[FFmpeg-devel] [RFC] swscale regression tests

Vitor Sessak vitor1001
Wed Jul 21 10:55:36 CEST 2010


On 07/21/2010 01:24 AM, Ramiro Polla wrote:
> 2010/7/20 M?ns Rullg?rd<mans at mansr.com>:
>> Ramiro Polla<ramiro.polla at gmail.com>  writes:
>>> 2010/7/20 M?ns Rullg?rd<mans at mansr.com>:
>>>> Ramiro Polla<ramiro.polla at gmail.com>  writes:
>>>>> I have attached an attempt at making swscale regression tests. This
>>>>> method creates one file for each input pixel format (currently 42) and
>>>>> compresses the output with bzip2. I have not tested with different
>>>>> platforms (x86_64 linux only yet), and IIRC it doesn't give the same
>>>>> result on all platforms. This is just a request for comments whether
>>>>> this approach is good. I have also removed the fast bilinear test
>>>>> since it works differently on many platforms and would make the
>>>>> regression tests more complex (like keeping one reference file for
>>>>> each platform).
>>>>>
>>>>> The output bzip2ed files add 280Kb of binary data to repository (5.6Mb
>>>>> uncompressed). As for disk usage (with 'du'), it takes 340Kb (5.8Mb
>>>>> uncompressed).
>>>>
>>>> Why can't you store a checksum instead?
>>>
>>> Storing a checksum makes it harder to find out which test failed. And
>>> the compressed files don't seem to take as much space as I had
>>> originally thought they would take.
>>
>> Then why are you bundling many tests together?
>
> There are 42 input/output pixel formats, 5 algorithms being tested,
> and all combinations of (64,96,128)->(64,96,128) conversions. That's
> 79380 tests. Ideas to rework the tests are welcome.

Why not group it in:

YUV -> YUV
YUV -> RGB
YUV -> EXOTIC
RGB -> RGB
RGB -> YUV
RGB -> EXOTIC
EXOTIC -> YUV
EXOTIC -> RGB
EXOTIC -> EXOTIC

And I would do MD5 for all them with the 5 algos and the bitexact flag 
(and it should match across all arches).

For testing without the bitexact flag, one might test if the sse does 
not increase more than 5% in comparison with the bitexact test.

-Vitor



More information about the ffmpeg-devel mailing list