[FFmpeg-soc] [soc]: r2856 - aac/aac.c

[Robert Swain]

2008/7/27 Michael Niedermayer <michaelni at gmx.at>:
> On Sun, Jul 27, 2008 at 11:53:32AM +0200, superdump wrote:
>> Author: superdump
>> Date: Sun Jul 27 11:53:32 2008
>> New Revision: 2856
>>
>> Log:
>> Add doxygen comments for band types
>>
>>
>> Modified:
>>    aac/aac.c
>>
>> Modified: aac/aac.c
>> ==============================================================================
>> --- aac/aac.c (original)
>> +++ aac/aac.c Sun Jul 27 11:53:32 2008
>> @@ -174,12 +174,12 @@ enum WindowSequence {
>>  };
>>
>>  enum BandType {
>> -    ZERO_BT        = 0,
>> +    ZERO_BT        = 0,     ///< Scalefactors and spectral data are all zero.
>
>>      FIRST_PAIR_BT  = 5,
>
> and this one?

I couldn't find any specific description from the spec only code
depending on it. And I couldn't find anything about the order of what
they call the huffman codebooks to be able to disseminate what the
'pair' is to be able to assign it some meaning.

>> -    ESC_BT         = 11,
>> -    NOISE_BT       = 13,
>> -    INTENSITY_BT2  = 14,
>> -    INTENSITY_BT   = 15,
>> +    ESC_BT         = 11,    ///< Spectral data are coded with an escape sequence.
>> +    NOISE_BT       = 13,    ///< Spectral data are scaled white noise not coded in the bitstream.
>
>> +    INTENSITY_BT2  = 14,    ///< Scalefactor data are intensity stereo positions.
>> +    INTENSITY_BT   = 15,    ///< Scalefactor data are intensity stereo positions.
>
> what is the difference between them?

if (sce1->band_type[g][i] == INTENSITY_BT || sce1->band_type[g][i] ==
INTENSITY_BT2) {
    c = -1 + 2 * (sce1->band_type[g][i] - 14);

It seems like this is the only occurrence that will cause any
difference and this appears to swap (BT2) or maintain (BT) the sign of
the intensity stereo position.

>>      ESC_FLAG       = 16,
>
> and this?

Unused, can be removed.

Rob