpeg need any external lib for this? If yes we would have to add it to our build script and to our mingw environment.

@herrnst
Member
herrnst commented Apr 1, 2014

@t-nelson Project/depend update commits moved to the top.

@jmarshallnz Probably there are not many files causing issues, but I have at least one candidate that does a garbling sound at one position, which isn't there when playing back with mpg123/ffmpeg (or even tools like winamp). Also - which honestly is a really "synthetic" case - if you use a tool like MP3Gain to increase the gain of a file and over-amplify to like +20db, libmad will produce really screetchy clipping noise, while anything other will clip.

@wsoltys Looking at ffmpeg.git/libavcodec/mpegaudio*, everythings there, but someone with proper ffmpeg knowledge should confirm this ;)

@fritsch
Member
fritsch commented Apr 1, 2014

@wsnipex Could you have a look if our ffmpeg is ready? I tested it with ffmpeg 2.2.x and here everything worked - but not sure if it's build especially cause of some -dev packages I have arround

@opdenkamp
Member

@jmarshallnz re audio bin codec add-on: correct, i'll clean it up and create PRs soon.

@wsnipex
Member
wsnipex commented Apr 1, 2014

configure.in needs cleanup as well. I wonder if we should remove the lame config option and switch cdrip over to ffmpeg as well.

@fritsch we use the default mpegaudio codec in ffmpeg2 and it just works.

@herrnst
Member
herrnst commented Apr 1, 2014

Appended some more smaller commits which need squashing:

docs/README* should also be cleaned up (also still mentions libflac stuff which is also gone already), but IMHO that's stuff for another PR.

@herrnst
Member
herrnst commented Apr 26, 2014

FWIW: Tested this on Linux with the FFmpeg 2.2 stuff. Noticed that a few frames are missing at the beginning of every file (even after trying to rewind to 00:00:00), but besides that everything works flawless.

Also squashed the additional commits as mentioned above a few days ago.

@FernetMenta
Member

why does rewind to 0 not work? would it make any difference if you set the backwards flag to true?
https://github.com/xbmc/xbmc/blob/master/xbmc/cores/paplayer/DVDPlayerCodec.cpp#L266

@herrnst
Member
herrnst commented Apr 27, 2014

Yep, that makes a difference (in fact, files are now played back complete without a larger number of missing frames at the beginning) - commit at herrnst@39e323f

These SeekTime lines gets pushed to xbmc.log before this (when rewinding to zero, 2940 being the last message, smells alot like the offset):

13:02:41 T:140626569565952   DEBUG: SeekTime - seek ended up on time 32600
13:02:41 T:140626569565952   DEBUG: SeekTime - seek ended up on time 2940

after (and without missing audio):

13:04:20 T:140395283470080   DEBUG: SeekTime - seek ended up on time 35566
13:04:21 T:140395283470080   DEBUG: SeekTime - seek ended up on time 5094
13:04:21 T:140395283470080   DEBUG: SeekTime - seek ended up on time 0

Looking at dvdplayer/DVDDemuxers/*.cpp, only the PVR demuxer actually makes use of the flag but that shouldn't be used from within paplayer. FLAC (now also handled by FFmpeg) still seeks fine, can't tell about the other DVDPlayerCodec/FFmpeg formats ATM.

EDIT: mentioned commit appended to this PR/branch.

@herrnst
Member
herrnst commented Apr 27, 2014

PR updated: Last two commits make the backwards flag in ::Seek() being set properly (fixing missing frames), and enables accurate seeking in VBR mp3 files, most notable when playing MP3/CUE files.

@herrnst
Member
herrnst commented Apr 29, 2014

Rebased onto master after ffmpeg bump (conflict in DVDDemuxFFmpeg resolved).

@FernetMenta
Member

jenkins build this please

still angry? (obviously, grrr)

@MartijnKaijser
Member

jenkins build this please

@herrnst
Member
herrnst commented Apr 30, 2014

Hm...

Fetching upstream changes from git://github.com/xbmc/xbmc.git
FATAL: Failed to fetch from git://github.com/xbmc/xbmc.git

... since jenkins doesn't listen to me, someone mind giving him a push again?

