a17337910c
During probing on a system with AMD GPU, mpv used to output the following messages if hardware decoding was enabled: [ffmpeg] AVHWFramesContext: Failed to create surface: 2 (resource allocation failed). [ffmpeg] AVHWFramesContext: Unable to allocate a surface from internal buffer pool. This commit removed the message, with hopefully no other side effects. Long explanations follow, better don't read them, it's just tedious drivel about the details. People should learn to write concise commit messages, not drone on and on endlessly all while they have no fucking point. The code probes supported hardware pixel format, and checks whether they can be mapped as textures. av_hwdevice_get_hwframe_constraints() returns a list of hardware pixel formats in the valid_sw_formats field (the "sw" means software, but they're still hardware pixel formats, makes sense). This contained the format yuv420p, even though this is not a valid hardware format. Trying to create a surface of this type results in VA surface creation failure, upon which FFmpeg prints the error messages above. We'd be fine with this, except FFmpeg has a global log callback, and there's no way to suppress these messages without creating other issues. It turns out that FFmpeg's vaapi implementation returns all formats from vaQueryImageFormats() if no "hwconfig" is provided. This list includes yuv420p, which is probably supported for surface upload/download, but not as native format. Following FFmpeg's logic, it should not appear in the valid_sw_formats list, because formats for transfers are returned by another roundabout API. Idiotically, there doesn't seem to be any vaapi call that determines whether a format is a valid surface format. All mechanisms to do this are bound to a VAConfigID (= video codec or video processor), all while the actual surface creation API strangely does not take a VAConfigID (a big WTF). Also, calling the vaCreateSurfaces() API ourselves for probing is out of the question, because that functions is utterly and idiotically complex. Look at the FFmpeg code and how much effort it requires to setup a complete set of attributes - we can't duplicate this. So the only way left to do this is the most idiotic and tedious way: enumerating all VAProfile (and VAEntrypoints) to create all possible VAConfigIDs. Each of the VAConfigIDs is associated with a list of formats, which FFmpeg can return (by passing the ID along with the "hwconfig"), and which is probed separately. Note that VAConfigID actually refers to a dynamic instance of something, and creating a VAConfigID takes not only the VAProfile and the VAEntrypoint, but also an arbitrary attribute array. In theory, this means our attempt to get to know all possible configurations cannot work, but in practice this attribute array seems to be pointless for decoding and video processing, and FFmpeg doesn't use it (though the encoding path does use it). This probably just makes it _barely_ OK to do it this way. Could we discard all this probing shit, and somehow do it another way? Probably not. The EGL API for mapping surfaces doesn't even seem to provide a way to enumerate supported formats, we may not even know whether DRM/dmabuf interop is actually supported (AFAIR the EGL extensions are present even if they don't work), nor do we know whether the VAAPI driver supports this interop (not sure). So actually trying is the only way. Further, mpv initializes the decoder on a another thread, where you can't just access OpenGL state. This suckage is mostly to be blamed on OpenGL itself and its crazy thread boundedness. In theory, this could be done anyway (see how software decoding "direct rendering" tries to get around this). But to make it worse, the decoder never cares about the list of supported formats determined by this code; instead, f_autoconvert.c tries to deal with it and insert a video processor (well, good luck with this crap, I bet it doesn't even work). So this whole endeavor might be pointless, other than the fact that failed probing can disable use of vaapi (which is correct and necessary). But if you have a shovel, you don't use it to smash the flat end on the heap of shit that's piled up before you, or do you? While this method probably works, it's still orgasmically tedious. It was tedious before: we had to create a real surface, create a GL texture, map the surface with it, then destroy everything again. But the added code is tedious on its own. Highlights include the need to malloc a FFmpeg struct just to pass a single damn integer, the need to enumerate "entrypoints" for