This was needlessly complicated and prone to breakage, because even the
references to the ring buffer could end up getting invalidated and
containing garbage data on e.g. shader cache flush. For much the same
reason why we can't keep around the *timer_pool, we're also forced to
hard-copy the entire sample buffer per pass per frame.
Not a huge deal, though. This is, what, a few kB per frame? We have more
pressing CPU performance concerns anyway.
Also simplified/fixed some other code.
This clearly highlights all out-of-gamut/clipped pixels. (Either too
bright or too saturated)
Has some (documented) caveats. Also make TONE_MAPPING_CLIP stop actually
clamping the value range (it's unnecessary and breaks this feature).
Redefining texture1D / texture3D seems to be illegal, they are already
built-in macros or something. So just use tex1D and tex3D instead.
Additionally, GL_KHR_vulkan_glsl requires using explicit vertex
locations and bindings, so make some changes to facilitate this. (It
also requires explicitly setting location=0 for the color attachment
output)
Vulkan compat. rgb16 doesn't exist on hardware anyway, might as well
just generate the 3DLUT against rgba16 as well. We've decided this is
the simplest way to do vulkan compatibility: just make sure we never
actually need 3-component textures.
This is mostly done so we can support using textures with more
components than the scaler LUTs have entries. But while we're at it,
also change the way the weights are packed so that they're always
sequential with no gaps. This allows us to simplify
pass_sample_separated_get_weights as well.
If a VO-area option changes, gl_video_resize() is called
unconditionally. This function does something even if the size does not
change (at least it discards buffered frames for interpolation), which
can lead to stutter when you keep firing option change events during
playback.
Check for an actual resize, and if nothing changes, exit early.
Could cause a crash if anything called ra_get_imgfmt_desc(imgfmt=0). Let
it fail correctly. This can happen if a hwdec backend does not set
hw_subfmt correctly.
This also introduces RA_CAP_GLOBAL_UNIFORM. If this is not set, UBOs
*must* be used for non-bindings. Currently the cap is ignored though,
and the shader_cache *always* generates UBO-using code where it can.
Could be made an option in principle.
Only enabled for drivers new enough to support explicit UBO offsets,
just in case...
No change to performance, which is probably what we expect.
This no longer concerns the API user except in as much as the API user
probably wants to know whether or not PBOs are active, so keep around
the CAP field even though it's mostly useless now.
Both vulkan and opengl distinguish between rendering to an image and
using an image as a storage attachment. So make this an explicit
capability instead of lumping it in with render_dst. (That way we could
support, for example, using an image as a storage attachment without
requiring a framebuffer)
The real reason for this change is that you can directly use the output
FBO as a storage attachment on vulkan but you can't on opengl, which
makes this param structly separate from render_dst.
I don't like the feeling of "reusing" the int binding for this. It
feels... wrong, somehow. I'd prefer to use an explicit "offset" field.
(Plus, I might re-use this for uniform buffers or something)
YMMV
I really wouldn't care much about this, but some parts of the core code
are under HAVE_GPL, so there's some need to get rid of it. Simply turn
the video equalizer from its current fine-grained handling with vf/vo
fallbacks into global options. This makes updating them much simpler.
This removes any possibility of applying video equalizers in filters,
which affects vf_scale, and the previously removed vf_eq. Not a big
loss, since the preferred VOs have this builtin.
Remove video equalizer handling from vo_direct3d, vo_sdl, vo_vaapi, and
vo_xv. I'm not going to waste my time on these legacy VOs.
vo.eq_opts_cache exists _only_ to send a VOCTRL_SET_EQUALIZER, which
exists _only_ to trigger a redraw. This seems silly, but for now I feel
like this is less of a pain. The rest of the equalizer using code is
self-updating.
See commit 96b906a51d for how some video equalizer code was GPL only.
Some command line option names and ranges can probably be traced back to
a GPL only committer, but we don't consider these copyrightable.
