Here's a sine wave in brstm, dsp, and adx format for people that want to put it into a game that uses those music types.
And yeah, I know Skyward Sword uses AX HLE, I was just confirming the issue in another game.
Anyone can confirm if it's about number of frames reproduced or clock time?
With good hw reaching 150fps or more with no frame limitation would be a lot easier to reach the static noise in 1/4 the time.
I've modified Dolphin so that it dumps sound both in both real-time and in emulator-time, and I've got a sine wave playing in Brawl at 800+% speed. Let's see what happens.
We just confirmed on IRC that the issue happens after 1h30 real time, not emulated time. That's... weird and throws off my theory.
Yeah, I just confirmed that too. Guess I'll let it run at 100% for 75 more minutes. That just makes it that much harder to test.
Have to go for the night, it would be really valuable if some people tried Zelda UCode games with DSP HLE and check if the problem happens there too (that would exclude the AI from the list of potential buggy modules).
I'm running Brawl playing a sine wave in both hle and lle interpreter, and Mario Sunshine with HLE all running at a solid 100%.
(07-08-2013, 09:33 AM)delroth Wrote: [ -> ]Have to go for the night, it would be really valuable if some people tried Zelda UCode games with DSP HLE and check if the problem happens there too (that would exclude the AI from the list of potential buggy modules).
I could reproduce the static on Mario Kart Double Dash, one time, but I don't remember if it was with HLE or LLE, but if you wanna know if the static happens with zelda ucode, I can confirm it happens.
I can confirm it happens in Twilight Princess around the 92 minute mark.
I can confirm the save state save and load eliminates the issue for a time. I tested that with Kirby Return to Dreamland awhile back.
I will check my dumped audio files to see if there is an exact time stamp.
Update:
My findings concur with JMC47. My static happened at exactly the 92 minute mark...to the second...weird.
Between 90 and 92 minutes the static starts to get more frequent and the pops start to "build on one another". At 92 minutes all hell breaks loose. Not sure if that helps, but that is what my 2 audio dumps show me.
(07-08-2013, 08:05 AM)RachelB Wrote: [ -> ]Has anyone tried with no audio output?
Forgive my ignorance:
How would we know if the issue occurred with no sound output? Does selecting that option still output a usable dump wave file?