1
0
mirror of https://github.com/mpv-player/mpv synced 2024-12-25 00:02:13 +00:00
mpv/DOCS/man/vo.rst

879 lines
36 KiB
ReStructuredText
Raw Normal View History

VIDEO OUTPUT DRIVERS
====================
Video output drivers are interfaces to different video output facilities. The
syntax is:
2013-07-08 16:02:14 +00:00
``--vo=<driver1[:suboption1[=value]:...],driver2,...[,]>``
Specify a priority list of video output drivers to be used.
2013-07-08 16:02:14 +00:00
If the list has a trailing ',', mpv will fall back on drivers not contained
in the list. Suboptions are optional and can mostly be omitted.
You can also set defaults for each driver. The defaults are applied before the
normal driver parameters.
``--vo-defaults=<driver1[:parameter1:parameter2:...],driver2,...>``
Set defaults for each driver.
2013-07-08 16:02:14 +00:00
.. note::
See ``--vo=help`` for a list of compiled-in video output drivers.
The recommended output drivers are ``--vo=vdpau`` and ``--vo=opengl-hq``.
All other drivers are just for compatibility or special purposes.
2013-07-08 16:02:14 +00:00
.. admonition:: Example
``--vo=opengl,xv,``
2013-07-08 16:02:14 +00:00
Try the ``opengl`` driver, then the ``xv`` driver, then others.
Available video output drivers are:
2013-07-08 16:02:14 +00:00
``xv`` (X11 only)
Uses the XVideo extension to enable hardware-accelerated display. This is
2013-07-08 16:02:14 +00:00
the most compatible VO on X, but may be low-quality, and has issues with
OSD and subtitle display.
.. note:: This driver is for compatibility with old systems.
2013-07-08 16:02:14 +00:00
``adaptor=<number>``
Select a specific XVideo adaptor (check xvinfo results).
2013-07-08 16:02:14 +00:00
``port=<number>``
Select a specific XVideo port.
2013-07-08 16:02:14 +00:00
``ck=<cur|use|set>``
Select the source from which the colorkey is taken (default: cur).
cur
The default takes the colorkey currently set in Xv.
use
Use but do not set the colorkey from mpv (use the ``--colorkey``
option to change it).
set
Same as use but also sets the supplied colorkey.
2013-07-08 16:02:14 +00:00
``ck-method=<man|bg|auto>``
Sets the colorkey drawing method (default: man).
man
Draw the colorkey manually (reduces flicker in some cases).
bg
Set the colorkey as window background.
auto
Let Xv draw the colorkey.
``colorkey=<number>``
Changes the colorkey to an RGB value of your choice. ``0x000000`` is
black and ``0xffffff`` is white.
``no-colorkey``
Disables colorkeying.
2013-07-08 16:02:14 +00:00
``x11`` (X11 only)
Shared memory video output driver without hardware acceleration that works
whenever X11 is present.
2013-07-08 16:02:14 +00:00
.. note:: This is a fallback only, and should not be normally used.
2013-07-08 16:02:14 +00:00
``vdpau`` (X11 only)
Uses the VDPAU interface to display and optionally also decode video.
Hardware decoding is used with ``--hwdec=vdpau``.
.. note::
Earlier versions of mpv (and MPlayer, mplayer2) provided sub-options
to tune vdpau postprocessing, like ``deint``, ``sharpen``, ``denoise``,
``chroma-deint``, ``pullup``, ``hqscaling``. These sub-options are
deprecated, and you should use the ``vdpaupp`` video filter instead.
2013-07-08 16:02:14 +00:00
``sharpen=<-1-1>``
(Deprecated. See note about ``vdpaupp``.)
For positive values, apply a sharpening algorithm to the video, for
negative values a blurring algorithm (default: 0).
2013-07-08 16:02:14 +00:00
``denoise=<0-1>``
(Deprecated. See note about ``vdpaupp``.)
2013-07-08 16:02:14 +00:00
Apply a noise reduction algorithm to the video (default: 0; no noise
reduction).
2013-07-08 16:02:14 +00:00
``deint=<-4-4>``
(Deprecated. See note about ``vdpaupp``.)
Select deinterlacing mode (default: 0). In older versions (as well as
MPlayer/mplayer2) you could use this option to enable deinterlacing.
This doesn't work anymore, and deinterlacing is enabled with either
the ``D`` key (by default mapped to the command ``cycle deinterlace``),
or the ``--deinterlace`` option. Also, to select the default deint mode,
you should use something like ``--vf-defaults=vdpaupp:deint-mode=temporal``
instead of this sub-option.
0
Pick the ``vdpaupp`` video filter default, which corresponds to 3.
1
2013-07-08 16:02:14 +00:00
Show only first field.
2
2013-07-08 16:02:14 +00:00
Bob deinterlacing.
3
2013-07-08 16:02:14 +00:00
Motion-adaptive temporal deinterlacing. May lead to A/V desync
with slow video hardware and/or high resolution.
4
2013-07-08 16:02:14 +00:00
Motion-adaptive temporal deinterlacing with edge-guided spatial
interpolation. Needs fast video hardware.
2013-07-08 16:02:14 +00:00
``chroma-deint``
(Deprecated. See note about ``vdpaupp``.)
Makes temporal deinterlacers operate both on luma and chroma (default).
Use no-chroma-deint to solely use luma and speed up advanced
deinterlacing. Useful with slow video memory.
2013-07-08 16:02:14 +00:00
``pullup``
(Deprecated. See note about ``vdpaupp``.)
Try to apply inverse telecine, needs motion adaptive temporal
deinterlacing.
2013-07-08 16:02:14 +00:00
``hqscaling=<0-9>``
(Deprecated. See note about ``vdpaupp``.)
0
Use default VDPAU scaling (default).
1-9
Apply high quality VDPAU scaling (needs capable hardware).
2013-07-08 16:02:14 +00:00
``fps=<number>``
Override autodetected display refresh rate value (the value is needed
for framedrop to allow video playback rates higher than display
refresh rate, and for vsync-aware frame timing adjustments). Default 0
means use autodetected value. A positive value is interpreted as a
refresh rate in Hz and overrides the autodetected value. A negative
value disables all timing adjustment and framedrop logic.
2013-07-08 16:02:14 +00:00
``composite-detect``
NVIDIA's current VDPAU implementation behaves somewhat differently
under a compositing window manager and does not give accurate frame
timing information. With this option enabled, the player tries to
detect whether a compositing window manager is active. If one is
detected, the player disables timing adjustments as if the user had
2013-07-08 16:02:14 +00:00
specified ``fps=-1`` (as they would be based on incorrect input). This
means timing is somewhat less accurate than without compositing, but
2013-07-08 16:02:14 +00:00
with the composited mode behavior of the NVIDIA driver, there is no
hard playback speed limit even without the disabled logic. Enabled by
2013-07-08 16:02:14 +00:00
default, use ``no-composite-detect`` to disable.
