Either OBS or Elgato is acting up, causing echoed game audio on playback

[Blackjack Gabbiani]

https://youtu.be/_bATeqyzE24 Here's my sound test (note, though this is TotK there are no spoilers here). Everything sounds clear to me, but when I play it back, the game audio is doubled up.

I expand more on the circumstance in the description, including why I'm rather...hesitant to come here for help given a bad experience with a moderator insulting me.

Anyway I'd really like to have this solved soon. The only thing I didn't say in the description is that I wasn't fully updated when the problem first happened yesterday, but updating didn't change anything.

 

[AaronD]

Are you Monitoring the game in OBS, with the same device that also has a Desktop Audio capture? That'll do it.

The game audio comes in through its intended source, goes out the Monitor to that device, mixes with everything else that goes to that device, and then gets picked up by the Desktop source that watches the same device. It takes some time to make that round trip, thanks to Windows' buffering, so you hear two distinct copies.

 

[Blackjack Gabbiani]

Ok I figured out what it was. OBS had added an extra audio channel that I never added--a channel for desktop audio. Though it's funny because I'm not recording TO the desktop. My footage goes straight from the Switch to OBS. I don't know why it would do either of those, add a channel on its own, or pick up audio from a source it shouldn't apply to.

 

[AaronD]

On a fresh install, OBS has two global sources - Mic and Desktop - that are both set to Default, which lets the operating system choose what device it is. That's to get someone going as soon as possible, who knows nothing, and just wants it to "work" right away...whatever that means. If you've never changed that, then it's still there.

And you really *should* change it - in fact, look at *everything*, understand it, and change a lot of it - so that it does what YOU want, and not what some programmer thinks "most people" would want...which probably fits nobody no matter what they set as default. But what it actually is, is not a half-bad approximation of the average. Still not particularly *good* though, nor can it be, given the wide variety of things that people use it for.