OBS and Descript issues

[unitz29]

I'm having some issues with recording online game streaming with OBS and using Descript to help edit/transcribe my audio.

My first issue that I discussed with Descript customer service was that the Descript program was not transcribing my mic audio very well. In fact, the timing was incredibly off most of the time. Customer service told me it was because the stereo was interfering with the transcription process, and they said:

"I suggest to only record in mono, split the channels to mono before importing one of the tracks, or assure the stereo channels are in-phase while recording (not always easy to accomplish depending upon your recording device). For your specific workflow, I would also recommend recording the camera and the game capture on separate tracks so that when your file is sent off for transcription, the game audio doesn't interfere with your spoken audio."

I followed guides to edit my audio settings, and have attached the file of the advanced settings I changed them to. I'm wondering if this set up is correct for not having any duplicate audio or audio interfering with each other when transcribing.
Also, after attempts to get Descript transcribing correctly by editing my OBS settings (AKA putting my mic and game audio onto separate tracks), I for some reason cannot upload any video to Descript. It gives me an error that says that "Too many audio streams were found: 2.. etc.".

ALSO: The way my settings are presently in the pictures only seems to record my mic audio, and not the game audio. I can only seem to hear the game audio in the background?
I just want OBS to record the game audio, my audio, and have Decript be able to transcribe it without me having to edit it for 3 days straight to have OK captions.

How do I fix this? I just want to record my stream on OBS and be able to upload that video to Descript to make captions easier.
I'll include some other pictures from OBS.

 

[unitz29]

*bump*
Really need help with this. What am I doing wrong?

 

[PaiSand]

For what you say, and I can guess from the error it gives, you need to use only 1 channel with all the audio.
Right now you have 2, channels 2 and 3.

More info is needed. And probably Decript support can give you the right answer.

 

[unitz29]

This is the last thing Descript support said,

"To clarify, you'd ideally want to have each input/source recorded as 1 file that contains a single mono or stereo audio stream. Alternatively, if you can set OBS to record a single video source but have separate files for the audio sources(game, microphone, etc.), that would work as well.

If the inputs being recorded from OBS are being encoded with 3 or more audio streams, I'd suggest reaching out to OBS Support or the forums for assistance on how to ensure that the files are encoded in a format that Descript accepts. I'm truly sorry that I'm not able to be of more help here."

I'm not sure if I understand everything they're saying here, but I'm pretty sure my OBS audio settings meet what they're describing?

 

[PaiSand]

Like I said before, use 1 track instead of 2. (sorry, I wrote channel instead of track)
You already sent all the audios to track 1, so you may want to record this one. Or, only use track 2 for the voice and the other audio you're recording. Or use something like Audacity (can't remember if it can do this) to extract the audio files individually and use only sent voice track.

 

[unitz29]

Wait, so the second option of only using track 2 for voice and the other audio. Do you mean it should look like this? And the result won't cause Descript to transcribe poorly because it's detecting the game and mic audio?

 

[koala]

To make your transcription software recognize just your mic and nothing else, your mic has to be on its own audio track, and nothing else must be on that audio track. It also seems your transcription software expects at most 1 audio track if you upload video. Not 2 or more. This is probably because it isn't able to guess which one is the mic track if you include multiple audio tracks.

However, you want more audio recorded than just your mic. Your game audio and whatever more. So you must record to multiple audio tracks with OBS, otherwise it isn't possible to get all audio you want. One audio track must contain your mic only for transcribing, and one or more audio tracks will contain the other audio.

To send just your mic audio to the transcribing service, you need to extract the mic audio track from the video file after recording has been completed. There are several tools for this, see https://www.google.com/search?q=extract+audio+from+video
Extract to some format your transcribing software understands, then upload just the extracted track instead of the whole video.
After extracting, you can delete that track from the video file, or you can start some audio postprocessing and mix it with the other track(s) you also wrote to the video. The result of that postprocessing is a mixed audio track with what you want to distribute, so replace the original audio tracks from your video with your mixed audio track.

You have more options.
In case you don't want to postprocess audio, you can write just 2 audio tracks to the video: one with the audio mix you want as final mix, and one with the mic only. You already prepared such a setup in the screenshot from your OP: You write every audio source to track 1 as mix, and additionally mic to track 2 and Stardew to track 3. You disabled writing track 1 to the video, so you don't get that mixed track. I recommend you activate this again, so the track numbers in OBS correspond to the track numbers in the video. If you do it this way, the track containing your mic is the 2nd track, so you need to extract the 2nd audio track from the video and upload this to your transcription software.
In track 1 all your audio is present as mix, including the mic as well, so you can use that track for further processing.
Or you take tracks 2 and 3 and create your own audio mix and ignore track 1.

 

[unitz29]

Very good answer, thank you!
My English is not always the best, but please tell me if I misunderstood.
You're giving me a few options here, but one option is to record my general audios in track 1. I should stream track 1, have my mic audio only in track 2. Then I make a video, and then use a program to extract just track 2, then upload just track to so it transcribes the audio. Then I can add the video?
I feel like the video will still have 2 audio sources (track 1 and track 2) and won't accept the audio file. Unless extracting track 2 reduces the video file to only have track 1?