``queuetime_windowed=<number>`` and ``queuetime_fs=<number>``
Use VDPAU's presentation queue functionality to queue future video
frame changes at most this many milliseconds in advance (default: 50).
See below for additional information.
2013-07-08 16:02:14 +00:00
``output_surfaces=<2-15>``
Allocate this many output surfaces to display video frames (default:
3). See below for additional information.
``colorkey=<#RRGGBB|#AARRGGBB>``
Set the VDPAU presentation queue background color, which in practice
is the colorkey used if VDPAU operates in overlay mode (default:
``#020507``, some shade of black). If the alpha component of this value
is 0, the default VDPAU colorkey will be used instead (which is usually
green).
``force-yuv``
Never accept RGBA input. This means mpv will insert a filter to convert
to a YUV format before the VO. Sometimes useful to force availability
of certain YUV-only features, like video equalizer or deinterlacing.
Using the VDPAU frame queueing functionality controlled by the queuetime
2013-07-08 16:02:14 +00:00
options makes mpv's frame flip timing less sensitive to system CPU load and
allows mpv to start decoding the next frame(s) slightly earlier, which can
reduce jitter caused by individual slow-to-decode frames. However, the
NVIDIA graphics drivers can make other window behavior such as window moves
choppy if VDPAU is using the blit queue (mainly happens if you have the
composite extension enabled) and this feature is active. If this happens on
your system and it bothers you then you can set the queuetime value to 0 to
disable this feature. The settings to use in windowed and fullscreen mode
are separate because there should be no reason to disable this for
fullscreen mode (as the driver issue should not affect the video itself).
You can queue more frames ahead by increasing the queuetime values and the
2013-07-08 16:02:14 +00:00
``output_surfaces`` count (to ensure enough surfaces to buffer video for a
certain time ahead you need at least as many surfaces as the video has
frames during that time, plus two). This could help make video smoother in
some cases. The main downsides are increased video RAM requirements for
the surfaces and laggier display response to user commands (display
changes only become visible some time after they're queued). The graphics
driver implementation may also have limits on the length of maximum
queuing time or number of queued surfaces that work well or at all.
2013-07-08 16:02:14 +00:00
``direct3d_shaders`` (Windows only)
Video output driver that uses the Direct3D interface.
.. note:: This driver is for compatibility with systems that don't provide
proper OpenGL drivers.
2013-07-08 16:02:14 +00:00
``prefer-stretchrect``
Use ``IDirect3DDevice9::StretchRect`` over other methods if possible.
2013-07-08 16:02:14 +00:00
``disable-stretchrect``
Never render the video using ``IDirect3DDevice9::StretchRect``.
2013-07-08 16:02:14 +00:00
``disable-textures``
Never render the video using D3D texture rendering. Rendering with
textures + shader will still be allowed. Add ``disable-shaders`` to
completely disable video rendering with textures.
2013-07-08 16:02:14 +00:00
``disable-shaders``
Never use shaders when rendering video.
2013-07-08 16:02:14 +00:00
``only-8bit``
Never render YUV video with more than 8 bits per component.
2013-07-08 16:02:14 +00:00
Using this flag will force software conversion to 8-bit.
2013-07-08 16:02:14 +00:00
``disable-texture-align``
Normally texture sizes are always aligned to 16. With this option
enabled, the video texture will always have exactly the same size as
the video itself.
2013-07-08 16:02:14 +00:00
Debug options. These might be incorrect, might be removed in the future,
might crash, might cause slow downs, etc. Contact the developers if you
actually need any of these for performance or proper operation.
2013-07-08 16:02:14 +00:00
``force-power-of-2``
Always force textures to power of 2, even if the device reports
non-power-of-2 texture sizes as supported.
2013-07-08 16:02:14 +00:00
``texture-memory=N``
Only affects operation with shaders/texturing enabled, and (E)OSD.
Values for N:
0
default, will often use an additional shadow texture + copy
1
2013-07-08 16:02:14 +00:00
use ``D3DPOOL_MANAGED``
2
2013-07-08 16:02:14 +00:00
use ``D3DPOOL_DEFAULT``
3
2013-07-08 16:02:14 +00:00
use ``D3DPOOL_SYSTEMMEM``, but without shadow texture
2013-07-08 16:02:14 +00:00
``swap-discard``
Use ``D3DSWAPEFFECT_DISCARD``, which might be faster.
Might be slower too, as it must(?) clear every frame.
2013-07-08 16:02:14 +00:00
``exact-backbuffer``
Always resize the backbuffer to window size.
2013-07-08 16:02:14 +00:00
``direct3d`` (Windows only)
Same as ``direct3d_shaders``, but with the options ``disable-textures``
and ``disable-shaders`` forced.
.. note:: This driver is for compatibility with old systems.
2013-07-08 16:02:14 +00:00
``corevideo`` (Mac OS X 10.6 and later)
Mac OS X CoreVideo video output driver. Uses the CoreVideo APIs to fill
PixelBuffers and generate OpenGL textures from them (useful as a fallback
2013-07-08 16:02:14 +00:00
for ``opengl``).
.. note:: This driver is for compatibility with old systems.
2013-07-08 16:02:14 +00:00
``opengl``
OpenGL video output driver. It supports extended scaling methods, dithering
and color management.
By default, it tries to use fast and fail-safe settings. Use the alias
2012-11-15 13:25:20 +00:00
``opengl-hq`` to use this driver with defaults set to high quality
rendering.
2013-07-08 16:02:14 +00:00
Requires at least OpenGL 2.1 and the ``GL_ARB_texture_rg`` extension. For
older drivers, ``opengl-old`` may work.
Some features are available with OpenGL 3 capable graphics drivers only
(or if the necessary extensions are available).
Hardware decoding over OpenGL-interop is supported to some degree. Note
that in this mode, some corner case might not be gracefully handled, and
colorspace conversion and chroma upsampling is generally in the hand of
the hardware decoder APIs.
2013-07-08 16:02:14 +00:00
``lscale=<filter>``
2013-07-08 16:02:14 +00:00
``bilinear``
Bilinear hardware texture filtering (fastest, mid-quality).
This is the default.
2013-07-08 16:02:14 +00:00
``lanczos2``
Lanczos scaling with radius=2. Provides good quality and speed.
2013-07-08 16:02:14 +00:00
``lanczos3``
Lanczos with radius=3.
``spline36``
This is the default when using ``opengl-hq``.
2013-07-08 16:02:14 +00:00
``bicubic_fast``
Bicubic filter. Has a blurring effect on the image, even if no
scaling is done.
2013-07-08 16:02:14 +00:00
``sharpen3``
Unsharp masking (sharpening) with radius=3 and a default strength
of 0.5 (see ``lparam1``).
2013-07-08 16:02:14 +00:00
``sharpen5``
Unsharp masking (sharpening) with radius=5 and a default strength
of 0.5 (see ``lparam1``).
