[FFmpeg-devel] [PATCH] als in mp4

[Jai Menon]

On Sat, Mar 21, 2009 at 5:16 PM, Justin Ruggles
<justin.ruggles at gmail.com> wrote:
> Jai Menon wrote:
>> On Sat, Mar 21, 2009 at 11:52 AM, Baptiste Coudurier
>> <baptiste.coudurier at gmail.com> wrote:
>>> Jai Menon wrote:
>>>> On Sat, Mar 21, 2009 at 11:24 AM, Baptiste Coudurier
>>>> <baptiste.coudurier at gmail.com> wrote:
>>>>> Hi Jai,
>>>>>
>>>>> Jai Menon wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Attached :
>>>>>>
>>>>>> 1) demux als streams from mp4, als codec id and minor lavc version bump
>>>>>> 2) mux als in to mp4, add isomedia type
>>>>>>
>>>>>> [...]
>>>>>>
>>>>>> @@ -1381,7 +1385,8 @@
>>>>>> ? ? ?uint8_t *buf;
>>>>>>
>>>>>> ? ? ?for (i = 0; i < s->nb_streams; i++)
>>>>>> - ? ? ? ?if (mov->tracks[i].enc->flags & CODEC_FLAG_BITEXACT) {
>>>>>> + ? ? ? ?if (mov->tracks[i].enc->flags & CODEC_FLAG_BITEXACT ||
>>>>>> + ? ? ? ? ? ?mov->tracks[i].enc->codec_id == CODEC_ID_MP4ALS) {
>>>>>> ? ? ? ? ? ? ?return 0;
>>>>>> ? ? ? ? ?}
>>>>>>
>>>>> ???? Please explain :>
>>>> ok :)
>>>> the muxer currently writes out a userdata box, with a meta box inside
>>>> (mp4 mode).
>>>> The meta box has a itunes hdlr tag which again leads to problems. So the patch
>>>> skips writing a udta box. maybe instead of this, we could write mov style udta.
>>>> what do you suggest?
>>>>
>>>> btw, very sorry most of this looks like a hack, the reference codec
>>>> mp4 parser has
>>>> limitations.
>>>>
>>> Humm, sorry too hackish, where can I get this reference codec mp4 parser ?
>>
>> http://www.nue.tu-berlin.de/forschung/projekte/lossless/mp4alsRM22.zip
>>
>> i dunno if there is a newer version which is publicly available.
>> maybe Justin knows?
>
> RM22 is the latest. ?Maybe you could try to fix whatever is weird about
> it and send a patch to Tilman (liebchen at nue.tu-berlin.de).

Sorry for the delay, no net access on weekends :)
The weirdness is basically that it does not write out the channel
configuration bits in the audio config object. Actually, according to
the spec, if the value is zero, then it should be read from the
decoder specific info object. So in this case we probably should parse
this value from the extradata. But, it would be easier if it was
written out in the audioconfig object. Also, the samplerate value is
correct so i don't understand why they can't do the same for channels
field.

> -Justin

-- 
Regards,

Jai