Separate Audio capture devices for separate scenes

thoward

New Member
I have two scenes which are cameras in two rooms and I want to switch between them.
I want a stereo audio feed from each room to go with the camera feed.

Am I right that OBS doesn't handle multi-channel sound cards particularly cleanly, so rather than buying a 4-input sound card and having inputs 1&2 for one scene, and inputs 3&4 for another scene, I'd be better to buy two stereo input cards and have a separate physical audio input device for each scene?

I can't see a way to separate out inputs from a multi-channel sound card within OBS. I am on Windows no OSX.
 

konsolenritter

Active Member
Look if your fourchannel-device's default sound driver presents the channels already in stereo-pairs to windows. That's the default behaviour often and exactly what you want, isn't it?

What is separated by the driver already, OBS can handle just this way.
 

thoward

New Member
I have to buy a card specially but don't think it's that commonplace, although it might be more common in 4 input devices. I have a variety of 16/24/32 input devices here and none of them seem to do that, but that would mean 16 pairs of stereo inputs showing up.
 

thoward

New Member
On OSX I could use Loopback and create a virtual sound card separating them in software into virtual pairs of 2x channels that OBS could use but I'm thinking adding additional software is more effort and it might be easier just to add 2x interfaces. The capture devices are 2x ATEM Minis with line ins but the encoders have a ton of noise on them being unbalanced.
 

konsolenritter

Active Member
...variety of 16/24/32 input devices here and none of them seem to do that, but that would mean 16 pairs of stereo inputs showing up.

That's right. Devices handling so much channels are more supposed to have asio drivers at hand.

Originally you wrote about four channels: What can possibly work in this case maybe VOICE MEETER (a re-routable capable software).
 

konsolenritter

Active Member
Aaaahm. I forgot: If you like to hand-over the audio-channels like the cams by switching scenes, you're strictly encouraged to look for a single 4-channel-device. OBS runs into trouble (over long runtimes) if it has to handle different devices with (slightly) different clocks/crystals at the same time. Even if they all are set to the same samplerate!

Problems along a described here or here. The latter post is given example of onboard-sound versus elgato (audio fetched in from HDMI). But the principle counts...
 

thoward

New Member
Thanks, is this problem true of video capture devices too? Or would still be a problem with the audio if I used the audio capture off the 2x atems separately?
I would only have one audio input capture active in each scene, does this avoid the problem?

I will look at Voice Meeter for separating the card into virtual 1-2 and 3-4, I have used VB-Cable for routing already so hopefully can be done simply if need be.
 

thoward

New Member
I have ordered a Behringer UMC404HD in the hope that the drivers will have 1-2 and 3-4 separately as two stereo inputs. There seem to be a lot of people having problems with Audacity recording all 4 channels and once and complaining it comes up as 2 stereo cards but in this instance that’s what I want. Our other cards are Motu 24Ai / 828s etc so none of these come up as stereo pairs. Hopefully can do it without needing 3rd party software and the analogue audio inputs on these atems are terrible.
 

konsolenritter

Active Member
Regarding video capture devices it could be said that the timing of delivered audio over HDMI depends on the clocking of the sending side. So its definitely an independent clock source.

On the other hand nobody can guarantee (or even say), in advance, if "the problem" will arise or not on specific machines with specific combination auf audio and capture devices. I'm afraid that you have to check and test what works for you. There are only hints to prevent some situations.

Activating and deactivating of single devices by switching the scene can help to clear buffers but may introduce silent lags and pauses that are unwanted at scene changes.
 

thoward

New Member
I meant more if this was also a known bug with clocks on video capture devices - I'm not sure what kind of clock speeds they run at so I assume it may be a different ball game. Only as I have 2x ATEMs, one for each area, and I switch between those as well. I'm just unhappy with the audio quality from the ATEMs so hoping to add a 4x balanced input sound card to capture the audio separately.

I am planning to use this with Multiview on OBS which I assume keeps the video inputs alive, as they are visible in the preview window. However if this causes problems I'll lose it.
 

konsolenritter

Active Member
Oooh yes. These are _two_ problems. I too heard that people are unhappy with the audio rails of the ATEMs, but i don't own one to check for myself. So - in generally terms - i appreciate your step to separate sound card. For church services we do so accordingly: Fetching sound in good studio grade level and quality by XLR from a mixing desk on-site, then going thru DI-boxes (possible earth decoupling to prevent ground loop based noises and hum) then routing into such external soundcards.

A last hint: Then, if you work with the Behringer (or whatever) try to disable the other sound devices (for instance onboard sound) and use only the behringer. Hope that works well for you.
 
Top