2013-07-08 16:02:14 +00:00
``mitchell``
Mitchell-Netravali. The ``b`` and ``c`` parameters can be set with
``lparam1`` and ``lparam2``. Both are set to 1/3 by default.
There are some more filters. For a complete list, pass ``help`` as
value, e.g.::
mpv --vo=opengl:lscale=help
2013-07-08 16:02:14 +00:00
``lparam1=<value>``
Set filter parameters. Ignored if the filter is not tunable. These are
unset by default, and use the filter specific default if applicable.
2013-07-08 16:02:14 +00:00
``lparam2=<value>``
See ``lparam1``.
2013-07-08 16:02:14 +00:00
``scaler-resizes-only``
Disable the scaler if the video image is not resized. In that case,
``bilinear`` is used instead whatever is set with ``lscale``. Bilinear
will reproduce the source image perfectly if no scaling is performed.
Note that this option never affects ``cscale``, although the different
processing chain might do chroma scaling differently if ``lscale`` is
disabled.
2013-07-08 16:02:14 +00:00
``stereo=<value>``
Select a method for stereo display. You may have to use
``--video-aspect`` to fix the aspect value. Experimental, do not expect
too much from it.
no
Normal 2D display
red-cyan
Convert side by side input to full-color red-cyan stereo.
green-magenta
Convert side by side input to full-color green-magenta stereo.
quadbuffer
Convert side by side input to quadbuffered stereo. Only supported
by very few OpenGL cards.
2013-07-08 16:02:14 +00:00
``srgb``
vo_opengl: Simplify and clarify color correction code This commit: - Changes some of the #define and variable names for clarification and adds comments where appropriate. - Unifies :srgb and :icc-profile, making them fit into the same step of the decoding process and removing the weird interactions between both of them. - Makes :icc-profile take precedence over :srgb (to significantly reduce the number of confusing and useless special cases) - Moves BT709 decompanding (approximate or actual) to the shader in all cases, making it happen before upscaling (instead of the old 0.45 gamma function). This is the simpler and more proper way to do it. - Enables the approx gamma function to work with :srgb as well due to this (since they now share the gamma expansion code). - Renames :icc-approx-gamma to :approx-gamma since it is no longer tied to the ICC options or LittleCMS. - Uses gamma 2.4 as input space for the actual 3DLUT, this is now a pretty arbitrary factor but I picked 2.4 mainly because a higher pure power value here seems to produce visually better results with wide gamut profiles, rather then the previous 1.95 or BT.709. - Adds the input gamma space to the 3dlut cache header in case we change it more in the future, or even make it user customizable (though I don't see why the latter would really be necessary). - Fixes the OSD's gamma when using :srgb, which was previously still using the old (0.45) approximation in all cases. - Updates documentation on :srgb, it was still mentioning the old behavior from circa a year ago. This commit should serve to both open up and make the CMS/shader code much more accessible and less confusing/error-prone and simultaneously also improve the performance of 3DLUTs with wide gamut color spaces. I would liked to have made it more modular but almost all of these changes are interdependent, save for the documentation updates. Note: Right now, the "3DLUT takes precedence over SRGB" logic is just coded into gl_lcms.c's compile_shaders function. Ideally, this should be done earlier, when parsing the options (by overriding the actual opts.srgb flag) and output a warning to the user. Note: I'm not sure how well this works together with real-world subtitles that may need to be color corrected as well. I'm not sure whether :approx-gamma needs to apply to subtitles as well. I'll need to test this on proper files later. Note: As of now, linear light scaling is still intrinsically tied to either :srgb or :icc-profile. It would be thinkable to have this as an extra option, :linear-scaling or similar, that could be used with or without the two color management options.
2014-03-05 02:56:30 +00:00
Convert and color correct the output to sRGB before displaying it on
the screen. This option enables linear light scaling. It also forces
the options ``indirect`` and ``gamma``.
2013-07-08 16:02:14 +00:00
vo_opengl: Simplify and clarify color correction code This commit: - Changes some of the #define and variable names for clarification and adds comments where appropriate. - Unifies :srgb and :icc-profile, making them fit into the same step of the decoding process and removing the weird interactions between both of them. - Makes :icc-profile take precedence over :srgb (to significantly reduce the number of confusing and useless special cases) - Moves BT709 decompanding (approximate or actual) to the shader in all cases, making it happen before upscaling (instead of the old 0.45 gamma function). This is the simpler and more proper way to do it. - Enables the approx gamma function to work with :srgb as well due to this (since they now share the gamma expansion code). - Renames :icc-approx-gamma to :approx-gamma since it is no longer tied to the ICC options or LittleCMS. - Uses gamma 2.4 as input space for the actual 3DLUT, this is now a pretty arbitrary factor but I picked 2.4 mainly because a higher pure power value here seems to produce visually better results with wide gamut profiles, rather then the previous 1.95 or BT.709. - Adds the input gamma space to the 3dlut cache header in case we change it more in the future, or even make it user customizable (though I don't see why the latter would really be necessary). - Fixes the OSD's gamma when using :srgb, which was previously still using the old (0.45) approximation in all cases. - Updates documentation on :srgb, it was still mentioning the old behavior from circa a year ago. This commit should serve to both open up and make the CMS/shader code much more accessible and less confusing/error-prone and simultaneously also improve the performance of 3DLUTs with wide gamut color spaces. I would liked to have made it more modular but almost all of these changes are interdependent, save for the documentation updates. Note: Right now, the "3DLUT takes precedence over SRGB" logic is just coded into gl_lcms.c's compile_shaders function. Ideally, this should be done earlier, when parsing the options (by overriding the actual opts.srgb flag) and output a warning to the user. Note: I'm not sure how well this works together with real-world subtitles that may need to be color corrected as well. I'm not sure whether :approx-gamma needs to apply to subtitles as well. I'll need to test this on proper files later. Note: As of now, linear light scaling is still intrinsically tied to either :srgb or :icc-profile. It would be thinkable to have this as an extra option, :linear-scaling or similar, that could be used with or without the two color management options.
2014-03-05 02:56:30 +00:00
This option is equivalent to using ``icc-profile`` with an sRGB ICC
profile, but it is implemented without a 3DLUT and does not require
LittleCMS 2. If both ``srgb`` and ``icc-profile`` are present, the
latter takes precedence, as they are somewhat redundant.
Note: When playing back BT.2020 content with this option enabled, out
of gamut colors will be numerically clipped, which can potentially
change the hue and/or luminance. If this is not desired, it is
recommended to use ``icc-profile`` with an sRGB ICC profile instead,
when playing back wide-gamut BT.2020 content.
2013-07-08 16:02:14 +00:00
``pbo``
Enable use of PBOs. This is slightly faster, but can sometimes lead to
sporadic and temporary image corruption (in theory, because reupload
is not retried when it fails), and perhaps actually triggers slower
paths with drivers that don't support PBOs properly.
