Emulator Issues #12989
open
Super Monkey Ball 2 sound effect distortion
0%
Description
Game Name?
Super Monkey Ball 2
Game ID? (right click the game in the game list, Properties, Info tab)
GM2E8P
MD5 Hash? (right click the game in the game list, Properties, Verify tab, Verify Integrity button)
29dd871cf4a00b2455c9d03e9328e7bd
What's the problem? Describe what went wrong.
Many sound effects, such as narrator voices, are distorted as if run through a pass filter. This is a regression from older dev versions and does not affect the first Super Monkey Ball (GMBE8P). Both the Cubeb and WASAPI backends have this problem.
What steps will reproduce the problem?
Affected sounds play when selecting in a menu, but the most noticeable reproduction is when the monkey flies up after completing a level. The sound should sound more distorted and lower quality compared to previous builds.
Is the issue present in the latest development version? For future reference, please also write down the version number of the latest development version.
Yes, 5.0-16974.
Is the issue present in the latest stable version?
No, 5.0 stable does not have this problem.
If the issue isn't present in the latest stable version, which is the first broken version? (You can find the first broken version by bisecting. Windows users can use the tool https://forums.dolphin-emu.org/Thread-green-notice-development-thread-unofficial-dolphin-bisection-tool-for-finding-broken-builds and anyone who is building Dolphin on their own can use git bisect.)
Unable to git bisect due to build errors on Externals/fmt/include/fmt/core.h after going older than 5.0-16453, however I was able to confirm that version has this issue as well.
If your issue is a graphical issue, please attach screenshots and record a three frame fifolog of the issue if possible. Screenshots showing what it is supposed to look like from either console or older builds of Dolphin will help too. For more information on how to use the fifoplayer, please check here: https://wiki.dolphin-emu.org/index.php?title=FifoPlayer
N/A, not a graphics issue
What are your PC specifications? (CPU, GPU, Operating System, more)
Windows 10 Home version 21H2 on Intel Core i7-10750H (2.6GHz) CPU and Nvidia GeForce GTX 1660 Ti GPU
Is there anything else that can help developers narrow down the issue? (e.g. logs, screenshots,
configuration files, savefiles, savestates)
[Anything else here]
Files
Updated by JMC4789 over 1 year ago
Are you sure that the game isn't just using the low-pass filter on those effects? Have you compared recordings to console to make sure? Dolphin didn't emulate the Low-Pass filter in old builds in HLE audio. If you try LLE audio in the old builds, does it sound like HLE audio in the new builds?
Updated by MandL27 over 1 year ago
Just tested LLE on 5.0-16177 (pre-regression) and it sounds the same as HLE, so I doubt it's that.
Updated by MandL27 over 1 year ago
Tested both HLE and LLE on 5.0-17021. LLE sounds fine, HLE still sounds filtered.
Updated by JMC4789 over 1 year ago
Sounds like HLE is busted in its low pass filter in some way. I remember this game being mentioned in the long, long ago.
Updated by MandL27 over 1 year ago
That'd explain it, then. LLE audio is a workaround in the meantime, at least.
Updated by JMC4789 over 1 year ago
- Status changed from New to Accepted
- Milestone set to Current
Sounds like the low pass filter is being applied to sounds that don't use the low pass filter in LLE audio.
Updated by flacs over 1 year ago
The HLE LPF was re-enabled in 5.0-14712, but Super Monkey Ball 2 doesn't use it. The main suspect is probably https://dolp.in/v5.0-16434 which is a change related to DPL2 mixing. Can you check? (For Windows you can just download old builds from the website.)
Updated by JMC4789 over 1 year ago
I'm stuck on Linux at this point, but I can probably compile it to make sure assuming no big dependencies break.
Updated by flacs over 1 year ago
I've tried to reproduce this but I'm failing to hear a difference. Can someone else bisect please?
Updated by AdmiralCurtiss over 1 year ago
Yeah this is definitely different between HLE and LLE, though I had to forcefully remove the music from the ISO to actually be sure.
Let's see where this comes from...
Updated by AdmiralCurtiss over 1 year ago
https://github.com/dolphin-emu/dolphin/commit/855dbf88cac56541162ed038127276f7177bec59 alters the output, though it's still different from LLE.
That's https://dolphin-emu.org/download/dev/27c6731a660fb301ec43db31b87d69ae6e66568a/ (5.0-16434) for reference, previous build https://dolphin-emu.org/download/dev/256e3b071e5807a9ff8c264491f82b775b345fe0/ (5.0-16431).
Updated by flacs over 1 year ago
I'm aware of the LLE vs HLE differences. LLE presumably uses proper DPL2 matrix encoding while HLE just mixes the surround channels without any phase shift. But according to OP there was an HLE regression which is what I'm failing to repro. How does the HLE audio compare before and after the DPL2 mixing change?
Updated by AdmiralCurtiss over 1 year ago
- File HLE old 1.wav HLE old 1.wav added
Here's how it sounds with https://github.com/dolphin-emu/dolphin/commit/855dbf88cac56541162ed038127276f7177bec59 reverted.
Updated by AdmiralCurtiss over 1 year ago
And yes, checking, that does match the output of 5.0-16177, which is the version indicated by our reporter as pre-regression -- once you factor out the dumping audio rate changes from https://github.com/dolphin-emu/dolphin/pull/10762.
Updated by AdmiralCurtiss over 1 year ago
fwiw I suspect this is actually a 'more accurate emulation reveals problem' kinda thing, in this case the fact that our HLE does not properly handle the Dolby Pro Logic stuff.
Updated by MandL27 over 1 year ago
Just tested both on my end and I can confirm 5.0-16434 is the first broken version.
Updated by flacs over 1 year ago
Two observations from earlier today:
- old HLE is identical to LLE downmixed to mono
- new HLE is identical to the left channel of LLE (Can someone confirm just to make sure I'm not misremembering?)
So either that's the effect of the missing phase shift and to be expected or we're doing something wrong with the right channel.
Updated by MandL27 over 1 year ago
(Can someone confirm just to make sure I'm not misremembering?)
Yep, that's what the waveforms look like to me.