So far, we had a thread-safe way to read options, but no option update
notification mechanism. Everything was funneled though the main thread's
central mp_option_change_callback() function. For example, if the
panscan options were changed, the function called vo_control() with
VOCTRL_SET_PANSCAN to manually notify the VO thread of updates. This
worked, but's pretty inconvenient. Most of these problems come from the
fact that MPlayer was written as a single-threaded program.
This commit works towards a more flexible mechanism. It adds an update
callback to m_config_cache (the thing that is already used for
thread-safe access of global options).
This alone would still be rather inconvenient, at least in context of
VOs. Add another mechanism on top of it that uses mp_dispatch_queue, and
takes care of some annoying synchronization issues. We extend
mp_dispatch_queue itself to make this easier and slightly more
efficient.
As a first application, use this to reimplement certain VO scaling and
renderer options. The update_opts() function translates these to the
"old" VOCTRLs, though.
An annoyingly subtle issue is that m_config_cache's destructor now
releases pending notifications, and must be released before the
associated dispatch queue. Otherwise, it could happen that option
updates during e.g. VO destruction queue or run stale entries, which is
not expected.
Rather untested. The singly-linked list code in dispatch.c is probably
buggy, and I bet some aspects about synchronization are not entirely
sane.
Also refactors the usage of tex_upload to make ra_tex_upload_pbo a
RA-internal thing again.
ra_buf_pool has the main advantage of being dynamically sized depending
on buf_poll, so for OpenGL we'll end up only using one buffer (when not
persistently mapping) - while for vulkan we'll use as many as necessary,
which depends on the swapchain depth anyway.
- tex_uploads args are moved to a struct
- the ability to directly upload texture data without going through a
buffer is made explicit
- the concept of buffer updates and buffer polling is made more explicit
and generalized to buf_update as well (not just mapped buffers)
- the ability to call tex_upload/buf_update on a tex/buf is made
explicit during tex/buf creation
- uploading from buffers now uses an explicit offset instead of
implicitly comparing *src against buf->data, because not all buffers
may actually be persistently mapped
- the initial_data = immutable requirement is dropped. (May be re-added
later for D3D11 if that ever becomes a thing)
This change helps the vulkan abstraction immensely and also helps move
common code (like the PBO pooling) out of ra_gl and into the
opengl/utils.c
This also technically has the side-benefit / side-constraint of using
PBOs for OSD texture uploads as well, which actually seems to help
performance on machines where --opengl-pbo is faster than the naive code
path. Because of this, I decided to hook up the OSD code to the
opengl-pbo option as well.
One drawback of this refactor is that the GL_STREAM_COPY hack for
texture uploads "got lost", but I think I'm happy with that going away
anyway since DR almost fully deprecates it, and it's not the "right
thing" anyway - but instead an nvidia-only hack to make this stuff work
somewhat better on NUMA systems with discrete GPUs.
Another change is that due to the way fencing works with ra_buf (we get
one fence per ra_buf per upload) we have to use multiple ra_bufs instead
of offsets into a shared buffer. But for OpenGL this is probably better
anyway. It's possible that in future, we could support having
independent “buffer slices” (each with their own fence/sync object), but
this would be an optimization more than anything. I also think that we
could address the underlying problem (memory closeness) differently by
making the ra_vk memory allocator smart enough to chunk together
allocations under the hood.
Instead of merging it into render_dst. This is better for vulkan,
because blitting in vulkan both does not require a FBO *and* requires a
different image layout.
Also less "hacky" for OpenGL, since now the weird blit=FBO requirement
is an implementation detail of ra_gl
This extracts non-ANGLE specific code to d3d11_helpers.c, which is
modeled after egl_helpers.c. Currently the only consumer is
context_angle.c, but in future this may allow the D3D11 device and
swapchain creation logic to be reused in other backends.