2013-07-08 16:02:14 +00:00
``dither-depth=<N|no|auto>``
Set dither target depth to N. Default: no.
no
Disable any dithering done by mpv.
auto
2013-07-08 16:02:14 +00:00
Automatic selection. If output bit depth cannot be detected,
8 bits per component are assumed.
8
Dither to 8 bit output.
Note that the depth of the connected video display device can not be
detected. Often, LCD panels will do dithering on their own, which
2013-07-08 16:02:14 +00:00
conflicts with ``opengl``'s dithering and leads to ugly output.
2013-07-08 16:02:14 +00:00
``dither-size-fruit=<2-8>``
Set the size of the dither matrix (default: 6). The actual size of
2013-05-27 21:41:37 +00:00
the matrix is ``(2^N) x (2^N)`` for an option value of ``N``, so a
value of 6 gives a size of 64x64. The matrix is generated at startup
time, and a large matrix can take rather long to compute (seconds).
2013-05-26 17:37:59 +00:00
Used in ``dither=fruit`` mode only.
2013-07-08 16:02:14 +00:00
``dither=<fruit|ordered|no>``
Select dithering algorithm (default: fruit).
2013-07-08 16:02:14 +00:00
``temporal-dither``
Enable temporal dithering. (Only active if dithering is enabled in
general.) This changes between 8 different dithering pattern on each
frame by changing the orientation of the tiled dithering matrix.
Unfortunately, this can lead to flicker on LCD displays, since these
have a high reaction time.
2013-07-08 16:02:14 +00:00
``debug``
Check for OpenGL errors, i.e. call ``glGetError()``. Also request a
debug OpenGL context (which does nothing with current graphics drivers
as of this writing).
2013-07-08 16:02:14 +00:00
``swapinterval=<n>``
2012-08-22 13:45:34 +00:00
Interval in displayed frames between two buffer swaps.
1 is equivalent to enable VSYNC, 0 to disable VSYNC.
2013-07-08 16:02:14 +00:00
``no-scale-sep``
When using a separable scale filter for luma, usually two filter
passes are done. This is often faster. However, it forces
conversion to RGB in an extra pass, so it can actually be slower
if used with fast filters on small screen resolutions. Using
this options will make rendering a single operation.
Note that chroma scalers are always done as 1-pass filters.
2013-07-08 16:02:14 +00:00
``cscale=<n>``
As ``lscale``, but for chroma (2x slower with little visible effect).
Note that with some scaling filters, upscaling is always done in
RGB. If chroma is not subsampled, this option is ignored, and the
luma scaler is used instead. Setting this option is often useless.
2013-07-08 16:02:14 +00:00
``fancy-downscaling``
When using convolution based filters, extend the filter size
when downscaling. Trades quality for reduced downscaling performance.
2013-07-08 16:02:14 +00:00
``no-npot``
Force use of power-of-2 texture sizes. For debugging only.
Borders will be distorted due to filtering.
2013-07-08 16:02:14 +00:00
``glfinish``
Call ``glFinish()`` before swapping buffers
2013-07-08 16:02:14 +00:00
``sw``
Continue even if a software renderer is detected.
2013-07-08 16:02:14 +00:00
``backend=<sys>``
The value ``auto`` (the default) selects the windowing backend. You
can also pass ``help`` to get a complete list of compiled in backends
(sorted by autoprobe order).
auto
auto-select (default)
cocoa
Cocoa/OSX
win
Win32/WGL
x11
X11/GLX
wayland
Wayland/EGL
2013-07-08 16:02:14 +00:00
``indirect``
Do YUV conversion and scaling as separate passes. This will first render
the video into a video-sized RGB texture, and draw the result on screen.
The luma scaler is used to scale the RGB image when rendering to screen.
The chroma scaler is used only on YUV conversion, and only if the video
is chroma-subsampled (usually the case).
This mechanism is disabled on RGB input.
Specifying this option directly is generally useful for debugging only.
2013-07-08 16:02:14 +00:00
``fbo-format=<fmt>``
Selects the internal format of textures used for FBOs. The format can
influence performance and quality of the video output. (FBOs are not
always used, and typically only when using extended scalers.)
``fmt`` can be one of: rgb, rgba, rgb8, rgb10, rgb10_a2, rgb16, rgb16f,
rgb32f, rgba12, rgba16, rgba16f, rgba32f.
Default: rgb.
``gamma=<0.0..10.0>``
Set a gamma value. If gamma is adjusted in other ways (like with
the ``--gamma`` option or keybindings and the ``gamma`` property), the
value is multiplied with the other gamma value.
Setting this value to 1.0 can be used to always enable gamma control.
(Disables delayed enabling.)
2013-07-08 16:02:14 +00:00
``icc-profile=<file>``
Load an ICC profile and use it to transform linear RGB to screen output.
vo_opengl: Simplify and clarify color correction code This commit: - Changes some of the #define and variable names for clarification and adds comments where appropriate. - Unifies :srgb and :icc-profile, making them fit into the same step of the decoding process and removing the weird interactions between both of them. - Makes :icc-profile take precedence over :srgb (to significantly reduce the number of confusing and useless special cases) - Moves BT709 decompanding (approximate or actual) to the shader in all cases, making it happen before upscaling (instead of the old 0.45 gamma function). This is the simpler and more proper way to do it. - Enables the approx gamma function to work with :srgb as well due to this (since they now share the gamma expansion code). - Renames :icc-approx-gamma to :approx-gamma since it is no longer tied to the ICC options or LittleCMS. - Uses gamma 2.4 as input space for the actual 3DLUT, this is now a pretty arbitrary factor but I picked 2.4 mainly because a higher pure power value here seems to produce visually better results with wide gamut profiles, rather then the previous 1.95 or BT.709. - Adds the input gamma space to the 3dlut cache header in case we change it more in the future, or even make it user customizable (though I don't see why the latter would really be necessary). - Fixes the OSD's gamma when using :srgb, which was previously still using the old (0.45) approximation in all cases. - Updates documentation on :srgb, it was still mentioning the old behavior from circa a year ago. This commit should serve to both open up and make the CMS/shader code much more accessible and less confusing/error-prone and simultaneously also improve the performance of 3DLUTs with wide gamut color spaces. I would liked to have made it more modular but almost all of these changes are interdependent, save for the documentation updates. Note: Right now, the "3DLUT takes precedence over SRGB" logic is just coded into gl_lcms.c's compile_shaders function. Ideally, this should be done earlier, when parsing the options (by overriding the actual opts.srgb flag) and output a warning to the user. Note: I'm not sure how well this works together with real-world subtitles that may need to be color corrected as well. I'm not sure whether :approx-gamma needs to apply to subtitles as well. I'll need to test this on proper files later. Note: As of now, linear light scaling is still intrinsically tied to either :srgb or :icc-profile. It would be thinkable to have this as an extra option, :linear-scaling or similar, that could be used with or without the two color management options.
2014-03-05 02:56:30 +00:00
Needs LittleCMS2 support compiled in. This option overrides the ``srgb``
property, as using both is somewhat redundant. It also enables linear
light scaling.
