9191aa5607
FakeESOut: fix uninitialized member
priv_control must be NULL if created from a module.
6 years ago
e9e46687f3
direct3d9: add an option to use DXVA-HD rather than StretchRect by default
The driver may lie about its conversion capabilities (like Y410 to X8R8G8B8 on
Intel) when DXVA-HD works correctly. So we use this option by default.
7 years ago
3d275aded4
decoder: rename input_Decoder*() to vlc_input_decoder_()
In order to match with vlc_input_decoder_t.
6 years ago
412d1b8a4c
vlc_decoder: use a specific opaque struct for input API
That way, the input API can't be used with user decoder_t modules.
This is a prerequisite of the packetizer API rework.
6 years ago
fa8c8eac8b
demux: adaptive: allow to override max buffering
7 years ago
41f5882377
demux: adaptive: recurse getPlaybackTimeDurationBySegmentNumber
6 years ago
1d0e48f08e
demux: adaptive: fix getPlaybackTimeDurationBySegmentNumber startnumber offset
7 years ago
8587d83877
demux: adaptive: fix live template period offset
6 years ago
680109a375
demux: adaptive: remove unused playlist update parameters
6 years ago
adcc27453b
nvdec: use the CudaFunctions and CUcontext from the decoder device
No need to load the same resources twice. We hold a reference to the decoder
device with these objects.
6 years ago
87c40237d9
nvdec: get the decoder device earlier
If the decoder device doesn't match we shouldn't be using this decoder anyway.
6 years ago
c2a9006a03
Revert "android: util: remove redundant strncmp"
This reverts commit 77aa424b72
7 years ago
50522ede98
opengl: fix GL_ASSERT_NOERROR macro
The macro assumed a local variable vgl existed, which is not always the
case.
Instead, pass a pointer to the opengl_vtable_t as parameter.
Signed-off-by: Steve Lhomme <robux4@ycbcr.xyz>
7 years ago
30e67033ba
qt: don't create the "window" variable in libvlc
The variable already already exists. If it doesn't exist the
var_SetString("window", "qt,any") should fail as the UI won't work properly
anymore.
6 years ago
f707c6f002
opengl: dxva2: fix indentation after previous commits
No functional changes
6 years ago
e595beb4e9
opengl: dxva2: use DXVA-HD instead of StretchRect for NVIDIA GPUs
The StretchRect we're using has the same issue as with the D3D9 vout.
6 years ago
b7a56bc889
opengl: dxva2: make sure we can handle the conversion in StretchRect
6 years ago
15132bbffb
opengl: dxva2: check we have a IDirect3DDevice9Ex earlier
6 years ago
886587dc85
direct3d9: move the DXVA2 OpenGL Interop module in its own file
It shares nothing with the D3D9 module and will never be loaded at the same time.
No functional changes.
6 years ago
7bad2a867f
demux: wav: fix wrong pts after a seek
7 years ago
6cdbd26806
demux: wav: don't increment pts past data chunk bundary
7 years ago
7adb4c9b05
demux: wav: use vlc_stream_GetSize()
6 years ago
2309a6e285
demux: wav: more robust chunk parsing
From the spec:
"There are no restrictions upon the order of the chunks within a WAVE file,
with the exception that the Format chunk must precede the Data chunk."
The Wav demuxer can now parse chunks in any order. I don't check that the fmt
chunk is before the data one in order to be more resilient to broken samples.
This commit will allow to add new chunks more easily, specially chunks that can
be after the 'data' one. I'm thinking about the ADM support, that need 2
chunks, 'chna' and 'axml' that is after the 'data' chunk.
7 years ago
9d7e963e05
demux: wav: change i_size type
Use the same type than GetDWLE() that returns an uint32_t.
7 years ago
6544b0598c
demux: wav: change i_data_size sign
7 years ago
4b4d3ccdee
demux: wav: change data_pos sign
Use the same type than vlc_stream_Tell().
7 years ago
9de829ec61
demux: wav: parse ds64 in a new function
7 years ago
8d23f74274
demux: wav: move fmt parsing in a new function
7 years ago
1510c8b50f
demux: wav: skip header in ChunkFind
7 years ago
fd15e5a03f
demux: wav: refactor chunk skip
7 years ago
fcea115014
demux: wav: reorder to avoid forward declarations
No functional changes.
7 years ago
03208617c5
d3d11_surface: fix missing local d3d11_device_t
6 years ago
3644c45e16
libvlc: add a way to disable the GPU callback engine
If set to none we fallback to normal playback into a custom window.
We need the variable in the regular player so the variable can be read even
when desktop VLC.
7 years ago
595878cddd
aom: remove unsupported AOM chroma
Also removed upstream in f1570f0c2f70832dd170285f8de60bd2379c8efa. See
the following link:
https://aomedia.googlesource.com/aom/+/f1570f0c2f70832dd170285f8de60bd2379c8efa%5E%21/#F0
Signed-off-by: Thomas Guillem <thomas@gllm.fr>
7 years ago
42c45b9603
dxva2: add more preferred decoder pixel format based on chroma sub-sampling
If this format is available for decoding it should give the best result and
performance. If not, something else will be picked.
7 years ago
7fbe530406
direct3d9: add support for Y410 (4:4:4 10 bit) rendering
Via DXVA-HD conversion to RGB or possibly StretchRect if it supports it.
6 years ago
bf7e3ee83b
direct3d9: add support for Y210 (4:2:2 10 bit) rendering
Via DXVA-HD conversion to RGB or possibly StretchRect if it supports it.
6 years ago
9ec0d83a05
direct3d9: add support for AYUV (4:4:4 8 bit) rendering
Via DXVA-HD conversion to RGB or possibly StretchRect if it supports it.
6 years ago
9b7df42f0d
direct3d9: add support for YUY2 (4:2:2 8 bit) rendering
Via DXVA-HD conversion to RGB or possibly StretchRect if it supports it.
7 years ago
920b9c1322
direct3d9: code cleaning
6 years ago
3a2aa3deeb
dxgi_fmt: allow hardware decoding to Y410 (4:4:4 10 bit)
6 years ago
62c0ea9ab7
direct3d11: support rendering from Y410 (4:4:4 10 bit) to RGB
6 years ago
002850be53
dxgi_fmt: allow hardware decoding to Y210 (4:2:2 10 bit)
6 years ago
5b7d9e132f
direct3d11: support rendering from Y210 (4:2:2 10 bit) to RGB
6 years ago
ed9988f015
dxgi_fmt: allow hardware decoding to YUY2 (4:2:2 8 bit)
6 years ago
84666be679
dxgi_fmt: allow hardware decoding to AYUV (4:4:4 8 bit)
6 years ago
6b0522a3b0
avcodec: va: allow 444/422 decoding with DXVA2/D3D11VA
Not sure why we should restrict the input we accept if lavc proposes to decode
that format. We can always check and refuse later if the hardware can't do it.
7 years ago
b693265a38
directx_va: add support for custom Intel HEVC GUIDs
Tested with the patched FFmpeg
7 years ago
ef3191943b
dxva2: add more known pixel format
This doesn't mean we can handle them, just that we log them properly.
7 years ago
2c5144cbdb
dxva2: log the chroma sub-sampling of the source when selecting decoder format
6 years ago
Alexandre Janniaux
Steve Lhomme
Thomas Guillem
Francois Cartegnie
Romain Vimont