Also includes small improvements to D3D11 device creation. It is now
possible to create feature level 11_1 devices (though ANGLE does not
support these,) and BGRA swapchains, which might be slightly more
efficient than ARGB, since its the same format used by the compositor.
This is for legacy GL: if VAOs are not available, the helper has to
specify vertex attributes again on every rendering. gl_vao_init() keeps
the vertex array for this purpose. Unfortunately, a temporary argument
was passed to the function, instead of the permanent copy.
Also, it didn't use num_entries (instead expected the array being
terminated by a {0} entry). Fix that source code indentation too.
In commit c6fafbffac we accidentally set the logical texture size to
the cropped video size, which is not correct. This caused rendering
artifacts in some cases.
Use the video surfaces size instead. Since the current mp_image_params
contains the cropped size only, wrapper texture creation has to be moved
to the _map function. Move the same code for the mixer case (strictly
speaking this is not needed, but seems more symmetric).
(Also there is no need to clear gl_textures on uninit - leftover from
the old hwdec mapper API. So we just drop that part.)
Fixes#4760.
Retrieve the depth for each component and internal texture format
separately. Only for 8 bit per component textures we assume that all
bits are used (or else we would in my opinion create too many probe
textures).
Assuming 8 bit components are always correct also fixes operation in
GLES3, where we assumed that each component had -1 bits depth, and this
all UNORM formats were considered unusable. On GLES, the function to
check the real bit depth is not available. Since GLES has no 16 bit
UNORM textures at all, except with the MPGL_CAP_EXT16 extension, just
drop the special condition for it. (Of course GLES still manages to
introduce a funny special case by allowing GL_LUMINANCE , but not
defining GL_TEXTURE_LUMINANCE_SIZE.)
Should fix#4749.
Runtime untested, because I get this:
[vo/rpi] Could not get DISPMANX objects.
This happened even when building older git versions, and on a RPI image
that hasn't changed in the recent years. I don't know how to make this
POS work again, so I guess if there's a bug in the new code, it will
remain broken.
This does two separate rather intrusive things:
1. Make the hwdec context (which does initialization, provides the
device to the decoder, and other basic state) and frame mapping
(getting textures from a mp_image) separate. This is more
flexible, and you could map multiple images at once. It will
help removing some hwdec special-casing from video.c.
2. Switch all hwdec API use to ra. Of course all code is still
GL specific, but in theory it would be possible to support other
backends. The most important change is that the hwdec interop
returns ra objects, instead of anything GL specific. This removes
the last dependency on GL-specific header files from video.c.
I'm mixing these separate changes because both requires essentially
rewriting all the glue code, so better do them at once. For the same
reason, this change isn't done incrementally.
hwdec_ios.m is untested, since I can't test it. Apart from superficial
mistakes, this also requires dealing with Apple's texture format
fuckups: they force you to use GL_LUMINANCE[_ALPHA] instead of GL_RED
and GL_RG. We also need to report the correct format via ra_tex to
the renderer, which is done by find_la_variant(). It's unknown whether
this works correctly.
hwdec_rpi.c as well as vo_rpi.c are still broken. (I need to pull my
RPI out of a dusty pile of devices and cables, so, later.)
Less flexible than GL_TIMESTAMP but supported by more platforms. This
will mean that nested queries have to be detected and silently omitted,
but oh well. Not much use for them anyway.
Fixes#4721.
Generic description of pixel formats is hard. In this case, the Apple
special format for packed YUV could have been interpreted as a RGB
format with funny packing.
Move multiple GL-specific things from the renderer to other places like
vo_opengl.c, vo_opengl_cb.c, and ra_gl.c.
The vp_w/vp_h parameters to gl_video_resize() make no sense anymore, and
are implicitly part of struct fbodst.
Checking the main framebuffer depth is moved to vo_opengl.c. For
vo_opengl_cb.c it always assumes 8. The API user now has to override
this manually. The previous heuristic didn't make much sense anyway.
The only remaining dependency on GL is the hwdec stuff, which is harder
to change.