``icc-profile-auto``
Automatically select the ICC display profile currently specified by
the display settings of the operating system.
NOTE: Only implemented on OS X with Cocoa.
2013-07-08 16:02:14 +00:00
``icc-cache=<file>``
Store and load the 3D LUT created from the ICC profile in this file.
This can be used to speed up loading, since LittleCMS2 can take a while
to create the 3D LUT. Note that this file contains an uncompressed LUT.
Its size depends on the ``3dlut-size``, and can be very big.
2013-07-08 16:02:14 +00:00
``icc-intent=<value>``
0
perceptual
1
relative colorimetric (default)
2
saturation
3
absolute colorimetric
vo_opengl: Simplify and clarify color correction code This commit: - Changes some of the #define and variable names for clarification and adds comments where appropriate. - Unifies :srgb and :icc-profile, making them fit into the same step of the decoding process and removing the weird interactions between both of them. - Makes :icc-profile take precedence over :srgb (to significantly reduce the number of confusing and useless special cases) - Moves BT709 decompanding (approximate or actual) to the shader in all cases, making it happen before upscaling (instead of the old 0.45 gamma function). This is the simpler and more proper way to do it. - Enables the approx gamma function to work with :srgb as well due to this (since they now share the gamma expansion code). - Renames :icc-approx-gamma to :approx-gamma since it is no longer tied to the ICC options or LittleCMS. - Uses gamma 2.4 as input space for the actual 3DLUT, this is now a pretty arbitrary factor but I picked 2.4 mainly because a higher pure power value here seems to produce visually better results with wide gamut profiles, rather then the previous 1.95 or BT.709. - Adds the input gamma space to the 3dlut cache header in case we change it more in the future, or even make it user customizable (though I don't see why the latter would really be necessary). - Fixes the OSD's gamma when using :srgb, which was previously still using the old (0.45) approximation in all cases. - Updates documentation on :srgb, it was still mentioning the old behavior from circa a year ago. This commit should serve to both open up and make the CMS/shader code much more accessible and less confusing/error-prone and simultaneously also improve the performance of 3DLUTs with wide gamut color spaces. I would liked to have made it more modular but almost all of these changes are interdependent, save for the documentation updates. Note: Right now, the "3DLUT takes precedence over SRGB" logic is just coded into gl_lcms.c's compile_shaders function. Ideally, this should be done earlier, when parsing the options (by overriding the actual opts.srgb flag) and output a warning to the user. Note: I'm not sure how well this works together with real-world subtitles that may need to be color corrected as well. I'm not sure whether :approx-gamma needs to apply to subtitles as well. I'll need to test this on proper files later. Note: As of now, linear light scaling is still intrinsically tied to either :srgb or :icc-profile. It would be thinkable to have this as an extra option, :linear-scaling or similar, that could be used with or without the two color management options.
2014-03-05 02:56:30 +00:00
``approx-gamma``
Approximate the actual gamma function as a pure power curve of
1.95. A number of video editing programs and studios apparently use this
for mastering instead of the true curve. Most notably, anything in the
Apple ecosystem uses this approximation - including all programs
compatible with it. It's a sound idea to try enabling this flag first
when watching movies and shows to see if things look better that way.
This only affects the output when using either ``icc-profile`` or ``srgb``.
vo_opengl: Simplify and clarify color correction code This commit: - Changes some of the #define and variable names for clarification and adds comments where appropriate. - Unifies :srgb and :icc-profile, making them fit into the same step of the decoding process and removing the weird interactions between both of them. - Makes :icc-profile take precedence over :srgb (to significantly reduce the number of confusing and useless special cases) - Moves BT709 decompanding (approximate or actual) to the shader in all cases, making it happen before upscaling (instead of the old 0.45 gamma function). This is the simpler and more proper way to do it. - Enables the approx gamma function to work with :srgb as well due to this (since they now share the gamma expansion code). - Renames :icc-approx-gamma to :approx-gamma since it is no longer tied to the ICC options or LittleCMS. - Uses gamma 2.4 as input space for the actual 3DLUT, this is now a pretty arbitrary factor but I picked 2.4 mainly because a higher pure power value here seems to produce visually better results with wide gamut profiles, rather then the previous 1.95 or BT.709. - Adds the input gamma space to the 3dlut cache header in case we change it more in the future, or even make it user customizable (though I don't see why the latter would really be necessary). - Fixes the OSD's gamma when using :srgb, which was previously still using the old (0.45) approximation in all cases. - Updates documentation on :srgb, it was still mentioning the old behavior from circa a year ago. This commit should serve to both open up and make the CMS/shader code much more accessible and less confusing/error-prone and simultaneously also improve the performance of 3DLUTs with wide gamut color spaces. I would liked to have made it more modular but almost all of these changes are interdependent, save for the documentation updates. Note: Right now, the "3DLUT takes precedence over SRGB" logic is just coded into gl_lcms.c's compile_shaders function. Ideally, this should be done earlier, when parsing the options (by overriding the actual opts.srgb flag) and output a warning to the user. Note: I'm not sure how well this works together with real-world subtitles that may need to be color corrected as well. I'm not sure whether :approx-gamma needs to apply to subtitles as well. I'll need to test this on proper files later. Note: As of now, linear light scaling is still intrinsically tied to either :srgb or :icc-profile. It would be thinkable to have this as an extra option, :linear-scaling or similar, that could be used with or without the two color management options.
2014-03-05 02:56:30 +00:00
2013-07-08 16:02:14 +00:00
``3dlut-size=<r>x<g>x<b>``
Size of the 3D LUT generated from the ICC profile in each dimension.
Default is 128x256x64.
Sizes must be a power of two, and 512 at most.
``alpha=<blend|yes|no>``
Decides what to do if the input has an alpha component (default: blend).
blend
Blend the frame against a black background.
yes
Try to create a framebuffer with alpha component. This only makes sense
if the video contains alpha information (which is extremely rare). May
not be supported on all platforms. If alpha framebuffers are
unavailable, it silently falls back on a normal framebuffer. Note
that when using FBO indirections (such as with ``opengl-hq``), an FBO
format with alpha must be specified with the ``fbo-format`` option.
no
Ignore alpha component.
2013-07-08 16:02:14 +00:00
``chroma-location=<auto|center|left>``
Set the YUV chroma sample location. auto means use the bitstream
flags (default: auto).
``rectangle-textures``
Force use of rectangle textures (default: no). Normally this shouldn't
have any advantages over normal textures. Note that hardware decoding
overrides this flag.
2013-07-08 16:02:14 +00:00
``opengl-hq``
Same as ``opengl``, but with default settings for high quality rendering.
2013-07-08 16:02:14 +00:00
This is equivalent to::
--vo=opengl:lscale=spline36:dither-depth=auto:fbo-format=rgb16
Note that some cheaper LCDs do dithering that gravely interferes with
2013-07-08 16:02:14 +00:00
``opengl``'s dithering. Disabling dithering with ``dither-depth=no`` helps.
