>
> Hi all.
>
> I'm an undergraduate computer engineer at UOW.
> For a project I have to manipulate the subbanded mdct
> transformed sample values when encoding in mp3 (layer3).
> I have been trying to understand the code dist10, but
> unfortunately without the standard, I can't seem to
> make much sense of it. I would appreciate _ANY_ help.
> Unfortuantely I seem to be going nowhere on my limited
> resources.
>
> For the Encoder, I am getting the values of xr[2][2][576]
> after the function mdct_sub() in musicin.c,
> where the first array index is the channel and the second and
> third represent the 1152 audio samples.
>
Actually, the first index is the 'granule' and the second index
is the channel. In mpeg1, there are two granules per frame
(1152 sample frames). In mpeg2, there is only one granule per frame
and the framesize is 576 sampels.
The 576 samples in xr represent 18 MDCT coefficients for each of the
32 subbands. Hence the dimensions of hybridIn[32][18].
If you didn't do any quantization or compression, I think hybridIn for the
decoder would be the same as xr[][] computed by the encoder,
except for maybe the ordering.
>
> --> If I have a mono audio signal, do I only worry about xr[0]??
>
yes, but it would be xr[gr][0].
Mark
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
