Re: [hatari-devel] Questions on Falcon DMA sound

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi Nicolas,

Was this checked after 2.3 release?

If not, would you have time to look into it after 2.5 release?


	- Eero

On 27.10.2020 11.17, Nicolas Pomarède wrote:
> Le 26/10/2020 à 23:08, Roger Burrows a écrit :
> > On 26 Oct 2020 at 19:39, Christian Zietz wrote:
> >
> > > Roger Burrows schrieb:
> > >
> > > > Quite possibly the documentation is just wrong.  However, in TOS4, 
> > > > there
> > > are
> > > > manipulations of 'mystery bits' that are not documented in the 
> > > > development
> > > > notes or the Compendium.  Specifically:
> > > >     . bit 3 of $ff8937 is toggled
> > > >     . bit 2 of $ff8938 is toggled
> > > >     . bit 3 of $ff8938 is cleared (only)
> > > > [...]
> > > > But meanwhile I would be interested to hear comments from the smart
> > > hardware
> > > > people about those mystery bits.
> > >
> > > There is this collection of technical documents that among other things
> > > covers the Falcon. Unfortunately, just the document that would contain
> > > these registers is damaged, so that only the text version (without
> > > formatting) survives, which is a bit hard to interpret.
> > >
> > > There, I can see:
> > ...
> > > Hence bit 3 of $FF8937 should be "Global Sound Reset". And...
> > Thanks!  That's a big help and expands on what I found by a test 
> > program (which
> > I have attached).  In the ZIP, Hatari developers please find the C 
> > source,
> > Lattice C .PRJ file, and binary for DMASOUND.TTP.
> >
> > Also there is FALCONHW.TXT which shows the results of running the 
> > program on
> > real Falcon hardware, and HATARI.TXT which shows the output from the same
> > program run on Hatari 2.2.1.  As you can see, toggling bit3 of $ff8937 
> > on real
> > hardware resets a lot of dmasound registers.
> >
> > I haven't tried running DMASOUND.TTP with the latest Hatari source in 
> > the repo,
> > but since the routines that handle that area of the hardware emulation 
> > do not
> > pay attention to that bit, I would be extremely surprised if the 
> > results were
> > different.
> >
> > I think that now would be a great time to fix this in Hatari, before the
> > upcoming 2.3, although I realise that there may be some concern about 
> > changing
> > the code at a late stage.  In any case, I will fix this on the EmuTOS 
> > side
> > shortly and see how that helps on real Falcon hardware.  There may be 
> > other
> > things I need to do, but this definitely needs doing.
>
> Hi
>
> thanks for your tests / programs, they can be used to improve hatari. 
> But as you say, release is near, I don't think that's a good idea to 
> change things at this point, it might cause regression, especially now 
> that laurent/eero went through a huge list of demos/games to update 
> their compatibility with next hatari 2.3 ; it would mean to recheck 
> eveything.
>
> Nicolas
>
> > ...
> > > Thus, bit 3 of $FF8938 would be "Expand" (whatever that is) and bit 2
> > > would be "mute". No idea why one would toggle "mute", though. Also, I
> > > don't know if everything is actually implemented in hardware or some
> > > bits were only planned for future revisions of the Falcon that never
> > > came to be.
> > I will see if I can determine what toggling bit 2 does and what 
> > setting bit 3
> > to 1 might do, but that's likely to be difficult.  The toggling of bit 
> > 2 is
> > done in conjunction with setting and resetting values in the attenuation
> > register, so it might be possible to figure it out logically.  If 
> > anyone would
> > like to see the relevant TOS4 disassembly, I can post it here (it's very
> > short).
> >
> > Thanks again,
> > Roger