Unlike ``opengl``, ``opengl-hq`` makes use of FBOs by default. Sometimes you
2013-07-08 16:02:14 +00:00
can achieve better quality or performance by changing the ``fbo-format``
suboption to ``rgb16f``, ``rgb32f`` or ``rgb``. Known problems include
Mesa/Intel not accepting ``rgb16``, Mesa sometimes not being compiled with
2013-07-08 16:02:14 +00:00
float texture support, and some OSX setups being very slow with ``rgb16``
but fast with ``rgb32f``.
2013-07-08 16:02:14 +00:00
``opengl-old``
OpenGL video output driver, old version. Video size must be smaller
than the maximum texture size of your OpenGL implementation. Intended to
work even with the most basic OpenGL implementations, but also makes use
2013-07-08 16:02:14 +00:00
of newer extensions which allow support for more color spaces.
The code performs very few checks, so if a feature does not work, this
2013-07-08 16:02:14 +00:00
might be because it is not supported by your card and/or OpenGL
implementation, even if you do not get any error message. Use ``glxinfo``
or a similar tool to display the supported OpenGL extensions.
.. note:: This driver is for compatibility with old systems.
2013-07-08 16:02:14 +00:00
``(no-)ati-hack``
ATI drivers may give a corrupted image when PBOs are used (when using
2013-07-08 16:02:14 +00:00
``force-pbo``). This option fixes this, at the expense of using a bit
more memory.
``(no-)force-pbo``
Always uses PBOs to transfer textures even if this involves an extra
2013-07-08 16:02:14 +00:00
copy. Currently this gives a little extra speed with NVIDIA drivers
and a lot more speed with ATI drivers. May need the ``ati-hack``
suboption to work correctly.
``(no-)scaled-osd``
Scales the OSD and subtitles instead of rendering them at display size
(default: disabled).
``rectangle=<0,1,2>``
Select usage of rectangular textures, which saves video RAM, but often
is slower (default: 0).
0
Use power-of-two textures (default).
1
Use the ``GL_ARB_texture_rectangle`` extension.
2
Use the ``GL_ARB_texture_non_power_of_two`` extension. In some
cases only supported in software and thus very slow.
2013-07-08 16:02:14 +00:00
``swapinterval=<n>``
Minimum interval between two buffer swaps, counted in displayed frames
(default: 1). 1 is equivalent to enabling VSYNC, 0 to disabling VSYNC.
Values below 0 will leave it at the system default. This limits the
framerate to (horizontal refresh rate / n). Requires
``GLX_SGI_swap_control`` support to work. With some (most/all?)
implementations this only works in fullscreen mode.
2013-07-08 16:02:14 +00:00
``ycbcr``
Use the ``GL_MESA_ycbcr_texture`` extension to convert YUV to RGB. In
most cases this is probably slower than doing software conversion to
RGB.
2013-07-08 16:02:14 +00:00
``yuv=<n>``
Select the type of YUV to RGB conversion. The default is
auto-detection deciding between values 0 and 2.
0
Use software conversion. Compatible with all OpenGL versions.
Provides brightness, contrast and saturation control.
1
2013-07-08 16:02:14 +00:00
Same as 2. This used to use NVIDIA-specific extensions, which
did not provide any advantages over using fragment programs, except
possibly on very ancient graphics cards. It produced a gray-ish
output, which is why it has been removed.
2
Use a fragment program. Needs the ``GL_ARB_fragment_program``
extension and at least three texture units. Provides brightness,
contrast, saturation and hue control.
3
2013-07-08 16:02:14 +00:00
Use a fragment program using the ``POW`` instruction. Needs the
``GL_ARB_fragment_program`` extension and at least three texture
units. Provides brightness, contrast, saturation, hue and gamma
control. Gamma can also be set independently for red, green and
blue. Method 4 is usually faster.
4
Use a fragment program with additional lookup. Needs the
``GL_ARB_fragment_program`` extension and at least four texture
units. Provides brightness, contrast, saturation, hue and gamma
control. Gamma can also be set independently for red, green and
blue.
5
Use ATI-specific method (for older cards). This uses an
ATI-specific extension (``GL_ATI_fragment_shader`` - not
``GL_ARB_fragment_shader``!). At least three texture units are
needed. Provides saturation and hue control. This method is fast
but inexact.
6
Use a 3D texture to do conversion via lookup. Needs the
``GL_ARB_fragment_program extension`` and at least four texture
units. Extremely slow (software emulation) on some (all?) ATI
cards since it uses a texture with border pixels. Provides
brightness, contrast, saturation, hue and gamma control. Gamma can
also be set independently for red, green and blue. Speed depends
more on GPU memory bandwidth than other methods.
2013-07-08 16:02:14 +00:00
``lscale=<n>``
Select the scaling function to use for luminance scaling. Only valid
for yuv modes 2, 3, 4 and 6.
0
Use simple linear filtering (default).
1
Use bicubic B-spline filtering (better quality). Needs one
additional texture unit. Older cards will not be able to handle
this for chroma at least in fullscreen mode.
2
Use cubic filtering in horizontal, linear filtering in vertical
direction. Works on a few more cards than method 1.
3
Same as 1 but does not use a lookup texture. Might be faster on
some cards.
4
Use experimental unsharp masking with 3x3 support and a default
2013-07-08 16:02:14 +00:00
strength of 0.5 (see ``filter-strength``).
5
Use experimental unsharp masking with 5x5 support and a default
2013-07-08 16:02:14 +00:00
strength of 0.5 (see ``filter-strength``).
2013-07-08 16:02:14 +00:00
``cscale=<n>``
Select the scaling function to use for chrominance scaling. For
2013-07-08 16:02:14 +00:00
details see ``lscale``.
``filter-strength=<value>``
Set the effect strength for the ``lscale``/``cscale`` filters that
support it.
``stereo=<value>``
Select a method for stereo display. You may have to use
``--video-aspect`` to fix the aspect value. Experimental, do not expect
too much from it.
0
Normal 2D display
1
Convert side by side input to full-color red-cyan stereo.
2
Convert side by side input to full-color green-magenta stereo.
3
Convert side by side input to quadbuffered stereo. Only supported
by very few OpenGL cards.
2013-07-08 16:02:14 +00:00
The following options are only useful if writing your own fragment programs.
2013-07-08 16:02:14 +00:00
``customprog=<filename>``
Load a custom fragment program from ``<filename>``.
2013-07-08 16:02:14 +00:00
``customtex=<filename>``
Load a custom "gamma ramp" texture from ``<filename>``. This can be used
in combination with ``yuv=4`` or with the ``customprog`` option.
``(no-)customtlin``
If enabled (default) use ``GL_LINEAR`` interpolation, otherwise use
``GL_NEAREST`` for customtex texture.
2013-07-08 16:02:14 +00:00
``(no-)customtrect``
If enabled, use ``texture_rectangle`` for the ``customtex`` texture.
