Great to see you here Raphael
Gabriel Bouvigne - France
[EMAIL PROTECTED]
icq: 12138873
MP3' Tech: www.mp3tech.org
----- Message d'origine -----
De : Rafael W. Luebbert <[EMAIL PROTECTED]>
� : <[EMAIL PROTECTED]>
Envoy� : vendredi 28 mai 1999 21:29
Objet : [MP3 ENCODER] LAME problems.
> Hello, all, I just thought I would throw in a few things after looking
> at your 3.04 code. Is 3.05 out?
>
> You use numlines[] to calculate your perceptual entropy, yet
> you are setting it here:
>
> (from l3psy.c)
>
> /* Read long block data */
>
> for(loop=0;loop<6;loop++)
> {
> freq_tp = *p++;
> cbmax_tp = (int) *p++;
> cbmax_tp++;
>
> if (sfreq == freq_tp )
> {
> cbmax = cbmax_tp;
> for(i=0,k2=0;i<cbmax_tp;i++)
> {
> j = (int) *p++;
> numlines[i] = (int) *p++; <<<<<<<<<<<<<<<<<<<
> minval[i] = *p++;
> qthr_l[i] = *p++;
> norm_l[i] = *p++;
> bval_l[i] = *p++;
> if (j!=i)
> {
> printf("1. please check \"psy_data\"");
> exit(-1);
> }
> for(k=0;k<numlines[i];k++)
> partition_l[k2++] = i ;
> }
> }
> else
> p += cbmax_tp * 6;
> }
>
>
> and then again, overwriting half of it here with the short block
information:
>
>
> /* Read short block data */
>
> for(loop=0;loop<6;loop++)
> {
> freq_tp = *p++;
> cbmax_tp = (int) *p++;
> cbmax_tp++;
>
> if (sfreq == freq_tp )
> {
> cbmax = cbmax_tp;
> for(i=0,k2=0;i<cbmax_tp;i++)
> {
> j = (int) *p++;
> numlines[i] = (int) *p++; <<<<<<<<<<<<<<<<<<<<<
> qthr_s[i] = *p++;
> norm_s[i] = *p++;
> SNR[i] = *p++;
> bval_s[i] = *p++;
> if (j!=i)
> {
> printf("3. please check \"psy_data\"");
> exit(-1);
> }
> for(k=0;k<numlines[i];k++)
> partition_s[k2++] = i ;
> }
> }
> else
> p += cbmax_tp * 6;
> }
>
> thus your pe ratios are going to be screwed.
>
> Also, the IS says to use values of the spreading function greater than
10e-6,
> thus
> shouldn't your spreading function equation be:
>
> if ((x+tempy) <= -60.0) s3_s[i][j] = 0.0;
> else s3_s[i][j] = exp( (x + tempy)*LN_TO_LOG10 );
>
> instead of:
>
> if (tempy <= -60.0) s3_s[i][j] = 0.0;
> else s3_s[i][j] = exp( (x + tempy)*LN_TO_LOG10 );
>
> ?
>
>
> Ill look around a bit more and see if I see any other blaring
irregularities.
>
> Hope this helps,
> Rafael
>
>
> --
> MP3 ENCODER mailing list
>
--
MP3 ENCODER mailing list
