This performs almost 50% faster on my machine (!!), from 4650μs down to
about 3176μs for ewa_lanczossharp.
It's possible we could use a similar approach to speed up the separable
scalers, although with vastly simpler code. For separable scalers we'd
also have the additional huge benefit of only needing padding in one
direction, so we could potentially use a big 256x1 kernel or something
to essentially compute an entire row at once.
This is done via compute shaders. As a consequence, the tone mapping
algorithms had to be rewritten to compute their known constants in GLSL
(ahead of time), instead of doing it once. Didn't affect performance.
Using shmem/SSBO atomics in this way is extremely fast on nvidia, but it
might be slow on other platforms. Needs testing.
Unfortunately, setting up the SSBO still requires OpenGL calls, which
means I can't have it in video_shaders.c, where it belongs. But I'll
defer worrying about that until the backend refactor, since then I'll be
breaking up the video/video_shaders structure anyway.
These can either be invoked as dispatch_compute to do a single
computation, or finish_pass_fbo (after setting compute_size_minimum) to
render to a new texture using a compute shader. To make this stuff all
work transparently, we try really, really hard to make compute shaders
as identical to fragment shaders as possible in their behavior.
Don't use FBOTEX_FUZZY where the FBO is sized according to
p->texture_w/h, since this changes infrequently (and when it does, we
need to reset everything anyway). No real reason to make this change
other than that it possibly prevents nasty surprises in the future, so I
feel more comfortable about it.
Seems like I really like this C99 idiom. No reason not to generalize it
do snprintf(). Introduce mp_tprintf(), which basically this idiom to
snprintf(). This macro looks like it returns a string that was allocated
with alloca() on the caller site, except it's portable C99/C11. (And
unlike alloca(), the result is valid only within block scope.)
Use it in 2 places in the vo_opengl code. But it has the potential to
make a whole bunch of weird looking code look slightly nicer.
Can be enabled via --vd-lavc-dr=yes. See manpage additions for what it
does.
This reminds of the MPlayer -dr flag, but the implementation is
completely different. It's the same basic concept: letting the decoder
render into a GPU buffer to avoid a copy. Unlike MPlayer, this doesn't
try to go through filters (libavfilter doesn't support this anyway).
Unless a filter can work in-place, DR will be silently disabled. MPlayer
had very complex semantics about buffer types and management (which
apparently nobody ever understood) and weird restrictions that mostly
limited it to mpeg2 style codecs. The mpv code does not do any of this,
and just lets the decoder allocate an arbitrary number of untyped
images. (No MPlayer code was used.)
Parts of the code based on work by atomnuker (starting point for the
generic code) and haasn (some GL definitions, some basic PBO code, and
correct fencing).
In addition to using the new VAO mechanism introduced in the previous
commit, this tries to keep the OSD code self-contained. This doesn't
work all too well (because of the pass and CMS stuff), but it's still
better than before.
This removes VAO handling from video.c. Instead the shader cache will
create the VAO as needed. The consequence is that this creates a VAO
per shader, which might be a bit wasteful, but doesn't matter anyway.
Reduce this to 1 draw call per OSD pass. This removes the need for some
annoying special handling regarding 3D video support (we supported
duplicating the OSD/subtitles for side-by-side 3D output etc.).
Remove the unneeded texture sampler uniform thing.
These are apparently expensive on some drivers which are not smart
enough to turn x/42 into x*1.0/42. So, do it for them.
My great test framework says it's okay
This is unnecessary to call from gl_video_resize, because the hooks only
(possibly) change when the actual vo_opengl options change. This used to
be required back when mpv still had prescaling built in, but since that
was all moved to user shaders and the code removed, this is a left-over
artifact.
The renderer code doesn't list a fixed set of supported formats, but
supports anything that is described by AVPixFmtDescriptor and follows a
number of constraints.
Plane order is not included in those constraints. This means the planes
could be in random order, rather than what the vo_opengl renderer
happens to assume. For example, it assumes that the 4th plane is alpha,
even though alpha could be on any plane. Likewise it assumes that plane
0 was always luma, and planes 2/3 chroma. (In earlier iterations of
format support, this was guaranteed by MP_IMGFLAG_YUV_P, but this is not
used anymore.)
Explicitly set the plane semantics (enum plane_type) by the component
descriptors and the used colorspace. The behavior should mostly not
change, but it's less likely to break when FFmpeg adds new pixel
formats.
In commit 6eb0bbe this was changed from xs[n] to use gl_format.chroma_w
indiscriminately, which broke chroma rendering when zooming/cropping.
The solution is to only use chroma_w for chroma planes.
Fixes#4592.
On optional hook points, we store to a temp FBO and then read from it
again to complete any operations that may still be left (e.g.
sigmoidization after MAIN/LINEAR).
