[hatari-devel] New ST(e) tracker bad audio

[Nicolas Pomarède]

Le 17/06/2011 11:01, Anders Eriksson a écrit :
> On Wed, 15 Jun 2011, Nicolas Pomarède wrote:
>
>> yes I think there're still some rounding errors or things like in ste
>> dma sound code. I'm trying to improve the precision for others bad
>> quality in some demos, so this will certainly benefit to this program
>> too.
>>
>> By the way, could you test on a real STE the program DMA_TEST that
>> Paulo posted on
>> http://atari-forum.com/viewtopic.php?f=28&t=20760&start=25 ?
>> The value slightly change on each run, but I'd like to know what you
>> get on a PAL STE.
>
>
> Hey hopp Nicolas!
>
> I've run the DMA_TEST program ten times and got the same value each
> time: 50016.
>
> Additionally here are recordings from a real STe with Hextracker for
> comparison:
>
> http://files.dhs.nu/party/outline2011/Dma_Sc-Hextracked-STe_Mono_12517Hz.mp3
>
> http://files.dhs.nu/party/outline2011/Dma_Sc-Hextracked-STe_Stereo_10013Hz.mp3
>
> http://files.dhs.nu/party/outline2011/Dma_Sc-Hextracked-ST_YM2149_Mono_10072Hz.mp3
>

Hello

thanks for the measure ; actually I get 50855 with my dev version of 
Hatari, so something is clearly wrong (as can be confirmed when 
comparing the sound quality of the mp3 with the result obtained in 
Hatari). Steem is much closer to a real STE with values around 50147.

I will try to code some very simple yet precise test programs to compare 
exactly a real STE to Hatari.

Hopefully this should help fixing things before releasing Hatari 1.5 
(I'd like to fix this before 1.5, releasing a new version 1 year after 
1.4 that wouldn't support the latest demos would certainly be 
disappointing to many users)

Nicolas

> As you can see it's the Dma-Sc zik he made for Outline music
> competition, 16 channels.
>
> http://files.dhs.nu/party/outline2011/ol2011stmod-dmasc-hextracked.zip
>
> --
> Anders Eriksson
> ae at dhs.nu http://www.dhs.nu/
> ae at atari.org http://www.atari.org/
>
>
>
> _______________________________________________
> hatari-devel mailing list
> hatari-devel at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/hatari-devel