Re: [hatari-devel] STe DMA sound

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 14/04/2018 à 18:56, Vincent Rivière a écrit :
> On 14/04/2018 at 18:23, Vincent Rivière wrote:
> > With current HG sources, I have also commented out 
> > DmaSnd_GenerateSamples() as a test.
> > Result: Good YM music, no DMA digidrums, and no 7 KHz issue
> > It seems to indicate that the 7 KHz artifact comes from the DMA, 
> > unless I'm mistaken...
>
> Other test:
> dmaSnd.c, line 553:
>              switch (microwire.mixing) {
>                  case 1:
>                      /* DMA and YM2149 mixing */
>                      MixBuffer[nBufIdx][0] = MixBuffer[nBufIdx][0] /*+ 
> dma.FrameLeft * -((256*3/4)/4)/4*/;
>                      break;
>
>
> Note that I have commented out the DMA part (I could have simply 
> commented out the whole line).
>
> Result in Lethal Xcess main menu:
> YM music, no digidrums... but annoying 7 KHz artifact present!
> So it seems to indicate that the noise comes from YM.
>
> Opposite conclusion of my previous test...
> I need to have a rest.
>
> NB: The only sane possibility would be that DmaSnd_GenerateSamples() 
> trashes the YM sound by some side effect.

This change will not use the DMA sample, but il will still apply the LMC 
filters. So, comparing with your previous test where 
DmaSnd_GenerateSamples() was not called at all, this would point to a 
problem in the LMC filters, not in the YM part.