@wsnipex
Member
wsnipex commented Apr 30, 2014

jenkins build this please

@FernetMenta FernetMenta merged commit 4060c71 into xbmc:master May 1, 2014

1 check passed

default Merged build #558 succeeded in 1 hr 35 min
Details
@MartijnKaijser MartijnKaijser added this to the Helix 14.0-alpha1 milestone May 1, 2014
@herrnst herrnst deleted the herrnst:remove-mp3codec branch May 1, 2014
@BlackIkeEagle BlackIkeEagle added a commit to herecura-archived/herecura that referenced this pull request May 5, 2014
@BlackIkeEagle BlackIkeEagle xbmc-git :: libmad is not a dependency anymore
see: xbmc/xbmc#4499

Signed-off-by: BlackEagle <ike.devolder@gmail.com>
1a067a5
@FernetMenta
Member

@herrnst looks like gapless is somehow broken. the previous and next track do not fit exactly together. the old code ignored the first 528 samples. do you think we need to do the same with dvdplayer codec for mp3?

Member

@FernetMenta I quick-checked with an official Helix 14.0 build on win32, there indeed is a small gap/audible sort-of blip between track changes even with MP3 files that contain zero-padding information (e.g. live-album ripped to separate tracks, play "real-gapless" in winamp). However, the same audible glitch is also there with FLAC files (which due to their bit-identical lossless compression do not suffer from gap problems by nature), so I suspect something in paplayer. I remember a longer time ago we had a talk about x-fade also having some blip upon end of fading, might be related?

Member

Oh, addendum: "Play with -> dvdplayer", no audible problems.

Member

eh? dvdplayer does not support gapless at all
there is an issue with mp3 and gapless because currently we don not consider decoder delays

Member

Well that's what I found out "as a user". paplayer = blip with both FLAC and MP3 (of course fine with e.g. WinAmp), dvdplayer = no blip, all fine, regardless of the file format.

@candrews candrews added a commit to candrews/xbmc that referenced this pull request Dec 11, 2016
@candrews candrews Remove remaining reference to libmad
libmad was removed in xbmc#4499 - this one reference erroneously remained.
25a5dc2
@candrews candrews referenced this pull request Dec 11, 2016
Merged

Remove remaining reference to libmad #11113

3 of 9 tasks complete
@candrews candrews added a commit to candrews/xbmc that referenced this pull request Dec 11, 2016
@candrews candrews Remove remaining references to libmad
libmad was removed in xbmc#4499
0ddad52
@candrews candrews added a commit to candrews/xbmc that referenced this pull request Dec 11, 2016
@candrews candrews Remove remaining references to libmad
libmad was removed in xbmc#4499
ed553bb
@candrews candrews referenced this pull request Dec 11, 2016
Merged

Remove remaining references to libmad #11132

3 of 9 tasks complete
@raptorjr raptorjr added a commit to raptorjr/xbmc that referenced this pull request Dec 12, 2016
@raptorjr raptorjr Bringing fork up to date (#2)
* VideoPlayer: rtmp fixes

* [PVR] Reintroduce filename sort for PVR recordings

* [PVR][guiinfo] Extend LISTITEM_HASTIMERSCHEDULE to support pvr timer info tag items.

* [Estouchy] update to match the new look of the default skin

* [listprovider] fix job cancelling

* [estuary] sync

* fixed: tighten vcd .dat file checks to avoid unwanted filtering

ref http://forum.kodi.tv/showthread.php?tid=298161

* [Estouchy] Song Info Dialog contributors in control list

* rbp: Don't close the display here it should be closed by the opener

* [DirectoryNodes] override GetChildType() for Episode DirectoryNodes

* aml: introduce speaking constants in H264 4K2K check

* register is deprecated and does nothing but throw a warning with c++11

* fix initializer list orders

* remove extra parantheses to quell compiler warning

* fix format specifiers to quell compiler warnings

* add default case to quell compiler warnings

* add extra braces indicating an initializer list

you just won't shut up will you clang?

* fixed: wrong signature for IoControl method in PipeFile

* fixed: wrong signature for OpenForWrite in RarFile

* fixed: wrong signature for method in GUIWindowPrograms

* fixed: prototype struct as struct

* [lang][kodi.core] automatic syntax corrections for the en_GB language file

* [python] allow setting the videoinfotag path

* CAddonsDirectory: Improve code clarity in GetScriptsAndPlugins()

* CAddonDirectory: Don't clear existing items from item list parameter

* DVDCodecUtils: Fix conversion to av_malloc missed by 09acfb8

* BinaryAddonCache: add function to get add-on by ID and type

* BinaryAddonCache: optimize Update() function

* DVDCodecUtils: Fix buffer overflow if height is odd

* StringUtils: template-ize Join() to support more container types

* PlayMedia() builtin: Don't clear video playlist if item is not video

* [cmake] fix build error when version tag is empty

* [cmake] use app name target for application manifest

* [win32] fix build when optical is disabled

* [win32] fix linking when airtunes or upnp is disabled

* [win32] remove unused has audio define

* [estuary] sync

* bump to v18.0 alpha 1

* [game-settings] Add 'select' as setting type

The newly added settings generator generates settings with type
'select'. These were then not loaded correctly.