In theory this mechanism should be reworked to avoid the temporary FBO
until the next time we actually need one - and also skip redundant
passes if we the next thing we need *is* a FBO - but both are those are
tricky. Anyway, in the meantime, at least we can label the
(semi-)redundant passes that get generated when using user shaders.
It's now possible to request non-dumb mode as a user, even when not
using any non-dumb features. This change is mostly intended for testing,
so I can easily switch between dumb and non-dumb mode on default
settings. The default behavior is unaffected.
This is exposed so that bjin/mpv-prescalers can use textureGatherOffset
for performance.
Since there are now quite a lot of parameters where it isn't quite clear
why they're all defined, add a paragraph to the man page that explains
them a bit.
This helps prevent unnaturally, weirdly colorized blown out highlights
for direct images of the sunlit sky and other way-too-bright HDR
content. I was debating whether to set the default at 1.0 or 2.0, but
went with the more conservative option that preserves more detail/color.
This logic doesn't really make sense. copy_img_tex already binds the
texture, so why would we bind it a second time? Furthermore, nothing
actually uses this return value. Must have been some left-over artifact
of a previous iteration of this function. Anyway, it's harmless, just
nonsensical. So remove it.
This is more efficient on my machine (nvidia), but only when applied to
groups of exactly 4 texels. So we switch to the more efficient
textureGather for groups of 4. Some notes:
- textureGatherOffset seems to be faster than textureGather by a
non-negligible amount, but for some reason, textureOffset is still
slower than a straight-up texture
- textureGather* requires GLSL 400; and at least on nvidia, this
requires actually allocating a GL 4.0 context.
- the code in opengl/common.c that clamped the GLSL version to 330 is
deprecated, because the old user shader style has been removed
completely in the meantime
- To combat the growing complexity of the polar sampling code, we drop
the antiringing functionality from EWA shaders completely, since it
never really worked well for EWA to begin with. (Horrific artifacting)
- change asserts to silent exits
- check all pointers before use
- move the p->pass initialization code to the right place
This should hopefully cut down on the amount of crashing by making the
code fundamentally more robust, while also fixing a concrete issue where
opengl-cb failed to initialize p->pass.
This allows filter functions to be prematurely cut off once their
contributions start becoming insignificant. This effectively prevents
wasted GPU time sampling from parts of the function that are essentially
reduced to zero by the window function, providing anywhere from a 10% to
20% speedup. (5700μs -> 4700μs for me)
2f41c4e8 exposed some other edge cases as well. Globally resetting the
pass info was not the right way to go about it, because we don't know in
advance what the frame type is going to be - at least not with the
current code structure. (In principle, we could separately indicate the
frame type and the pass type and then only reset it on the first
actual pass_describe call, but that's annoying as well)
Also fixes a latent issue where p->pass was never initialized, which
broke the MP_DBG debugging code in some cases.
Since all existing code does gl_video_upload immediately followed by
pass_render_frame, we can just move the upload into pass_render_frame
itself, which arguably makes more sense anyway.
This replaces `vo-performance` by `vo-passes`, bringing with it a number
of changes and improvements:
1. mpv users can now introspect the vo_opengl passes, which is something
that has been requested multiple times.
2. performance data is now measured per-pass, which helps both
development and debugging.
3. since adding more passes is cheap, we can now report information for
more passes (e.g. the blit pass, and the osd pass). Note: we also
switch to nanosecond scale, to be able to measure these passes
better.
4. `--user-shaders` authors can now describe their own passes, helping
users both identify which user shaders are active at any given time
as well as helping shader authors identify performance issues.
5. the timing data per pass is now exported as a full list of samples,
so projects like Argon-/mpv-stats can immediately read out all of the
samples and render a graph without having to manually poll this
option constantly.
Due to gl_timer's design being complicated (directly reading performance
data would block, so we delay the actual read-back until the next _start
command), it's vital not to conflate different passes that might be
doing different things from one frame to another. To accomplish this,
the actual timers are stored as part of the gl_shader_cache's sc_entry,
which makes them unique for that exact shader.
Starting and stopping the time measurement is easy to unify with the
gl_sc architecture, because the existing API already relies on a
"generate, render, reset" flow, so we can just put timer_start and
timer_stop in sc_generate and sc_reset, respectively.
The ugliest thing about this code is that due to the need to keep pass
information relatively stable in between frames, we need to distinguish
between "new" and "redrawn" frames, which bloats the code somewhat and
also feels hacky and vo_opengl-specific. (But then again, this entire
thing is vo_opengl-specific)
Instead of setting up a weird swizzle (which is linked to how the
internal renderer code works, rather than the generic format code), add
per-component mapping to gl_imgfmt_desc.
The renderer still computes the weird swizzle, but at least it's
confined to itself. Also, it appears the hwdec backends don't need this
anymore.
It's really nice that the messy init_format() goes away too.
The changes to path list options is basically getting rid of the need to
pass multiple paths to a single option. Instead, you can use the option
multiple times. The old behavior can be used by using the -set suffix
with the option.
