[MPlayer-users] DVB stream including LATM AAC Audio issues

[Steven Ellis]

> Il Monday 21 January 2008 09:17:20 Steven Ellis ha scritto:
>
>> Nico - Awesome work. Thank you very much. Almost all of my samples
>> work perfectly now. I'll grab a bunch of fresh tests tonight from
>> the Auckland DTT transmissions as all of the channels have now
>> switched to AAC.
>>
>> Couple of bits of feedback.
>>
>> 1. I needed the following additional patch for my compile to work
>>
>> Index: libmpcodecs/ad_faad.c
>> ===================================================================
>> --- libmpcodecs/ad_faad.c       (revision 25826)
>> +++ libmpcodecs/ad_faad.c       (working copy)
>> @@ -26,6 +26,8 @@
>>  #ifndef USE_FAAD_INTERNAL
>>  #include <faad.h>
>>  #else
>> +#include "libfaad2/common.h"
>> +#include "libfaad2/bits.h"
>>  #include "libfaad2/faad.h"
>>  #endif
>
>
> ad_faad.c should remain unchanged respect to the version in svn.
> I'll check it.

If I remove my patch I get the following compile error

In file included from ad_faad.c:29:
../libfaad2/faad.h:29: warning: ignoring #pragma message
In file included from ../libfaad2/faad.h:32,
                 from ad_faad.c:29:
../libfaad2/neaacdec.h:221: error: expected declaration specifiers or
'...' before 'bitfile'
make[1]: *** [ad_faad.o] Error 1



>>
>> 2. I had to use the following configure option
>>
>> ./configure --with-extraincdir=../libfaad2
>>
>>
>> There appeared to be an issue with some of the newer definitions in
>> the faad2 headers.
>>
>>
>> I do have one sample that produces a couple of interesting debug
>> message, but it might be an issue with the broadcaster which is why
>> I want to grab some fresh samples
>>
>> Opening audio decoder: [faad] AAC (MPEG2/4 Advanced Audio Coding)
>>
>> Unsupported LATM configuration: 1 programs/ 42 subframes, 2 layers,
>> allstreams: 0
>>
>
> most likely there's some data that the parsing code identifies as
> a correct startcode (maybe I already wrote that such a short
> syncword is far too short and error prone) .
> Does decoding go on correctly after that point?
> Ah, btw, please test that decoding of ordinary AAC (non-latm)
> still works correctly: with that code base you  can never be sure
> if you broke something or not :)

All of the channels have now switched to LATM encapsulated AAC and I
hasn't kept any of my earlier AAC samples to test with.

Running through the NZ DTT test channels shows minimal issues except for
one channel where it looks like the whole channel is in a mess and the
encoder needs resetting.

So I'm assuming now that you need to push your changes upstream to the
faad2 developers for inclusion?

Steve