* [addon/depends] Handle autocrlf for depends

On Windows, if we have addon dependencies, we need to forcefully disable
autocrlf because otherwise the patches will not apply.

Patches are typically in LF format (and the build system automatically adds
--binary if necessary). But the patches will also not apply if the source files
are in CLRF which can happen if the user has core.autocrlf=true in the
global git config.

A better way of fixing would be to pass the config already to clone:
https://gitlab.kitware.com/cmake/cmake/issues/15799

* [addon/depends] Clone only the last commit of dependencies

This has been introduced with CMake 3.6 but is ignored on older versions.

* use video disptime only in case video is running

* FileBrowser: fix multiple item selection

* [cmake] Move downloading of patch.exe tool into generic tool script

* [cmake] Support @MINGW_TOOLCHAIN_FILE@ in dependency's flags.txt

* [binary addons] Support platform specific deps.txt

* [binary addons] Add msys and mingw dependencies (from kodi's mingw/msys)

* [guilib] Change System.HasModalDialog to System.HasActiveModalDialog and add System.HasVisibleModalDialog

* [estuary] system.hasmodaldialog -> system.hasactivemodaldialog

* [addon] remove not needed header comment in CScreenSaver

The comment on code has confused me a bit and is not really required on so small code.
To prevent misunderstand becomes it removed.

* fixed: do not use move into member variable here

clang points out that it actually hinders copy elusion,
not aids it.

* [doxygen] add pydocs v18 support

* [cleanup][pvr] get rid of obsolete guide views

* [cleanup][settings] remove dead code - options filler 'epgguideviews'

* [Estouchy] update

* Mir windowing system

* [retroplayer] Game settings

* [retroplayer] Keyboard players

* [retroplayer] Game info tags

This is mostly a stop-gap implementation until PyRomInfo
(https://github.com/garbear/pyrominfo) is included.

Thanks to fetzerch for fixing a PVR-related crash.

* [retroplayer] IPlayer: Add game flag to player interface

* [retroplayer] RetroPlayer core

Thanks to elpendor for RGB565 support, poisson for RAII improvements,
ChrisMyhre for catching a compile error, notspiff for CMake fixes,
acmiyaguchi for video and audio codec support, and popcornmix for
Raspberry Pi support (PR 62).

TODO: Is a call to `g_renderManager.IsStarted()` needed?

* [retroplayer] Game add-ons

Thanks to Themaister for rewind functionality, fetzerch for mouse support,
file length check and cmake modifications, topfs2 for fixing a crash when
loading game clients, eibma for fixing linux compilation errors, a1rwulf
for catching a missing callback symbol and fixing some rebase errors, and
to notspiff for helping with the rebrand and cmake.

* [retroplayer/api] Expose GameInfoTag properties to Python list items

* [retroplayer] MyGames window

* [PVR] Context menu rewrite & major refactoring of gui actions.

* [PVR] Context menu rewrite, gui actions refactoring: adsp settings

* [PVR] Context menu rewrite, gui actions refactoring: pvr client menu hook

* [PVR] Context menu rewrite, gui actions refactoring: play recording, play channel, resume recording

* [PVR] Context menu rewrite, gui actions refactoring: recording info

* [PVR] Context menu rewrite: micro opt

* [PVR] Context menu rewrite, gui actions refactoring: delete recording

* [PVR] Context menu rewrite, gui actions refactoring: rename recording

* [PVR] Context menu rewrite, gui actions refactoring: undelete recording

* [PVR] Context menu rewrite, gui actions refactoring: mark watched/unwatched

* [PVR] Context menu rewrite, gui actions refactoring: rename timer

* [PVR] Context menu rewrite, gui actions refactoring: activate / deactivate timer

* [PVR] gui actions refactoring: hide channel

* [PVR] refactored channels window 'Manage...' context menu item code

* [PVR] refactored direct channel number input code (removed logic of derived class from base class)

* [PVR] gui actions refactoring: delete all recordings from trash

* [PVR] gui actions refactoring: cleanup strings.po

* [PVR] Context menu rewrite: Use CStaticContextMenuAction for menu items with static labels.

