[FFmpeg-devel] [PATCH] tests/fate/cbs: add tests for APV

[James Almer]

On 5/5/2025 5:59 PM, Mark Thompson wrote:
> On 05/05/2025 18:13, James Almer wrote:
>> Ensure bitexact passthrough using the apv_metadata bsf.
>>
>> Signed-off-by: James Almer <jamrial at gmail.com>
>> ---
>>   tests/fate/cbs.mak                    | 11 +++++++++++
>>   tests/ref/fate/cbs-apv-profile_422-10 |  1 +
>>   2 files changed, 12 insertions(+)
>>   create mode 100644 tests/ref/fate/cbs-apv-profile_422-10
>>
>> diff --git a/tests/fate/cbs.mak b/tests/fate/cbs.mak
>> index dfd9e7d93c..729a559327 100644
>> --- a/tests/fate/cbs.mak
>> +++ b/tests/fate/cbs.mak
>> @@ -25,6 +25,17 @@ FATE_CBS_$(1)_DISCARD += fate-cbs-$(1)-discard-$(2)
>>   fate-cbs-$(1)-discard-$(2): CMD = md5 -i $(TARGET_SAMPLES)/$(3) -c:v copy -y -bsf:v filter_units=discard=$(2) -f $(4)
>>   endef
>>   
>> +# APV read/write
>> +
>> +FATE_CBS_APV_SAMPLES =             \
>> +    profile_422-10.apv
>> +
>> +$(foreach N,$(FATE_CBS_APV_SAMPLES),$(eval $(call FATE_CBS_TEST,apv,$(basename $(N)),apv,apv/$(N),apv)))
>> +
>> +FATE_CBS_APV-$(call FATE_CBS_DEPS, APV, APV, APV, APV, APV) = $(FATE_CBS_apv)
>> +FATE_SAMPLES_AVCONV += $(FATE_CBS_APV-yes)
>> +fate-cbs-apv: $(FATE_CBS_APV-yes)
>> +
>>   # AV1 read/write
>>   
>>   FATE_CBS_AV1_CONFORMANCE_SAMPLES = \
>> diff --git a/tests/ref/fate/cbs-apv-profile_422-10 b/tests/ref/fate/cbs-apv-profile_422-10
>> new file mode 100644
>> index 0000000000..897520f5f1
>> --- /dev/null
>> +++ b/tests/ref/fate/cbs-apv-profile_422-10
>> @@ -0,0 +1 @@
>> +646c61e4e96b83146da8140c4c2a68b6
> 
> Add fate-cbs-apv as a dep of fate-cbs at the top of the file as well.
> 
> Test LGTM.  We should add some more once we have samples with better feature coverage (multiple frames, AU info, more metadata), but this is good for what we have now.

The latter two can be made with the av1_metadata bsf, but I'd leave AU 
info alone until we can get (or make) a sample with more than one frame 
per AU.

The spec doesn't really go into detail about the other frame types, but 
can we assume that Alpha is, like in other codecs, just a separate 
stream with a luma plane?

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 495 bytes
Desc: OpenPGP digital signature
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20250505/854dad59/attachment.sig>