8-bit Sound Asset Creation Questions

Ask about editing graphics, sounds, models, music, etc here!
Shaders (GLSL) and SNDINFO questions also go here!

Moderators: GZDoom Developers, Raze Developers

Forum rules
Before asking on how to use a ZDoom feature, read the ZDoom wiki first. If you still don't understand how to use a feature, then ask here.

8-bit Sound Asset Creation Questions

Postby Hambourgeois » Wed Jul 08, 2020 8:24 am

I'll preface this by saying this problem is 100% caused by me wanting to be a lo-fi retro gaming dipshit and if its unavoidable I can just suck it up and use modern source port features.

I have a yeti microphone so I figured I might as well get some use out of it. I am trying to make a spirit wail/hiss sound thats quite a bit more subtle than the Wraithverge spirits sound so I hissed into my microphone and made a 5 second loop of it in Audacity. Ideally, I'd like it to be able to be in the normal doom sound format so it doesn't sound out of place among the other sounds. Anyway, here's what I got as 32-bit WAV files before downsampling:

Unprocessed loop - Link
Low-pass and reverbed version I'd like to use - Link

Now, when I used Audacity to export the files as 8-bit unsigned PCM files, they get eaten by white noise, especially the processed version. See here:

Unprocessed loop - Link
Low-pass/reverb loop - Link

Anyway, obviously these aren't really that usable in 8-bit and I am trying to figure out to what degree that's inherent to the bit depth, especially since this is a quiet, noisy sound which is, to my understanding pretty much the worst case scenario for low bit depth audio. I've tried most of the available dithering options in the Audacity quality pane and have had no luck with that. In this thread on Doomworld Xaser mentions that he had this problem with Audacity too and sticks to Goldwave. Should I look into that or is there a dithering plug-in/VST that people have had good luck with for this sort of thing? This admittedly isn't my area of expertise so if there's any sound people on here I would greatly appreciate any advice or thoughts.
User avatar
Hambourgeois
 
Joined: 06 Jul 2020
Operating System: Windows 10/8.1/8/201x 64-bit

Re: 8-bit Sound Asset Creation Questions

Postby Chris » Wed Jul 08, 2020 7:59 pm

The sound seems to peak at about -54dB, while a dithered 8-bit sound has a noise floor (the white noise) around -70 to -80dB. There's only about 15 to 20dB between the peak and static.

Generally with sounds, you'll want to have a recording use the full dynamic range available, and have the engine play it at a lower volume dynamically. In this case, if you instead have the sound peak at -4dB there will be 65 to 70dB between the peak and static, and if you set the sound' definition's volume (in SNDINFO) to be something like 0.003, it'll play back normally at about -54dB with the noise floor lowered with it, keeping the difference.
User avatar
Chris
 
Joined: 17 Jul 2003

Re: 8-bit Sound Asset Creation Questions

Postby Hambourgeois » Wed Jul 08, 2020 10:14 pm

Ok that actually helps a lot lol thank you
User avatar
Hambourgeois
 
Joined: 06 Jul 2020
Operating System: Windows 10/8.1/8/201x 64-bit


Return to Assets (and other stuff)

Who is online

Users browsing this forum: Steve5563 and 0 guests