* [PVR] Context menu rewrite, gui actions refactoring: fileitem usage optimizations.

* [PVR] refactoring: put all pvr and epg shared_ptr typedefs into one header file instead to typedef again and again in several header files.

* [PVR] C++ basics: const vs non-const, reference vs value. :-/

* [settings] move unregister calls into separate methods (similar to register calls)

* [settings] move powermanager defaults from CApplication::Create() to CSettings::InitializeDefaults()

* [cmake] linux: install game headers

* [settings] integrate CSettings into CServiceManager/CServiceBroker

* Revert "This makes the SubtitleDialog remember the last service used, and also the last search instead of always using, and searching, the first service in the list."

This reverts commit 17525c5.

* [aml] windowing: drop EGLNativeTypeAmlAndroid

* [aml] amlcodec: use libamcodec.so

... amlplayer on android is no more.

* [aml] xbmcapp: drop aml workarounds

* [aml] depends: drop libamplayer

* [aml] amlutils: android cleanup

* Enable VAAPI support for Mir using DRM

* [gui] return "" instead of "0" for playlist.position in case nothing is playing

* [doxygen] improve coding guidelines for header includes

* AML: Make DTS-HD, TrueHD work on IEC61937 backport work (more verbose so we can distinguish)

* VideoPlayer: make sure streams are not discarded after a program change

* update splash for v18 alpha

* [NFS] - in case we resolve a symlink - don't accidently manipulate the original dirent pointer because the memory belongs to libnfs and it will remember the change. Instead use a copy of the dirent instance - fixes strange behavior in some symlink scenarios. thx to ronbaby for making it reproducable

* [PVR] Refactor CApplication::PlayMedia to use new PVRGUIActions functionality for playback of PVR channels and recordings.

* [OSX/SDL/Windowing] - ignore resize events that exceed the screen resolution (possible bug in SDL related to OSX HiDP scaled resolutions)

* [WinSystem/OSX] - ensuer the window origin is reset after resizing - fixes windowed mode appearing on the lower bottom of the screen while ignoring the former origin

* [PVR] Fix PVR channel OSD display conditions.

* BinaryAddonCache: Fix deadlock

Release lock before calling CAddonMgr::IsAddonDisabled(), as this may block
in CAddonMgr.

* allow absolute item focus

* FIX: [droid;dyload] never dyload dependent sytem libs

* [epg] Skinning engine: epg grid: add support for alternative (horizontal) grid layout with channels as columns and timeline as rows.

* fixed: tag updates on shoutcast streams with low meta data intervals

a new tag in came in and updated the position before the previous
tag had been processed, meaning the tag never updated after
stream start

* Fix CID 157138: Control flow issues (DEADCODE)

Defect was:

AddonsOperations.cpp: 65 in CAddonsOperations::GetAddons(const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, JSONRPC::ITransportLayer *, JSONRPC::IClient *, const CVariant &, CVariant &)()

59         case ADDON_AUDIO:
60           content = CPluginSource::AUDIO;
61           break;
62         case ADDON_IMAGE:
63           content = CPluginSource::IMAGE;
64           break;
>>>     CID 157138:  Control flow issues  (DEADCODE)
>>>     Execution cannot reach this statement: "case ADDON::ADDON_GAME:".
65         case ADDON_GAME:
66           content = CPluginSource::GAME;
67           break;
68         case ADDON_EXECUTABLE:
69           content = CPluginSource::EXECUTABLE;
70           break;

* Fix CID 157137: Error handling issues (CHECKED_RETURN)

Defect was:

GUIWindowGames.cpp: 258 in GAME::CGUIWindowGames::GetDirectory(const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, CFileItemList &)()

252       if (!CGUIMediaWindow::GetDirectory(strDirectory, items))
253         return false;
254
255       // Set label
256       std::string label;
257       if (items.GetLabel().empty())
>>>     CID 157137:  Error handling issues  (CHECKED_RETURN)
>>>     Calling "IsSource" without checking return value (as is done elsewhere 6 out of 7 times).
258         m_rootDir.IsSource(items.GetPath(), CMediaSourceSettings::GetInstance().GetSources("games"), &label);
259
260       if (!label.empty())
261         items.SetLabel(label);
262
263       // Set content

* Update GameInfoTag

This also fixes defects with CID 157143 amd 157144. Defect was:

