Need Help Resolving Audio Issues in OBS

[TudorAlexander]

Hi guys, I am hoping you can help me figure out a few things with my OBS setup because it is driving me a little crazy. I use the vertical plugin to be able to record my camera and screen simultaneously for two video files. However the audio is the issue, and there are several issues.

1. The first issue is that I record the audio in Audacity for my good mic, and in OBS I record using the webcam mic (just so I can line things up easier). The audio from audacity is ALWAYS shorter by a certain amount of frames depending on the length. On a 3 hour recording it is like 10 seconds. People have told me this is because the frame rate is off, but there is no frame rate setting in Audacity. The bit rate is the same as OBS - 48k - so I don't see why this is happening. It is an extra step for me to match the audios, then change the duration so they are the same length and I would love to just not have to do that. I used to record the audio for my main mic in OBS, but for some reason it was creating TWO instances of this microphone - one on the video and one on the screen. Even though the source for the video did not have any input from microphone, it was only places in the display sources. Anyway, doing this with my main mic on OBS would cause funky artifacts to happen in the audio so that is why I decided to do what I am doing now. But I have the duration problem.

1. A second issue is that I record my screen and sometimes have audio that is played on a video or whatever else I am using. I would like the audio for OUTPUT to be in its own track, separate from INPUT audio. But I can't figure out how to do that.

Attached are my OBS settings. Strangely I don't even have the input or output audio added in "sources" yet it still records the input and output, so I don't know what that's all about.

 

[konsolenritter]

To the first point:
Audio differs because you are using two "soundcards". Whatever your "good mic" is bound to (in terms of what soundcard its connected to), the webcam mic via usb will be another audio source (so to call "soundcard") for your computer. These two soundcards do have their own clocks. Well, even driven both at 48k they drift over the time from each other. A couple of seconds is typical for minutes or hours of recording.

These kind of issue is normally fixed by using an (well, expensive) multichannel audio card at the computer. Different inputs to differnt outputs (software instances) routings possible, but all is going thru a single one soundcard/clocking device. So the windows audio system don't see different clocks that would drift over time. All receiving software will be clocked identically.

Don't know what kind of funky artifacts you mean. You should try to solve the origin issue and get rid of the additional Audacity step.
A common issue is to have the same audio source twice in the system: once as soon as the source is activated in the current scene, and maybe a second time via the global audio device settings. You should prefer one of them and drop the other, keeping the global entry XOR the scene dependent source, not both.

To the second point:
Routing differnt input tracks in OBS to different output tracks in the recording isn't difficult. You find the solution in your second screenshot, right hand. Right from "audio monitoring" you find the matrix of the "tracks" where you can route to or switch off single sources.

Don't forget to enable the count/number of the used output tracks in the global obs settings for the recording output (allow 2nd or 3rd stereo output track on the mkv file). Beware: You have to remix those tracks afterwards in a post production like Davinci Resolve, Premiere or something alike. In the VLC player you can check at least by switching between the different recorded audio tracks while playing.

 

[TudorAlexander]

To the first point:
Audio differs because you are using two "soundcards". Whatever your "good mic" is bound to (in terms of what soundcard its connected to), the webcam mic via usb will be another audio source (so to call "soundcard") for your computer. These two soundcards do have their own clocks. Well, even driven both at 48k they drift over the time from each other. A couple of seconds is typical for minutes or hours of recording.

These kind of issue is normally fixed by using an (well, expensive) multichannel audio card at the computer. Different inputs to differnt outputs (software instances) routings possible, but all is going thru a single one soundcard/clocking device. So the windows audio system don't see different clocks that would drift over time. All receiving software will be clocked identically.

Don't know what kind of funky artifacts you mean. You should try to solve the origin issue and get rid of the additional Audacity step.
A common issue is to have the same audio source twice in the system: once as soon as the source is activated in the current scene, and maybe a second time via the global audio device settings. You should prefer one of them and drop the other, keeping the global entry XOR the scene dependent source, not both.

To the second point:
Routing differnt input tracks in OBS to different output tracks in the recording isn't difficult. You find the solution in your second screenshot, right hand. Right from "audio monitoring" you find the matrix of the "tracks" where you can route to or switch off single sources.

Don't forget to enable the count/number of the used output tracks in the global obs settings for the recording output (allow 2nd or 3rd stereo output track on the mkv file). Beware: You have to remix those tracks afterwards in a post production like Davinci Resolve, Premiere or something alike. In the VLC player you can check at least by switching between the different recorded audio tracks while playing.

Yes it seems that you can specify which tracks to record to and then rip them out with Handbrake and that will solve the issue of using multiple programs to record.

To be clear the issue wasn't conflict between good mic and webcam mic - the issue was that my microphone (good mic) was being recorded by both my vertical plugin (webcam video) and display (screen video). So I have two videos being produced, both recording the good mic audio. This created artifacts in the production, like when you do too much noise removal. That kind of thing. Times when the sound would just cut out, as if it's being pinged or something. Never had those issues in audio recording just straight one audio track in Audacity, for example.

My confusion still lies in how exactly this is configured. I have two scenes - the normal one and the vertical one. On the normal I have 1 source - the display capture source. On the vertical source I have the video capture device. Why are they even recording audio if neither of them have my good mic as an added source? This I don't understand, because I thought they would record only based on the sources you put.

