Re: [hatari-devel] Sample frame buffer is incorrectly restarted

[ Thread Index | Date Index | More lists.tuxfamily.org/hatari-devel Archives ]


Hi,

On 3.2.2024 19.04, Miro Kropáček wrote:
On Fri, 2 Feb 2024 at 23:48, Eero Tamminen <oak@xxxxxxxxxxxxxx> wrote:

$ hatari --machine falcon --fpu 68882 ./play.tos

     - sample is played back

Sound quality is rather scratchy, even when played back from GEMDOS HD
i.e. zero disk overhead.  Using 32Mhz CPU instead of default 16Mhz did
not really help.

Interesting.  I have Hatari compiled on Oct 18, 2023, 4+64 MB RAM, 32 MHz
*but* TOS 4.04. Can you repeat your test on TOS? For me the sound is pretty
good, except the weird repeat part.

Neither TOS4 nor TT-RAM / 32-bit addressing changed it, sound is still scratchy compared to PC playback of the raw file.


	- Eero

PS.  Having bad flu / fever, so I've not been at computer for a while...

Is it possible that Hatari with just EmuTOS is lacking lacking things
that e.g. initialize Falcon sound matrix, that could happen on real
machine with lots of SW running at startup (e.g. audio ACC / CPX).

For the bad replay quality maybe but for the repeat part (which you seem to
hear as well if I understood correctly) this seems to be a Hatari issue (as
I'm setting registers directly in the interrupt handler).

Strangely, I could get the same bug in Aranym as well. But if I put it in a
real world use case (i.e. the replay buffer and the delay is much smaller),
it worked well. So to me it looks like some timing issue within Hatari
(Aranym).




Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/