Completely unnecessary, we can just update the uniforms immediately
after creating the program. In theory, for GLSL 4.20+, we could even
skip this, but oh well.
The vp_w/vp_h variables and parameters were not really used anymore
(they were redundant with ra_tex w/h) - but vp_h was still used to
identify whether rendering should be done mirrored.
Simplify this by adding a fbodst struct (some bad naming), which
contains the render target texture, and some parameters how it should be
rendered to (for now only flipping). It would not be appropriate to make
this a member of ra_tex, so it's a separate struct.
Introduces a weird regression for the first frame rendered after
interpolation is toggled at runtime, but seems to work otherwise. This
is possibly due to the change that blit() now mirrors, instead of just
copying. (This is also why ra_fns.blit is changed.)
Fixes#4719.
When using dumb mode, we can actually redraw a frame without uploading
it. Marking this as fresh as well results in unpredictable pass
behavior, which is confusing and makes debugging harder. So mark it as a
redraw instead, in that case.
Since the GL *gl is no longer needed for the timers, we can get rid of
the sc->gl dependency. This requires moving a utility function (which is
not GL-specific anyway) out of gl_utils.h and into utils.h
In the past, this always measured the per-shader execution times of the
individual OSD parts, which was thrown off because the shader was reused
anyway. (And apparently recording the OSD shader execution times was
removed completely, probably because of them being so unrealiably
anyway)
Since ra_timer no longer has the restriction of not allowing timers to
run concurrently, we can just wrap the entire OSD block inside a single
osd_timer now, and record that. (Technically, this can still be off when
using --blend-subtitles=video/yes and showing a full-screen OSD at the
same time. Maybe this can be done better?)
In order to prevent code duplication and keep the ra abstraction as
small as possible, `ra` only implements the actual timer queries,
it does not do pooling/averaging of the results. This is instead moved
to a ra-neutral struct timer_pool in utils.c.
This code is pretty much for the sake of vo_opengl_cb API users. It
resets certain state that either the user or our code doesn't reset
correctly. This is somewhat outdated. With GL implicit state being
so awfully large, it seems more reasonable require that any code
restores the default state when returning to the caller. Some
exceptions are defined in opengl_cb.h.
Now all GL-specifics of shader compilation are abstracted through ra.
Of course we still have everything hardcoded to GLSL - that isn't going
to change.
Some things will probably change later - in particular, the way we pass
uniforms and textures to the shader. Currently, there is a confusing
mismatch between "primitive" uniforms like floats, and others like
textures.
Also, SSBOs are not abstracted yet.
Instead of having a mutable ra_tex field (and the only one), move the
flag to struct ra, since we have only 2 tex_upload user calls anyway,
and both want the same PBO behavior. (At first I considered making it
a RA_TEX_UPLOAD_ flag, but why bother. PBOs are a terribly GL-specific
thing, so we can't expect a reasonable abstraction of it anyway.)
This requires a silly extension to ra_fns.tex_upload: since the OSD
texture can be much larger than the actual OSD image data to upload, a
mechanism for uploading only to a small part of the texture is needed.
Otherwise, we'd have to realloc/copy the data, just to pad it, and then
pay for uploading the padding too.
The RA_TEX_UPLOAD_DISCARD flag is not interpreted by GL (not sure how
you'd tell GL about this), but it clarifies the API and might be
helpful if we support other backend APIs in the future.
Actually GL-specific parts go into gl_utils.c/h, the shader cache
(gl_sc*) into shader_cache.c/h.
No semantic changes of any kind, except that the VAO helper is made
public again as part of gl_utils.c (all while the goal for gl_utils.c
itself is to be included by GL-specific code).
Another "small" step towards removing GL dependencies from the renderer.
This commit generally passes ra_tex objects instead of GL FBO integer
IDs to various rendering functions. video.c still manually binds the
FBOs when calling shaders.