I also don't understand which "global audio devices" to disable or enable. if I enable one mic, it applies it to both of these scenes, and recording with two instances like this seems to cause problems for whatever reason.

 

[konsolenritter]

You nailed it, without doubt.

Again: Sources are scene-dependent (defined within a scene or referenced at more than one scene), or they are globally. Thats exactly that fact.
So again: You should bring in audio resources as global audio device EXCLUSIVELY OR as a scene source, not both.

Regarding the thing with "as too much noise removal". Maybe drivers get unstable in the windows audio architecture when called/used by more than one software/plugin at a single point in time, or (maybe additionally) its the windows system which does unwanted "enhancements" to the audio. Within the windows audio settings in the deeper dialogues of audio device specific settings: Whenever you see a checkbox like "enable windows sound enhancements" you strongly should DISBALE it.

 

[TudorAlexander]

What I am looking to do is so simple, and yet despite watching countless tutorials and talking on forums I cannot produce it.

I need to record TWO video files - one of my display capture and one of my webcam - simultaneously. I need to also record my microphone and the sound coming out of the computer. I don't care if both of these videos capture the microphone, because that seems to happen (when I was using the vertical plugin) - but by doing so it was causing those issues. I do not see windows enhancements anywhere and my sound settings in windows are all kosher. I don't understand why I was experiencing random shut offs of the audio for a second (almost like it was reacting to loud noise and dampening it, but that wasn't what was happening). But this is what happened before when I was using the vertical plugin to record the webcam video separate from the display video.

I have now removed the vertical plugin, and tried using the source record plugin. I set it up exactly as they say, to be on "Always" and set up a hotkey to trigger the filter of source recording to come on or off. Well, if you hit the trigger like this: ON OFF ON in succession, it crashes OBS. It also isn't triggering the program to work and produce two videos simultaneously like they show in the videos.

I don't see why this is complicated. I managed to get the video I do produce to have separate audio tracks, which is useful, but now I need two separate videos.

 

[TudorAlexander]

I recorded a quick video about the Source Record plugin not working:

Setting up Source Record Issues

Hey there! In this video, I walk you through setting up source record for my scenes. I encountered an issue where despite following all instructions, only one video file was recorded instead of two as expected. I explain my setup with audio devices, display scenes, and video camera filters. Your...

www.loom.com

 

[konsolenritter]

Trying to solve an issue you open up another one.

I clearly see why all this is complicated. You want to produce two videos at once with different content. This ISN'T EASY. In the world of video mixers this is requiring a 2 M/E mixer, what you are looking for. Quiet expensive tech stuff, believe.

Hopefully anyone with experience of those two plugins may provide help here for you, but stick with it: It IS COMPLICATED. The (way more) natural solution instead would be to go with PiP (Picture in Picture), showing the display capture in full size while having an small overlay of your webcam in one of the corners. Hence ONE video file containing both in the same area.

Some people do record an (one) OBS video with twice the width in pixels. Being an extreme wide video, with one content in the left half, the other in the right half. Later then (in post-production by Premiere or Resolve) they may split (render) this into two target videos with normal aspect ratio. Maybe that idea helps you out?

By the way: Random shut-offs of audio for a second or so is exactly my experience when one soundsource is tried to be captured by software(s) two times at the same time. It works a couple of seconds and the other second not.

 

[TudorAlexander]

Trying to solve an issue you open up another one.

I clearly see why all this is complicated. You want to produce two videos at once with different content. This ISN'T EASY. In the world of video mixers this is requiring a 2 M/E mixer, what you are looking for. Quiet expensive tech stuff, believe.

Hopefully anyone with experience of those two plugins may provide help here for you, but stick with it: It IS COMPLICATED. The (way more) natural solution instead would be to go with PiP (Picture in Picture), showing the display capture in full size while having an small overlay of your webcam in one of the corners. Hence ONE video file containing both in the same area.

Some people do record an (one) OBS video with twice the width in pixels. Being an extreme wide video, with one content in the left half, the other in the right half. Later then (in post-production by Premiere or Resolve) they may split (render) this into two target videos with normal aspect ratio. Maybe that idea helps you out?

By the way: Random shut-offs of audio for a second or so is exactly my experience when one soundsource is tried to be captured by software(s) two times at the same time. It works a couple of seconds and the other second not.

I will take a look at that, sounds like a good idea.

Regarding the soundsource being captured two times - this is the problem with recording two videos. They both have the sound and I think that's the issue, which is why I was using audacity to get the audio and OBS just with webcam (to match it) - but audacity records at a different clock for some stupid reason.

How do you set the camera to record on a large canvas?

 

[TudorAlexander]

Nevermind, figured it out and works perfectly now. Thank you for your help!!

 

[konsolenritter]

Glad to hear! :D
Let's participate other people (who stumble upon this thread later) what's the solution for you now...

 

[TudorAlexander]

I set the screen and canvas to 3840x1080 (double width)
I created scenes for each source (audio devices + video camera)
I nested these scenes in a main scene
I then duplicated the camera scene inside the main scene, so two displays
I modified the camera and cropped it so I have one side of the video screen + camera and another side just camera
In Resolve I just crop one side out from the other by duplicating the original video

Easy peasy!