Use DXGI/D3DFMT 16bit pixel format to compatible with 12bit Y212/XV36
since there is no 12bit pixel defined in D3D11/D3D9.
Fix cmdline on Windows:
$ ffmpeg.exe -hwaccel qsv -init_hw_device qsv:hw,child_device_type=
{d3d11, dxva2} -i input_12bit.bin -f null -
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Fei Wang <fei.w.wang@intel.com>
Also include the hardware feature flags like the other archs do and
clean up the code a bit.
Tested on Linux POWER9.
Signed-off-by: Brad Smith <brad@comstyle.com>
Fixes compile error on windows with decklink:
In file included from ./libavformat/os_support.h:175,
from ./libavformat/internal.h:30,
from libavdevice/decklink_common.cpp:25:
./libavutil/wchar_filename.h: In function 'int wchartocp(unsigned int, const wchar_t*, char**)':
./libavutil/wchar_filename.h:59:32: error: invalid conversion from 'void*' to 'char*' [-fpermissive]
59 | *filename = av_malloc_array(num_chars, sizeof **filename);
| ~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| void*
Regression since e9e8bea2e7.
Fixes ticket #11103.
Signed-off-by: Marton Balint <cus@passwd.hu>
This can be useful to simplify certain processes that need to know how many
reserved bits there are and where they are placed, even if they are ultimately
unused, as will be shown in the next commit.
For any other case where the user simply looks at nb_components components, it
will make no difference.
Signed-off-by: James Almer <jamrial@gmail.com>
This function is only used for filtering and generic compute.
The issue is that a view inherits the usage flags from the image
by default, and the spec says the view format must be compatible
with the usage. VkImageViewUsageCreateInfo allows us to filter out
the indeded uses of the imageview.
Pffff.
When using **integer** images inside shaders, it turns out
that conversion doesn't automatically happen, but we need to
explicitly use the imageviews to get the image exposed as
a suitable representation for the shader.
Finally enables bitexact image representations.
This format is useful for doing certain lossless transforms on images,
RCT in particular, which require you to escalate the size from 16 to
32 bits to avoid overflows.
APIchanges will be done alongside when comitting.
The original either reported 8 or 16-bit conversion from the
original, rather than being able to return the actual original.
This makes it usable in a situation where preserving exactness
is required.
The flexible array member struct can have padding added by
the compiler which was not taken into account properly, which
could lead to a heap buffer overflow.
Fei Wang
Sean McGovern
Brad Smith
James Almer
Marton Balint
Lynne
Anton Khirnov
Martin Schitter
Marvin Scholz