Re: [hatari-devel] Regression in Falcon sound

[Laurent Sallafranque <laurent.sallafranque@xxxxxxx>]

Hi all,

> - Old bug: When starting a second time, there's old sample data played for a frame or two, VGA+RGB (also present in 2.4.1)

I've also noticed the old sample played (and I confirm  it's not recent).

The falcon have two  32 bytes buffers (FIFO if i'm right) for sound buffering.

It's not emulated in hatari.

When this fifo is empty, the "sound value" should be 0 and this would clear the buffer after sound is played.

I'll have a look at the problem too (but I don't have a lot of time actually, my mother in law is at hospital).

Regards

Laurent

Le 17/04/2024 à 13:57, Daniel Hedberg a écrit :

On Wed, Apr 17, 2024 at 10:22 AM Eero Tamminen <oak@xxxxxxxxxxxxxx> wrote:

On 2024-04-03, Anders comment that:
-----------------------------------
after some further tests, it looks like the stuttering only happens "live".
I tried to video record the problem (using Hataris built-in recording),
and to my surprise there was no stuttering in the recorded video. I used
the MP2 example for testing.

So Hatari seems to be rendering the sound correct but somewhere in the
audio output it goes a bit bananas.
-----------------------------------

I.e. those issues seem related to host performance and/or SDL2 audio
processing rather than to Falcon emulation?

It's definitely not host performance. I tried running our Falcon030 demo "It's that time of the year" on my computer at work (i9-13900@5.8GHz/Windows) and I get the audio stuttering described by Anders in its original post.

--Daniel