On 04/26/2013 02:36 PM, Diego Biurrun wrote:
> On Fri, Apr 26, 2013 at 01:05:30PM -0400, Justin Ruggles wrote:
>> On 04/26/2013 10:48 AM, Diego Biurrun wrote:
>>> The function requires increasing the fuzz factor for the ac3/eac3 encode
>>> tests and even so makes fate fail. It only provides a slight encoding
>>> speedup for legacy CPUs that do not support SS2. Thus its benefit is not
>>> worth the trouble it creates and fixing it would be a waste of time.
>>> ---
>>>  libavcodec/x86/ac3dsp.asm    |   36 ------------------------------------
>>>  libavcodec/x86/ac3dsp_init.c |    1 -
>>>  tests/fate/ac3.mak           |    2 --
>>>  3 files changed, 0 insertions(+), 39 deletions(-)
>> [...]
>>> diff --git a/tests/fate/ac3.mak b/tests/fate/ac3.mak
>>> index 246dd7c..23974ba 100644
>>> --- a/tests/fate/ac3.mak
>>> +++ b/tests/fate/ac3.mak
>>> @@ -52,14 +52,12 @@ fate-ac3-encode: CMD = enc_dec_pcm ac3 wav s16le $(REF) 
>>> -c:a ac3 -b:a 128k
>>>  fate-ac3-encode: CMP_SHIFT = -1024
>>>  fate-ac3-encode: CMP_TARGET = 404.53
>>>  fate-ac3-encode: SIZE_TOLERANCE = 488
>>> -fate-ac3-encode: FUZZ = 3
>>>  
>>>  FATE_EAC3-$(call ENCDEC, EAC3, EAC3) += fate-eac3-encode
>>>  fate-eac3-encode: CMD = enc_dec_pcm eac3 wav s16le $(REF) -c:a eac3 -b:a 
>>> 128k
>>>  fate-eac3-encode: CMP_SHIFT = -1024
>>>  fate-eac3-encode: CMP_TARGET = 516.94
>>>  fate-eac3-encode: SIZE_TOLERANCE = 488
>>> -fate-eac3-encode: FUZZ = 3
>>>  
>>>  fate-ac3-encode fate-eac3-encode: CMP = stddev
>>>  fate-ac3-encode fate-eac3-encode: REF = 
>>> $(SAMPLES)/audio-reference/luckynight_2ch_44kHz_s16.wav
>>
>> Are you sure these pass with FUZZ = 1 on all platforms? e.g. x86-32 with
>> no asm.
> 
> It passes on my K6-3 with no asm and I added the fuzz for that
> machine back in 4cc8e153715ed9757974e4976c5766d2e731bee0, so I
> expect no problems.

Patch is ok with me as long as the tests still pass on x86-32 and x86-64
with and without asm.

-Justin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to