each VA profile, even though all profiles have exactly 1 entrypoint, and the kind of obnoxious way how vaapi requires you to preallocate arrays for returned things, even they could for example reasonably be returned as immutable arrays or have some other simpler API. The main grand fuckup is of course that vaapi requires a VAConfigID to query surface properties, but not for creating surfaces. This awkwardness even affected the FFmpeg API design, which has a "hwconfig" concept that is only used by vaapi (vaapi is only 1 out of 10 hardware decoding APIs supported by the FFmpeg hwcontext stuff). Maybe I'm just missing something. It's as if vaapi required setting radioactive shit on fire. Look how clean the native D3D11 code is instead. (Even the ANGLE code manages to avoid being this fucked up. Or the VDPAU code, despite supporting multiple mapping methods.) Another only barely related change is that the valid_sw_formats field can be NULL, and the API explicitly documents this. Technically, the mpv code was buggy for not checking this, although until now the FFmpeg implementation so far could not return it when we still passed NULL for the hwconfig parameter. |
||
|---|---|---|
| .github | ||
| audio | ||
| ci | ||
| common | ||
| demux | ||
| DOCS | ||
| etc | ||
| filters | ||
| input | ||
| libmpv | ||
| misc | ||
| options | ||
| osdep | ||
| player | ||
| stream | ||
| sub | ||
| ta | ||
| test | ||
| TOOLS | ||
| video | ||
| waftools | ||
| .gitignore | ||
| .travis.yml | ||
| appveyor.yml | ||
| bootstrap.py | ||
| Copyright | ||
| LICENSE.GPL | ||
| LICENSE.LGPL | ||
| mpv_talloc.h | ||
| README.md | ||
| RELEASE_NOTES | ||
| VERSION | ||
| version.sh | ||
| wscript | ||
| wscript_build.py | ||
mpv
- External links
- Overview
- System requirements
- Downloads
- Changelog
- Compilation
- FFmpeg vs. Libav
- FFmpeg ABI compatibility
- Release cycle
- Bug reports
- Contributing
- Relation to MPlayer and mplayer2
- License
- Contact
External links
Overview
mpv is a media player based on MPlayer and mplayer2. It supports a wide variety of video file formats, audio and video codecs, and subtitle types.
Releases can be found on the release list.
System requirements
- A not too ancient Linux, Windows 7 or later, or OSX 10.8 or later.
- A somewhat capable CPU. Hardware decoding might help if the CPU is too slow to
decode video in realtime, but must be explicitly enabled with the
--hwdecoption. - A not too crappy GPU. mpv is not intended to be used with bad GPUs. There are
many caveats with drivers or system compositors causing tearing, stutter,
etc. On Windows, you might want to make sure the graphics drivers are
current. In some cases, ancient fallback video output methods can help
(such as
--vo=xvon Linux), but this use is not recommended or supported.
Downloads
For semi-official builds and third-party packages please see mpv.io/installation.
Changelog
There is no complete changelog; however, changes to the player core interface are listed in the interface changelog.
Changes to the C API are documented in the client API changelog.
The release list has a summary of most of the important changes on every release.
Changes to the default key bindings are indicated in restore-old-bindings.conf.
Compilation
Compiling with full features requires development files for several external libraries. Below is a list of some important requirements.
The mpv build system uses waf, but we don't store it in the
repository. The ./bootstrap.py script will download the latest version
of waf that was tested with the build system.
For a list of the available build options use ./waf configure --help. If
you think you have support for some feature installed but configure fails to
detect it, the file build/config.log may contain information about the
reasons for the failure.
NOTE: To avoid cluttering the output with unreadable spam, --help only shows
one of the two switches for each option. If the option is autodetected by
default, the --disable-*** switch is printed; if the option is disabled by
default, the --enable-*** switch is printed. Either way, you can use
--enable-*** or --disable-** regardless of what is printed by --help.
To build the software you can use ./waf build: the result of the compilation
will be located in build/mpv. You can use ./waf install to install mpv
to the prefix after it is compiled.
Example:
./bootstrap.py
./waf configure
./waf
./waf install
Essential dependencies (incomplete list):
- gcc or clang
- X development headers (xlib, xrandr, xext, xscrnsaver, xinerama, libvdpau, libGL, GLX, EGL, xv, ...)
