Also test 8bit formats, and try bitdepth conversion paths.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
The fate test reference changes are due to the conversion being a simple
lossless deinterleave, instead of going through a RGB -> YUV -> RGB roundtrip.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
This commit also reduces the number of times ff_sws_init_scale() gets
called (only once per bit depth), and the number of times randomize_buffers()
gets called (only if the function must be checked).
Benchmarks are only performed on bit depths 8 and 16 (since they are
different functions, and not only different constants).
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Reduce input sizes to 8 (to test that the function works with widths
smaller than the vector length) and 1920 (raising the largest input
size to improve benchmark results).
Signed-off-by: Paul B Mahol <onemda@gmail.com>
-u and -q doesn't work together for diff on macOS.
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Much like XV30 and XV36 in d75c4693fe,
XV48 is added to support 16bit 4:4:4 as defined by Microsoft.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
And preserve the public SwsContext as separate name. The motivation here
is that I want to turn SwsContext into a public struct, while keeping the
internal implementation hidden. Additionally, I also want to be able to
use multiple internal implementations, e.g. for GPU devices.
This commit does not include any functional changes. For the most part, it is
a simple rename. The only complications arise from the public facing API
functions, which preserve their current type (and hence require an additional
unwrapping step internally), and the checkasm test framework, which directly
accesses SwsInternal.
For consistency, the affected functions that need to maintain a distionction
have generally been changed to refer to the SwsContext as *sws, and the
SwsInternal as *c.
In an upcoming commit, I will provide a backing definition for the public
SwsContext, and update `sws_internal()` to dereference the internal struct
instead of merely casting it.
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Niklas Haas <git@haasn.dev>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
This test utilizes an MPEG-TS sample from FATE suite, yet was
marked as not requiring samples.
Reviewed-by: Jan Ekström <jeebjp@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
This helps detecting inconsistencies in swscale input/output code and
a av_read_image_line()/av_write_image_line() pass.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
This makes the unscaled output of p010le and p016le match the generic path.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Ensure those bits are copied, which will result in the output being the same as
the input, where swscale set them to the equivalent of fully opaque.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
And add a few more tests to ensure all the pixfmts affected by this change
are tested.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Forgotten in 601a3479466f4f03d27487dfc6fcb9430efb8fc6.
Found-by: ramiro
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Mistakenly written in native endianness in 42098f92662c47659df85173d8076b62a23269aa.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
As with other pixel formats, set the undefined alpha bits to opaque.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
As with other pixel formats, set the undefined alpha bits to opaque.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Should fix fate-checkasm-sw_rgb under gcc-ubsan.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Ramiro Polla <ramiro.polla@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
The demuxer's 'missing_streams' private option is used to communicate
information from the demuxer to avformat_find_stream_info(). However,
that is not only unnecessarily complicated, it also leaks internal
information to users, e.g. this option appears in the results of the
fate-flv-demux test.
Use a new field in FFFormatContext to communicate this information
instead.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
stddev and PSNR values change by the removal of format conversion and because
of the added sws_flags. Either or will have the same effect even on their own.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul B Mahol <onemda@gmail.com>
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.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