This also happens to fix a memory leak with output_fbo.
Further work removing GL dependencies from the actual video renderer,
and moving them into ra backends.
Use of glInvalidateFramebuffer() falls away. I'd like to keep this, but
it's better to readd it once shader runs are in ra.
This currently only works when using lcms-based color management
(--icc-profile-*).
In principle, we could also support using lcms even when the user has
not specified an ICC profile, by generating the profile against a fixed
reference (--target-prim/--target-trc) instead. I still might do that
some day, simply because 3dlut provides a higher quality conversion than
our simple gamut mapping does for stuff like BT.2020, and also because
it's now needed to enable embedded ICC profiles. But that would be a
separate change, so preserve the status quo for now.
(Besides, my opinion is still that you should be using an ICC profile if
you care about colors being accurate _at all_)
mesa won't pick client storage unless this bit is set, and we
*absolutely* want to be using client storage for our DR PBOs.
Performance is shit on AMD otherwise. (Nvidia always uses client storage
for persistent coherent buffers whether you tell it it or not, probably
because it's way faster and nvidia doesn't trust users to figure that
out on their own)
It makes no sense to have this on an already created buffer.
If anything, the ra backend would have to export this as a global value
(e.g. struct ra field), so that whatever allocates the buffer can
account for the required alignment. Since this code is in vo_opengl.c in
the first place, and since GL doesn't dictate any special alignment
here, it doesn't make sense in the first place to export this. (Maybe
something like this will be required later.)
Breaks on mesa for whatever reason... even though it doesn't generate a
GLSL shader compiler error
Shouldn't make a performance difference for us because we cache `pos`
anyway, and most compute shaders will probably cache all of their
samples to shmem. Might have to re-visit this when we have an actual use
case for repeated sampling inside CS though. (RAVU + anti-ringing is a
possible candidate for that)
Or less appropriate, as some would argue. The new name is short for
"Apple YUV packed".
(This format is needed only for hardware decoding on rather old Apple
hardware, and a very annoying special case.)
This broke float textures, which were actually used by some shaders.
There were probably some other bugs as well.
Lots of code can be avoided by using ra_tex_params directly, so do that.
The main change is that COMPONENT/FORMAT are replaced by a single FORMAT
directive, which takes different parameters now. Due to the mess with
16/32 bit float textures, and because we want to support other APIs than
just GL in the future, it's not really clear how this should be handled,
and the nice component/type separation makes things actually harder. So
just jump the gun and use the ra_format.name names, which were
originally meant mostly for debugging. (This is probably something that
will be regretted later.)
Still only superficially tested, but seems to work.
Fixes#4708.
Since this code was already written for HDR, and is now per-channel
(because it works better for HDR as well), we can actually reuse this to
get very high quality gamut mapping without clipping. The only required
change is to move the tone mapping from before the gamut map to after
the gamut map. Additonally, we need to also account for changes in the
signal range as a result of applying the CMS when we compute ref_peak,
which is fortunately pretty easy because we only need to consider the
case of primaries mapping to themselves.
Since `HDR` no longer really makes sense as a label, rename it to
`--tone-mapping` in general. Also fits better with
`--tone-mapping-desat` etc.
Arguably we could also rename `--hdr-compute-peak`, but that option is
basically only useful for HDR content anyway because we don't need
information about the signal range for gamut mapping.
This (finally!) gives us reasonably high quality gamut mapping even in
the absence of an ICC profile / 3DLUT.
Huge thanks to @rusxg for finding this solution, which was previously
believed not to exist. Of course, we still don't actually need it, but I
don't want to leave this half-implemented in case somebody does in the
future.
So far, switching between integrated and discrete GPU would cause the
kernel to kill mpv due to an indecipherable buffer error. The technical
note TN2229 from Apple recommends to enable OpenGL Offline Renderers for
every Mac with more GPUs than displays to handle the switch between GPU.
