[FFmpeg-devel] [PATCH] QCLP wrong sampling rate

Baptiste Coudurier baptiste.coudurier
Wed Sep 24 19:44:28 CEST 2008


Michael Niedermayer wrote:
> On Wed, Sep 24, 2008 at 10:10:02AM -0700, Baptiste Coudurier wrote:
>> Michael Niedermayer wrote:
>>> On Sat, Aug 16, 2008 at 09:58:08PM +0900, KO Myung-Hun wrote:
>>>> Baptiste Coudurier wrote:
>>>>> Hi,
>>>>>
>>>>> KO Myung-Hun wrote:
>>>>>   
>>>>>> Reynaldo H. Verdejo Pinochet wrote:
>>>>>>     
>>>>>>> Hi KO Myung-Hun
>>>>>>>
>>>>>>> KO Myung-Hun wrote:
>>>>>>>  
>>>>>>>       
>>>>>>>> Hi/2.
>>>>>>>>
>>>>>>>> Sampling rate of QCLP audio is always set to 8000Hz. But it's wrong.
>>>>>>>>
>>>>>>>> Test clips : http://sketchup.google.com/gsu6vtvideos.html
>>>>>>>>     
>>>>>>>>         
>>>>>>> I had problems with these kind of streams in the past
>>>>>>>
>>>>>>> If this QCLP tag identifies TIA/EIA/ISS-733 conforming
>>>>>>> data then this is just wrong. See, these are sampled at
>>>>>>> 8kHz -- period. this is no optional recomendation but
>>>>>>> a requirement (You can take a look at section 2.4.1
>>>>>>> of the all infamous C.S0020-0with3Gcover.pdf if you
>>>>>>> feel like it) Now, if this patch works out decoding
>>>>>>> for you I'd suggest adding a warning to state out-spec
>>>>>>> data, at the very least.
>>>>>>>   
>>>>>>>       
>>>>>> I've added a warning to the patch as your suggestion.
>>>>>>     
>>>>> Does this file decodes correctly with ffmpeg ?
>>>>> With soc decoder or non redistributable one ?
>>>>>
>>>>>   
>>>> Ah, MPlayer with QuickTime support.
>>> baptiste?
>>> IMHO the mov demuxer should not override the stored sampling rate for
>>> qcelp & amr
>> Well, I'm not disagreeing, the code was this way a long time ago, should
>> be fixed.
> 
> thanks
> 
> Though iam realizing now that the overriding may be needed for .mp4, though
> thats just a feeling based on the reliability of such fields for other
> codecs like aac in .mp4 ...
> 

Channels overriding is needed for mp4/3gp because it is commonly set to
2. Sample rate has to be set correctly in both formats, so I guess this
should be ok. About aac nightmare, it is fetched in esds, which is the
one the decoder uses, so I guess this should be ok too. If anything goes
wrong I'll fix it anyway.

-- 
Baptiste COUDURIER                              GnuPG Key Id: 0x5C1ABAAA
Smartjog USA Inc.                                http://www.smartjog.com
Key fingerprint                 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA




More information about the ffmpeg-devel mailing list