- Audio output development headers (libasound/ALSA, pulseaudio)
- FFmpeg libraries (libavutil libavcodec libavformat libswscale libavfilter and either libswresample or libavresample)
- zlib
- iconv (normally provided by the system libc)
- libass (OSD, OSC, text subtitles)
- Lua (optional, required for the OSC pseudo-GUI and youtube-dl integration)
- libjpeg (optional, used for screenshots only)
- uchardet (optional, for subtitle charset detection)
- vdpau and vaapi libraries for hardware decoding on Linux (optional)
Libass dependencies:
- gcc or clang, yasm on x86 and x86_64
- fribidi, freetype, fontconfig development headers (for libass)
- harfbuzz (optional, required for correct rendering of combining characters, particularly for correct rendering of non-English text on OSX, and Arabic/Indic scripts on any platform)
FFmpeg dependencies:
- gcc or clang, yasm on x86 and x86_64
- OpenSSL or GnuTLS (have to be explicitly enabled when compiling FFmpeg)
- libx264/libmp3lame/libfdk-aac if you want to use encoding (have to be explicitly enabled when compiling FFmpeg)
- For native DASH playback, FFmpeg needs to be built with --enable-libxml2 (although there are security implications).
- For good nvidia support on Linux, make sure nv-codec-headers is installed and can be found by configure.
- Libav support is broken. (See section below.)
Most of the above libraries are available in suitable versions on normal Linux distributions. For ease of compiling the latest git master of everything, you may wish to use the separately available build wrapper (mpv-build) which first compiles FFmpeg libraries and libass, and then compiles the player statically linked against those.
If you want to build a Windows binary, you either have to use MSYS2 and MinGW, or cross-compile from Linux with MinGW. See Windows compilation.
FFmpeg vs. Libav
Generally, mpv should work with the latest release as well as the git version of FFmpeg. Libav support is currently broken, because they did not add certain FFmpeg API changes which mpv relies on.
FFmpeg ABI compatibility
mpv does not support linking against FFmpeg versions it was not built with, even if the linked version is supposedly ABI-compatible with the version it was compiled against. Expect malfunctions, crashes, and security issues if you do it anyway.
The reason for not supporting this is because it creates far too much complexity with little to no benefit, coupled with absurd and unusable FFmpeg API artifacts.
Newer mpv versions will refuse to start if runtime and compile time FFmpeg library versions mismatch.
Release cycle
Every other month, an arbitrary git snapshot is made, and is assigned a 0.X.0 version number. No further maintenance is done.
The goal of releases is to make Linux distributions happy. Linux distributions are also expected to apply their own patches in case of bugs and security issues.
Releases other than the latest release are unsupported and unmaintained.
See the release policy document for more information.
Bug reports
Please use the issue tracker provided by GitHub to send us bug reports or feature requests. Follow the template's instructions or the issue will likely be ignored or closed as invalid.
Using the bug tracker as place for simple questions is fine but IRC is recommended (see Contact below).
Contributing
Please read contribute.md.
For small changes you can just send us pull requests through GitHub. For bigger changes come and talk to us on IRC before you start working on them. It will make code review easier for both parties later on.
You can check the wiki or the issue tracker for ideas on what you could contribute with.
Relation to MPlayer and mplayer2
mpv is a fork of MPlayer. Much has changed, and in general, mpv should be considered a completely new program, rather than a MPlayer drop-in replacement.
For details see FAQ entry.
If you are wondering what's different from mplayer2 and MPlayer, an incomplete and largely unmaintained list of changes is located here.
License
GPLv2 "or later" by default, LGPLv2.1 "or later" with --enable-lgpl.
See details.
Contact
Most activity happens on the IRC channel and the github issue tracker.
- GitHub issue tracker: issue tracker (report bugs here)
- User IRC Channel:
#mpvonirc.freenode.net - Developer IRC Channel:
#mpv-develonirc.freenode.net