Default is disabled.
``(no-)mipmapgen``
If enabled, mipmaps for the video are automatically generated. This
2013-07-08 16:02:14 +00:00
should be useful together with the ``customprog`` and the ``TXB``
instruction to implement blur filters with a large radius. For most
OpenGL implementations, this is very slow for any non-RGB formats.
Default is disabled.
2013-07-08 16:02:14 +00:00
Normally there is no reason to use the following options; they mostly
exist for testing purposes.
2013-07-08 16:02:14 +00:00
``(no-)glfinish``
Call ``glFinish()`` before swapping buffers. Slower but in some cases
more correct output (default: disabled).
2013-07-08 16:02:14 +00:00
``(no-)manyfmts``
Enables support for more (RGB and BGR) color formats (default: enabled).
Needs OpenGL version >= 1.2.
``slice-height=<0-...>``
Number of lines copied to texture in one piece (default: 0). 0 for
whole image.
2013-07-08 16:02:14 +00:00
``sw``
Continue even if a software renderer is detected.
2013-07-08 16:02:14 +00:00
``backend=<sys>``
auto
auto-select (default)
cocoa
Cocoa/OSX
win
Win32/WGL
x11
X11/GLX
2013-03-02 11:11:12 +00:00
wayland
Wayland/EGL
2013-07-08 16:02:14 +00:00
``sdl``
SDL 2.0+ Render video output driver, depending on system with or without
2013-07-08 16:02:14 +00:00
hardware acceleration. Should work on all platforms supported by SDL 2.0.
For tuning, refer to your copy of the file ``SDL_hints.h``.
.. note:: This driver is for compatibility with systems that don't provide
proper graphics drivers, or which support GLES only.
2013-07-08 16:02:14 +00:00
``sw``
Continue even if a software renderer is detected.
2013-07-08 16:02:14 +00:00
``switch-mode``
Instruct SDL to switch the monitor video mode when going fullscreen.
video: add vaapi decode and output support This is based on the MPlayer VA API patches. To be exact it's based on a very stripped down version of commit f1ad459a263f8537f6c from git://gitorious.org/vaapi/mplayer.git. This doesn't contain useless things like benchmarking hacks and the demo code for GLX interop. Also, unlike in the original patch, decoding and video output are split into separate source files (the separation between decoding and display also makes pixel format hacks unnecessary). On the other hand, some features not present in the original patch were added, like screenshot support. VA API is rather bad for actual video output. Dealing with older libva versions or the completely broken vdpau backend doesn't help. OSD is low quality and should be rather slow. In some cases, only either OSD or subtitles can be shown at the same time (because OSD is drawn first, OSD is prefered). Also, libva can't decide whether it accepts straight or premultiplied alpha for OSD sub-pictures: the vdpau backend seems to assume premultiplied, while a native vaapi driver uses straight. So I picked straight alpha. It doesn't matter much, because the blending code for straight alpha I added to img_convert.c is probably buggy, and ASS subtitles might be blended incorrectly. Really good video output with VA API would probably use OpenGL and the GL interop features, but at this point you might just use vo_opengl. (Patches for making HW decoding with vo_opengl have a chance of being accepted.) Despite these issues, decoding seems to work ok. I still got tearing on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also tested with the vdpau vaapi wrapper on a nvidia system; however this was rather broken. (Fortunately, there is no reason to use mpv's VAAPI support over native VDPAU.)
2013-08-09 12:01:30 +00:00
``vaapi``
Intel VA API video output driver with support for hardware decoding. Note
that there is absolutely no reason to use this, other than wanting to use
hardware decoding to save power on laptops, or possibly preventing video
tearing with some setups.
.. note:: This driver is for compatibility with crappy systems. You can
use vaapi hardware decoding with ``--vo=opengl`` too.
video: add vaapi decode and output support This is based on the MPlayer VA API patches. To be exact it's based on a very stripped down version of commit f1ad459a263f8537f6c from git://gitorious.org/vaapi/mplayer.git. This doesn't contain useless things like benchmarking hacks and the demo code for GLX interop. Also, unlike in the original patch, decoding and video output are split into separate source files (the separation between decoding and display also makes pixel format hacks unnecessary). On the other hand, some features not present in the original patch were added, like screenshot support. VA API is rather bad for actual video output. Dealing with older libva versions or the completely broken vdpau backend doesn't help. OSD is low quality and should be rather slow. In some cases, only either OSD or subtitles can be shown at the same time (because OSD is drawn first, OSD is prefered). Also, libva can't decide whether it accepts straight or premultiplied alpha for OSD sub-pictures: the vdpau backend seems to assume premultiplied, while a native vaapi driver uses straight. So I picked straight alpha. It doesn't matter much, because the blending code for straight alpha I added to img_convert.c is probably buggy, and ASS subtitles might be blended incorrectly. Really good video output with VA API would probably use OpenGL and the GL interop features, but at this point you might just use vo_opengl. (Patches for making HW decoding with vo_opengl have a chance of being accepted.) Despite these issues, decoding seems to work ok. I still got tearing on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also tested with the vdpau vaapi wrapper on a nvidia system; however this was rather broken. (Fortunately, there is no reason to use mpv's VAAPI support over native VDPAU.)
2013-08-09 12:01:30 +00:00
``scaling=<algorithm>``
default
Driver default (mpv default as well).
fast
Fast, but low quality.
hq
Unspecified driver dependent high-quality scaling, slow.
nla
``non-linear anamorphic scaling``
``deint-mode=<mode>``
Select deinterlacing algorithm. Note that by default deinterlacing is
initially always off, and needs to be enabled with the ``D`` key
(default key binding for ``cycle deinterlace``).
This option doesn't apply if libva supports video post processing (vpp).
In this case, the default for ``deint-mode`` is ``no``, and enabling
deinterlacing via user interaction using the methods mentioned above
actually inserts the ``vavpp`` video filter. If vpp is not actually
supported with the libva backend in use, you can use this option to
forcibly enable VO based deinterlacing.
video: add vaapi decode and output support This is based on the MPlayer VA API patches. To be exact it's based on a very stripped down version of commit f1ad459a263f8537f6c from git://gitorious.org/vaapi/mplayer.git. This doesn't contain useless things like benchmarking hacks and the demo code for GLX interop. Also, unlike in the original patch, decoding and video output are split into separate source files (the separation between decoding and display also makes pixel format hacks unnecessary). On the other hand, some features not present in the original patch were added, like screenshot support. VA API is rather bad for actual video output. Dealing with older libva versions or the completely broken vdpau backend doesn't help. OSD is low quality and should be rather slow. In some cases, only either OSD or subtitles can be shown at the same time (because OSD is drawn first, OSD is prefered). Also, libva can't decide whether it accepts straight or premultiplied alpha for OSD sub-pictures: the vdpau backend seems to assume premultiplied, while a native vaapi driver uses straight. So I picked straight alpha. It doesn't matter much, because the blending code for straight alpha I added to img_convert.c is probably buggy, and ASS subtitles might be blended incorrectly. Really good video output with VA API would probably use OpenGL and the GL interop features, but at this point you might just use vo_opengl. (Patches for making HW decoding with vo_opengl have a chance of being accepted.) Despite these issues, decoding seems to work ok. I still got tearing on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also tested with the vdpau vaapi wrapper on a nvidia system; however this was rather broken. (Fortunately, there is no reason to use mpv's VAAPI support over native VDPAU.)