By ordering the array from the least commonly rejected to the most,
we can sequentially remove PixelFormat attributes to fit the host.
Fixes#2371
Also add some more helpers.
Fix the broken math.h include statement.
utils.c uses ra_gl.h internals, which it shouldn't, and which will be
removed again as soon as this code gets converted to ra fully.
The dither texture data is created as a float array, but uploaded to a
texture with GL_R16 as internal format. We relied on GL to do the
conversion from float to uint16_t. Not all GL variants even support
this: GLES does not provide this conversion (one of the reasons why this
code has a float16 code path). Also, ra is not going to do this. So just
convert on the fly.
Still keep the float16 texture format fallback, because not all GLES
implementations provide GL_R16.
There is some possibility that we'll need to provide some kind of upload
conversion anyway for float->float16. We still rely on GL doing this
implicitly, and all GL variants support it, but with RA there might be
the need for explicit conversion. Even then, it might be best to reduce
the number of conversion cases. I'll worry about this later.
Format handling via ra_* was added earlier, but the format negotiation
part was forgotten.
Actually move some aspects of it to ra_get_imgfmt_desc(). Also make sure
the unorm and float formats selected by the common format lookup
functions are linear filterable. (For OpenGL, this is implicitly
guaranteed, so it wasn't done before.) Whether these assumptions should
be checked/enforced in the ra code at all is a bit fuzzy, but with ra
being helper code only for the actual video renderer, it's probably
justified.
Parsing the texture data as raw strings makes the textures the most
portable and self-contained. In order to facilitate different types of
shaders, the parse_user_shader interaction has been changed to instead
have it loop through blocks and call the passed functions for each valid
block parsed. This is more modular and also cleaner, with better code
separation.
Closes#4586.
- Each struct tex_hook now stores multiple hooks, this allows us to
avoid the awkward way of the current code has to add the same pass
multiple times.
- As a consequence, SHADER_MAX_HOOKS was split up into SHADER_MAX_PASSES
(number of tex_hooks) and SHADER_MAX_HOOKS (number of hooked textures
per tex_hook), and both numbers decreased correspondingly.
- Instead of having a weird free() callback, we can just leverage
talloc's recursive free behavior. The only user is the user shaders code
anyway.
This actually makes sure we don't decolor due to clipping even when the
signal itself exceeds the luma by a significant factor, which was pretty
common for saturated blues (and to a lesser degree, reds) - most
noticeable in skies etc.
This prevents the turn-the-sky-cyan effect of mobius tone mapping, and
should also improve the other tone mapping modes in quality.
This starts work on moving OpenGL-specific code out of the general
renderer code, so that we can support other other GPU APIs. This is in
a very early stage and it's only a proof of concept. It's unknown
whether this will succeed or result in other backends.
For now, the GL rendering API ("ra") and its only provider (ra_gl) does
texture creation/upload/destruction only. And it's used for the main
video texture only. All other code is still hardcoded to GL.
There is some duplication with ra_format and gl_format handling. In the
end, only the ra variants will be needed (plus the gl_format table of
course). For now, this is simpler, because for some reason lots of hwdec
code still requires the GL variants, and would have to be updated to
use the ra ones.
Currently, the video.c code accesses private ra_gl fields. In the end,
it should not do that of course, and it would not include ra_gl.h.
Probably adds bugs, but you can keep them.
The radius check was not strict enough, especially not for all
platforms. To fix this, actually check the hardware capabilities instead
of relying on a hard-coded maximum radius.
The textures not having an FBO actually caused regressions when trying
to render the subtitles on top of this texture (--blend-subtitles),
which still relied on an FBO.
So just kill off the logic entirely. Why worry about a single FBO wasted
when we're allocating like 10 anyway.
Fixes#4657.
According to the OpenGL spec, atomic access to SSBO variables is *not*
guaranteed to be coherent, even when reusing the same SSBO attached to
the same shader across different frames. So we actually need a
glMemoryBarrier here, at least in theory.
