 No.22382
Tafoid Administrator
Oct 23, 2024, 00:14
|
We need specific names of machines which are affected, please. The LLE version of QSound is supposed to be hardware intensive and it is my opinion that if your CPU isn't fast enough to run something, it doesn't make it an emulation error or issue to report. |
|---|---|
 No.22383
hap Developer
Oct 23, 2024, 11:03
|
I tried "dino" and yes, music is slow and I hear some other glitches. MAME runs at 100% speed. |
 No.22386
VasiliyFamiliya Tester
Oct 23, 2024, 14:01
|
The Punisher and Warriors of Fate also shows absolutely same results. |
 No.22392
StHiryu Tester
Oct 25, 2024, 12:04
|
Really not related but, yesterday I played The King of Dragons and noticed some game slowdowns, the most important after defeating the final boss of the game, when a lot of items appears on the screen. This never happened to me in previous versions... |
|
No.24139
VasiliyFamiliya Tester
10 days ago
|
Can it be related with an incorrect Z80-QSound connection timings? |
 No.24140
cuavas Administrator
9 days ago
|
With the cycle-granular Z80 and new memory system features, it might be time to revisit this stuff. |
|
No.24150
VasiliyFamiliya Tester
2 days ago
|
Z80 clock is same in both Dash and CPS2... So looks like in Dash case some mysterious extraneous factor presents. |
 No.24151
Kale Developer
2 days ago
|
vgmplay has sound issues with these games too (and don't think the "bad rip detected" I'm getting is related). |
|
No.24152
VasiliyFamiliya Tester
2 days ago
|
This comment from cps1.cpp catched my eye: const attotime audio_irq_period = attotime::from_hz(XTAL(8'000'000) / 32000); // measured 250Hz (cps2.cpp) Does CPS Dash really have no any own timing nuances? |
|
No.24153
cuavas Administrator
2 days ago
|
Just stop. You aren’t helping |
|
No.24154
cuavas Administrator
2 days ago
|
> vgmplay has sound issues with these games too (and don't think the "bad rip detected" I'm getting is related). The trouble with the VGM rips is they’re recorded with QSound HLE, so the Z80 could write to the window in DSP memory as fast as it wants to. In reality, Z80 can only process one write per sample interval, and the Z80 has to wait until it’s ready before doing another write. So you take the VGM recorded with the magical QSound HLE that accepts writes as fast as you push them and try to play them with the real DSP program running and writes get missed because they’re coming too fast. It might be possible to implement a FIFO in vgmplay to work around it. What’s a good VGM for testing? |
 No.24155
JimCarlTay Senior Tester
2 days ago
|
Maybe 1944: The Loop Master (vgmplay:1944), Cadillacs and Dinosaurs (vgmplay:dino), Tetris: The Grand Master (vgmplay:tgmj), and X-Men vs. Street Fighter (vgmplay:xmvsf). I could be just pulling those out of a magic hat, but I'd bet they'd be the best possible testing candidates. However, only Cadillacs and Dinosaurs is the only CPS Dash soundtrack in this group, as the rest are CPS2. |
|
No.24156
JimCarlTay Senior Tester
2 days ago
|
EDIT: Except tgmj, which is ZN, actually. |
|
No.24157
cuavas Administrator
2 days ago
|
Tetris: The Grand Master is not on CPS at all, and most of those aren’t on CPS1 Dash. Have you actually tested that these exhibit the issue? I’m getting really sick of wasting my time having to address stuff that doesn’t hold up to basic critical thinking. All the time I have to waste on that gets in the way of actually getting stuff done. If this doesn’t change, I’m going to avoid MAME Testers as well. Please think before you comment. |
|
No.24162
JimCarlTay Senior Tester
2 days ago
|
I'm sorry for being such an idiot. |