Re: [Zynaddsubfx-user] a mystery

[Abrolag <ab...@us...>]

On Tue, 17 Aug 2010 21:53:58 +1000
cal <ca...@gr...> wrote:

> There's a little piece in ADnote::ComputeVoiceOscillatorRingModulation() that
> puzzles me -
> 
>      // if I use VoiceOut[] as modullator
>      for(int k = 0; k < unison_size[nvoice]; k++) {
>          REALTYPE *tw = tmpwave_unison[k];
>          for(i = 0; i < SOUND_BUFFER_SIZE; i++) {
>              amp = INTERPOLATE_AMPLITUDE(FMoldamplitude[nvoice],
>                                          FMnewamplitude[nvoice],
>                                          i,
>                                          SOUND_BUFFER_SIZE);
>              int FMVoice = NoteVoicePar[nvoice].FMVoice;
>              for(i = 0; i < SOUND_BUFFER_SIZE; i++)
>                  tw[i] *=
>                       (1.0 - amp) + amp * NoteVoicePar[FMVoice].VoiceOut[i];
>          }
>      }
> 
> As I read it, the second for(i = 0; i < SOUND_BUFFER_SIZE; i++) loop means
> that the first for(i = 0 ...) loop will never run beyond the i = 0 case. So
> far I can't figure out for sure just what the intended/correct behaviour might
> be. Any suggestions?
> 

Definitely looks wonky. Also goes back to Zyn 2.2.1.
I wonder if it's simply a wrong bracket position and the two loops were
intended to be separate.

This link gives a clue as to what Paul might have been thinking of at
the time he wrote it.

http://www.codeguru.com/cpp/cpp/cpp_mfc/portability/article.php/c4051

-- 
Will J Godfrey
http://www.musically.me.uk
Say you have a poem and I have a tune.
Exchange them and we can both have a poem, a tune, and a song.