This bug slipped past my attention because nvidia ignores memory
barriers, but this is not necessarily always the case. Since
image_load_store is incoherent (specifically, writing to images from
compute shaders is incoherent) we need to insert a memory barrier to
make it coherent again. Since we only care about texture fetches, that's
the only barrier we need.
Two changes, compounded into one since they affect the same logic:
1. Never use linearization for HDR downscaling
2. Always use linearization for interpolation
Instead of fixing p->use_linear at the beginning of pass_render_frame,
we flip it on "dynamically" as needed. I plan on killing this
p->use_linear frame (along with other per-pass metadata) and moving them
into their own struct for tracking the "current" state of the video, but
that's a separate/upcoming refactor.
As a small bonus, reduce some code duplication in the interpolation
logic.
Fixes#4631
Mesa 17.1 supports compute shader but not full specs of OpenGL 4.3.
Change the code to detect OpenGL extension "GL_ARB_compute_shader"
rather than OpenGL version 4.3.
HDR peak detection requires SSBO, and polar scaler requires 2D array
extension. Add these extensions as requirement as well.
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
Performance seems pretty much unchanged but I no longer get nasty spikes
on NUMA systems, probably because glBufferSubData runs in the driver or
something.
As a simplification of the code, we also just size the PBO to always
have the full size, even for cropped textures. This seems slower but not
by relevant amounts, and only affects e.g. --vf=crop. It also slightly
increases VRAM usage for textures with big strides.
This new code path is especially nice because it no longer depends on
GL_ARB_map_buffer_range, and no longer uses any functions that can
possibly fail, thus simplifying control flow and seemingly deprecating
the manpage's claim about possible image corruption.
In theory we could also reduce NUM_PBO_BUFFERS since it doesn't seem
like we're streaming uploads anyway, but leave it in there just in
case some drivers disagree...
STREAM is better than DYNAMIC because we're only using it once per
frame. As for COPY vs DRAW, that was pretty much incorrect to begin with
- but surprisngly, COPY is actually faster (sometimes significantly so,
e.g. on my NUMA system).
After testing, the best I can gather is that it has to do with the fact
that COPY requires fewer redundant memcpy()s, and also 3x reduce RAM
bandwidth (in theory).
Anyway, that bit shouldn't introduce any regressions, it's just a
documentation update. Maybe I'll change my mind about the comment again
the future, it's really hard to tell. Vulkan, please save us!
Instead of allocating three PBOs and cycling through them, we allocate
one PBO that's three times as large, and cycle through the subregion
offsets.
This results in arguably simpler code and faster initialization
performance. Especially for 4K textures, initializing PBOs can take
quite some time (e.g. 180ms -> 110ms). For 1080p, it's more like 66ms ->
52ms for me.
The alignment to 4096 is completely unnecessary by spec, but we do it
anyway just for peace of mind.
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.
With some newer ANGLE builds, mapping can fail with "Failed to create
EGL surface" during playback. The reason is unknown, and it might just
be an ANGLE bug. Probe whether it works at init time to avoid the
problem.
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.
This just indicates a fixed linear coefficient to multiply into the
signal, similar to the old option --target-brightness (but the inverse
thereof). Good for testing purposes, which is why I added it. (This also
corresponds somewhat to what zimg does)
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 backend is selected if vaapi is available, but vaapi-over-EGL is
not. This causes various issues around the forced RGB conversion, which
is done with fixed, usually incorrect parameters.
It seems the existing auto probing check is too weak, and doesn't really
prevent it from getting loaded. Fix this by adding a flag to not ever
load this during auto probing.
I'm still not deleting it, because it's useful for testing on nvidia
machines.
See #4555.
The current algorithm blew up when the color was negative, such as the
case when downscaling with dscale=mitchell or other algorithms that
introduce negative ringing. The simplest solution is to just slightly
change the calculation to force both parameters to be in-range.
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.