Skip to main content
Participant
May 12, 2025

Transcription does not detect pauses or filler words for channel 2 audio of a mp4 I recorded

  • May 12, 2025
  • 6 replies
  • 398 views

I have a mp4 with 2 audio channels. Each channel has a different person talking. I transcribed and it transcribed what person 2 was saying but it did not detect any pauses or filler words for them.

 

Now when I exported a .wav of just channel 2 and brought that into premiere, it detected the pauses and filler words fine.

6 replies

Ross01ACAuthor
Participant
May 19, 2025

I re-transcribed but still getting the same problem. the channel 2 speaker's pauses and filler words are not detected at all.

Ross01ACAuthor
Participant
May 19, 2025

I just hit the transcribe button after I imported the video. I'll try again following your example.

Stan Jones
Community Expert
Community Expert
May 17, 2025

@Ross01AC,

 

Thanks for sharing the file.

 

MediaInfo shows this to be a 54 minute, 2 mono channel file. On Win10, PR 25.2.3, it imports correctly for me, showing mono 2. I tested the transcriptions before creating a sequence, and when I dragged the file to the new icon, it created a sequence with the video track and 2 mono tracks.

 

I transcribed by right-clicking on the file in the Project Panel, picking transcribe, and channel 1. Transcription works, and the transcript shows pauses and filler words. And then re-transcribed, this time, picked channel 2. This transcription also worked, and included pauses and filler words.

 

How did you create your transcriptions?

 

Not directly related to your problem, but just documenting some issues in the workflow: You can only have one transcription per file. There is more than one way to keep both. This time, I extracted channel 2, which creates a new file on disk, and transcribed that. This allowed me to compare the two a bit. Accuracy: Speaker 1 transcript involved many instances of incorrectly labeling as Speaker 2 - this is not the speaker 2 in the other channel; this is the transcription perceiving a second speaker on channel 1. There is, of course, only one speaker on channel 1. On both transcripts, there were many "chuckles," which were always treated as pauses. Fillers were genrally accurate as "ums" or similar.

 

I'll also note that the extracted file transcription resulted in a .prmi file being created along with it. I have my Media Analysis & Transcription preference set to sidecar.

 

Stan

Stan Jones
Community Expert
Community Expert
May 16, 2025

@Ross01AC,

 

The fact that it transcribed at all eliminates most issues. Simply omitting pauses and filler words is odd.

 

Can you post MediaInfo in tree view?

 

Or better yet, share the file? Or one that shows the same problem?

 

Stan

 

Ross01ACAuthor
Participant
May 16, 2025

 

Here's the codec info from VLC on the mp4 in question. It is both video and audio.

 

Community Manager
May 15, 2025

Hi @Ross01AC

 

I'm sorry to see you're having issues with transcription.  Can you tell me what your audio settings are?  Is it set to mono or stereo?  Is the mp4 audio only or audio and video?  

We need a few more details to try to help with the issue. Please see: How do I write a bug report?  Sorry for the frustration and thanks for reaching out.