*** CID 157143:  Uninitialized members  (UNINIT_CTOR)
GameInfoTag.h: 35 in GAME::CGameInfoTag::CGameInfoTag()()
29     {
30       class CGameInfoTag : public IArchivable,
31                            public ISerializable,
32                            public ISortable
33       {
34       public:
>>>     CID 157143:  Uninitialized members  (UNINIT_CTOR)
>>>     Non-static class member "m_year" is not initialized in this constructor nor in any functions that it calls.
35         CGameInfoTag() { Reset(); }
36         CGameInfoTag(const CGameInfoTag& tag) { *this = tag; }
37         const CGameInfoTag& operator=(const CGameInfoTag& tag);
38         ~CGameInfoTag() { }
39         void Reset();
40

* [estuary] Add missing visible condition

* Settings: Enable keyboard players by default

This allows the simple 2048 game add-on to be played out-of-the-box.

* [cmake] Fix FindXSLT.cmake typo

* [PVR] Fix trac #17108 (duplicate context menu entries for PVR recordings)

* MMALRender: Reduce log spam

* cec: fixup strings after PR 10775

* [Estouchy] Game support

* [estuary] sync

* [cmake] fix: command not found "GIT_SHALLOW 1"

* We should explicitly check/find libva-drm

* CGameLoop: Fix missing destructor

* [cmake] Set correct output directory for core_add_shared_library

Set the correct output-directory for shared libraries on all platforms
(module vs. shared library with different extensions on Darwin) and
for all configurations (Debug, Release, RelWithDebInfo, MinSizeRelease).

* [cmake/ios] Enable Xcode generator for IOS

- Adapt Toolchain file to build a bundles with xcode.
  This change has to be done in the Toolchain file because CMake
  otherwise fails to run some compile checks.
- Unify path where bundle is built for Makefiles and Xcode
  to build/$(CONFIGURATION)-$(EFFECTIVE_PLATFORM_NAME)
- Set only one architecture to prevent building multi-arch bundles.
- Simplify Install.cmake. Existing scrips are called now as post build
  steps so that the bundle can be executed with Xcode
- Apply xcconfig settings.

Note: Two adaptions to the existing "autotools/Xcode" scripts were
needed.
- CMake cannot have the bundle name be different to the binary
  and we cannot call the binary "Kodi.bin". The script has been adapted
  to adapt the library execution path directly in place.
- CMake calls mkdeb.sh with "Debug-iphoneos", hence the script has been
  adapted to match strings starting with "debug" or "release".

Reference:
 - https://cmake.org/Bug/view.php?id=15329
 - http://stackoverflow.com/questions/33660608/how-can-i-disable-xcode-bitcode-within-a-cmake-project

* Fix performace problem with CFileItem::IsGame unnecessarily accessing addon db for PVR content.

* [depends] taglib 1.11.1

* [depends] libpng 1.6.26

* [depends] libass 0.13.4

* [depends] curl 7.51.0

* ListItems: Rename "emulator" property to "gameclient"

This change apparently got dropped in a RetroPlayer rebase. The reason is
because gameclient is more generic than emulator, as the game might be a
launchable asset instead of a ROM. Also, I would expect "emulator" to be
a human-readable string, e.g. "PCSX Re-armed", whereas "gameclient" is
more obviously an add-on ID.

* [pvr] remove add-on lib function typedef's

* [addon] change namespace from V1::KodiAPI:: to KodiAPI::V1::

* [pvr] bye, bye libXBMC_pvr.cpp

* [dependencies] remove remaining references to libmodplug

libmodplug was removed in xbmc#9995 so remove remaining references

* Remove reference to libxinerama-dev

libxinerama-dev is not used so don't document it as a dependency

* Remove remaining references to libmad

libmad was removed in xbmc#4499

* [cmake] fix building when fribidi was built with glib

* VideoPlayer: fix cc after xbmc@03aa244

* [estuary] sync

* [gui] fix coverity issue for playlist.position

* Remove remaining references to glew

glew was (mostly) removed in 03ff0d5; this commit removes any remaining references to it.

* [inputstream] bye, bye libKODI_inputstream.cpp

* [inputstream] remove add-on lib function typedef's

* [addons] remove dead code on CAddonDll

* [cmake] Add support for Bluetooth

* [cec] missed change in cmake file after cc3ae03

* [lang][skin.estuary] automatic syntax corrections for the en_GB language file

* [lang][skin.estuary] updated language files from Transifex

* [lang][skin.estouchy] updated language files from Transifex

* [cmake] Add support for libcap if we find it
d8b085d
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment