Emulator

I've noticed that too, maybe our clock cycles/op are wrong?

Our cycle estimates are extremely unrealistic: the chip is superscalar, and we don't model the pipeline at all.

Could this explain why the emulator doesn't slow down properly in one of the Spyro games, causing a few tricks to be impossible?

Possibly? You can experiment with the emulated CPU speed by modifying the setting of CPU_CORE_CLOCK in PreInit() in SystemTimers.cpp (line 260).

The other big potential source of slowdown which we don't model is the graphics pipeline: we currently assume graphics rendering completes instantly.

magu: only on single core. On dual core, the game may wait for the gpu and waste emulated time. So is it required to overclock the emulated cpu on single core?

I tried this same trick (OCing the GC) on MegaMan X Collection, and it doesn't seem to help the stuttering audio in those FMVs. The stuttering audio seems to kick in after 15 - 20 seconds, and then goes away eventually, usually after another 15 - 20 seconds.

It seems as though something is going on regarding the CPU Timing. Two people have confirmed it's working in this Pull Request. https://github.com/dolphin-emu/dolphin/pull/1316

https://dl.dolphin-emu.org/prs/pr-1316-dolphin-latest-x64.7z

It's a hacky WIP and not guaranteed to be the correct fix, so not marking as fixed in PR or anything like that.

In 4.0-4424 audio stutters two times on the oppening video (with HLE and LLE). I've tried with pr-1316 and the problem disappears.

Issue 8382 has been merged into this issue.

Hey, I'm having the audio stutter on Dolphin 5.0, I tried doing some research, found something interesting: disc 1 stutters, but the same intro on disc 2 has no lag or stutter at all. I used HxD to compare the title.tpl (I'm assuming that's the intro, no clue), and they both have the exact same binary.

It almost seems like there's a setting affecting disc 1 or 2 differently, even though they both are gqseaf. Hopefully disc 1 can read audio propperly eventual. Hope this helps

Just so people don't get confused by this comment in the future, let me clarify.

Disc 2 doesn't stutter because it uses a lower quality video than disc 1. For some reason the opening is significantly smaller (just over half the file size of disc 1's opening) which makes it less intensive to decode.

The problem here is that our CPU timings are incorrect and something that uses the near maximum of the GameCube's processor is lagging in this situation due to timing differences.