2013-08-09 12:01:30 +00:00
no
Don't allow deinterlacing (default for newer libva).
video: add vaapi decode and output support This is based on the MPlayer VA API patches. To be exact it's based on a very stripped down version of commit f1ad459a263f8537f6c from git://gitorious.org/vaapi/mplayer.git. This doesn't contain useless things like benchmarking hacks and the demo code for GLX interop. Also, unlike in the original patch, decoding and video output are split into separate source files (the separation between decoding and display also makes pixel format hacks unnecessary). On the other hand, some features not present in the original patch were added, like screenshot support. VA API is rather bad for actual video output. Dealing with older libva versions or the completely broken vdpau backend doesn't help. OSD is low quality and should be rather slow. In some cases, only either OSD or subtitles can be shown at the same time (because OSD is drawn first, OSD is prefered). Also, libva can't decide whether it accepts straight or premultiplied alpha for OSD sub-pictures: the vdpau backend seems to assume premultiplied, while a native vaapi driver uses straight. So I picked straight alpha. It doesn't matter much, because the blending code for straight alpha I added to img_convert.c is probably buggy, and ASS subtitles might be blended incorrectly. Really good video output with VA API would probably use OpenGL and the GL interop features, but at this point you might just use vo_opengl. (Patches for making HW decoding with vo_opengl have a chance of being accepted.) Despite these issues, decoding seems to work ok. I still got tearing on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also tested with the vdpau vaapi wrapper on a nvidia system; however this was rather broken. (Fortunately, there is no reason to use mpv's VAAPI support over native VDPAU.)
2013-08-09 12:01:30 +00:00
first-field
Show only first field (going by ``--field-dominance``).
bob
bob deinterlacing (default for older libva).
video: add vaapi decode and output support This is based on the MPlayer VA API patches. To be exact it's based on a very stripped down version of commit f1ad459a263f8537f6c from git://gitorious.org/vaapi/mplayer.git. This doesn't contain useless things like benchmarking hacks and the demo code for GLX interop. Also, unlike in the original patch, decoding and video output are split into separate source files (the separation between decoding and display also makes pixel format hacks unnecessary). On the other hand, some features not present in the original patch were added, like screenshot support. VA API is rather bad for actual video output. Dealing with older libva versions or the completely broken vdpau backend doesn't help. OSD is low quality and should be rather slow. In some cases, only either OSD or subtitles can be shown at the same time (because OSD is drawn first, OSD is prefered). Also, libva can't decide whether it accepts straight or premultiplied alpha for OSD sub-pictures: the vdpau backend seems to assume premultiplied, while a native vaapi driver uses straight. So I picked straight alpha. It doesn't matter much, because the blending code for straight alpha I added to img_convert.c is probably buggy, and ASS subtitles might be blended incorrectly. Really good video output with VA API would probably use OpenGL and the GL interop features, but at this point you might just use vo_opengl. (Patches for making HW decoding with vo_opengl have a chance of being accepted.) Despite these issues, decoding seems to work ok. I still got tearing on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also tested with the vdpau vaapi wrapper on a nvidia system; however this was rather broken. (Fortunately, there is no reason to use mpv's VAAPI support over native VDPAU.)
2013-08-09 12:01:30 +00:00
``scaled-osd=<yes|no>``
If enabled, then the OSD is rendered at video resolution and scaled to
display resolution. By default, this is disabled, and the OSD is
rendered at display resolution if the driver supports it.
2013-07-08 16:02:14 +00:00
``null``
Produces no video output. Useful for benchmarking.
2013-07-08 16:02:14 +00:00
``caca``
Color ASCII art video output driver that works on a text console.
.. note:: This driver is a joke.
2013-07-08 16:02:14 +00:00
``image``
2012-08-06 17:15:04 +00:00
Output each frame into an image file in the current directory. Each file
takes the frame number padded with leading zeros as name.
2013-07-08 16:02:14 +00:00
``format=<format>``
2012-08-06 17:15:04 +00:00
Select the image file format.
jpg
2012-11-15 13:25:20 +00:00
JPEG files, extension .jpg. (Default.)
2012-08-06 17:15:04 +00:00
jpeg
JPEG files, extension .jpeg.
png
PNG files.
ppm
Portable bitmap format.
pgm
Portable graymap format.
pgmyuv
Portable graymap format, using the YV12 pixel format.
tga
Truevision TGA.
2013-07-08 16:02:14 +00:00
``png-compression=<0-9>``
2012-08-06 17:15:04 +00:00
PNG compression factor (speed vs. file size tradeoff) (default: 7)
2013-07-08 16:02:14 +00:00
``png-filter=<0-5>``
Filter applied prior to PNG compression (0 = none; 1 = sub; 2 = up;
3 = average; 4 = Paeth; 5 = mixed) (default: 5)
2013-07-08 16:02:14 +00:00
``jpeg-quality=<0-100>``
JPEG quality factor (default: 90)
2013-07-08 16:02:14 +00:00
``(no-)jpeg-progressive``
2012-11-15 13:25:20 +00:00
Specify standard or progressive JPEG (default: no).
2013-07-08 16:02:14 +00:00
``(no-)jpeg-baseline``
2012-11-15 13:25:20 +00:00
Specify use of JPEG baseline or not (default: yes).
2013-07-08 16:02:14 +00:00
``jpeg-optimize=<0-100>``
2012-08-22 13:45:34 +00:00
JPEG optimization factor (default: 100)
2013-07-08 16:02:14 +00:00
``jpeg-smooth=<0-100>``
smooth factor (default: 0)
2013-07-08 16:02:14 +00:00
``jpeg-dpi=<1->``
2012-08-06 17:15:04 +00:00
JPEG DPI (default: 72)
2013-07-08 16:02:14 +00:00
``outdir=<dirname>``
2012-08-06 17:15:04 +00:00
Specify the directory to save the image files to (default: ``./``).
``wayland`` (Wayland only)
Wayland shared memory video output as fallback for ``opengl``.
.. note:: This driver is for compatibility with systems that don't provide
working OpenGL drivers.
``alpha``
Use a buffer format that supports videos and images with alpha
2013-10-16 10:36:34 +00:00
information
2014-02-11 17:55:25 +00:00
``rgb565``
Use RGB565 as buffer format. This format is implemented on most
platforms, especially on embedded where it is far more efficient then
RGB8888.
``triple-buffering``
Use 3 buffers instead of 2. This can lead to more fluid playback, but
uses more memory.