Change some options to path lists. For example --script is now append by
default, and if you use --script-set, you need to use ":"/";" as
separator instead of ",".
--sub-paths/--audio-file-paths is a deprecated alias now, and will break
if the user tries to pass multiple paths to it. I'm assuming that if
these are used, most users will pass only 1 path anyway.
--opengl-shaders has more compatibility handling, since it's probably
rather common that users pass multiple options to it.
Also document all that in the manpage.
I'll probably regret this later, as it somewhat increases the complexity
of the option parser, rather than increasing it.
Commit 3fb6380 was supposed to increase MAX_TEXTURE_HOOKS but instead
increased SHADER_MAX_HOOKS, since I forgot that they were separate (for
whatever reason).
To prevent this mistake from happening again, and to unify the location
in which user_shader-specific #defines are placed, get rid of the two
constants in opengl/video.c and move/reuse them from user_shaders.h
instead.
Also bump up MAX_SAVED_TEXTURES (now SHADER_MAX_SAVED) slightly as a
precaution against adding more passes to vo_opengl. I think we're
already flirting with the limit.
These can never be uninitialized because the enum cases are exhaustive and
the fallback is in the correct order, but GCC is too dumb to understand
this.
Also explicitly initialize tex_type, because while GCC doesn't warn
about it (for some reason), maybe it will in the future.
This introduces (yet another..) mp_colorspace members, an enum `light`
(for lack of a better name) which basically tells us whether we're
dealing with scene-referred or display-referred light, but also a bit
more metadata (in which way is the scene-referred light expected to be
mapped to the display?).
The addition of this parameter accomplishes two goals:
1. Allows us to actually support HLG more-or-less correctly[1]
2. Allows people playing back direct “camera” content (e.g. v-log or
s-log2) to treat it as scene-referred instead of display-referred
[1] Even better would be to use the display-referred OOTF instead of the
idealized OOTF, but this would require either native HLG support in
LittleCMS (unlikely) or more communication between lcms.c and
video_shaders.c than I'm remotely comfortable with
That being said, in principle we could switch our usage of the BT.1886
EOTF to the BT.709 OETF instead and treat BT.709 content as being
scene-referred under application of the 709+1886 OOTF; which moves that
particular conversion from the 3dlut to the shader code; but also allows
a) users like UliZappe to turn it off and b) supporting the full HLG
OOTF in the same framework. But I think I prefer things as they are
right now.
st2084 and std-b67 are really weird names for PQ and HLG, which is what
everybody else (including e.g. the ITU-R) calls them. Follow their
example.
I decided against naming them bt2020-pq and bt2020-hlg because it's not
necessary in this case. The standard name is only used for the other
colorspaces etc. because those literally have no other names.
List of changes:
1. Kill nom_peak, since it's a pointless non-field that stores nothing
of value and is _always_ derived from ref_white anyway.
2. Kill ref_white/--target-brightness, because the only case it really
existed for (PQ) actually doesn't need to be this general: According
to ITU-R BT.2100, PQ *always* assumes a reference monitor with a
white point of 100 cd/m².
3. Improve documentation and comments surrounding this stuff.
4. Clean up some of the code in general. Move stuff where it belongs.
In GLES 2 mode, we can do dither, but "fruit" dithering is still out of
the question, because it does not support any high depth textures.
(Actually we probably could use an 8 bit texture too for this, at least
with small matrix sizes, but it's still too much of a pain to convert
the data, so why bother.)
This is actually a regression; before this, forcibly enabling dumb mode
due to low GL caps actually happened to avoid this case.
Fixes#4519.
I call it `mobius` because apparently the form f(x) = (cx+a)/(dx+b) is
called a Möbius transform, which is the algorithm this is based on. In
the extremes it becomes `reinhard` (param=0.0 and `clip` (param=1.0),
smoothly transitioning between the two depending on the parameter.
This is a useful tone mapping algorithm since the tunable mobius
transform allows the user to decide the trade-off between color accuracy
and detail preservation on a continuous scale. The default of 0.3 is
already far more accurate than `reinhard` while also being reasonably
good at preserving highlights, without suffering from the overall
brightness drop and color distortion of `hable`.
For these reasons, make this the new default. Also expand and improve
the documentation for these tone mapping functions.
This reverts commit 142b2f23d4293, and replaces it with another try. The
previous attempt removed the overlay on every rendering, because the
normal rendering path actually unrefs the mp_image. Consequently,
unmap_current_image() was completely inappropriate for removing the
overlay.
Mostly because of ANGLE (sadly).
The implementation became unpleasantly big, but at least it's relatively
self-contained.
I'm not sure to what degree shaders from different drivers are
compatible as in whether a driver would randomly misbehave if it's fed
a binary created by another driver. The useless binayFormat parameter
won't help it, as they can probably easily clash. As usual, OpenGL is
pretty shit here.