Occasionally users may encounter a stereo source file that contains a significant amount of phasing or delay between the left and right channels, which prevents your file from being transcribed properly. These steps can help work around the problem and get your transcription up and running.
Understanding Descript Transcription
For context, each time a file is added to Descript it creates a Compressed version of the file (used for transcription) and an Optimized version of the file (used for playback). Compressed files are summed to mono and sent off for transcription to speed up the transcription process. In very rare cases, files with severe phase or delay issues can cause degradation of the transcription file, resulting in partial or missing transcript.
Resolving phase issues before they happen are the best way to remediate (more on this in the "Understanding Phase section below).
Getting Started
First, you'll want to create a new source file by clicking on the dropdown arrow next to the project name and selecting Project Settings. If it's not already, check the box to Download Original Files then click Save Changes when complete. This will make sure your newly created files are original quality, instead of optimized.
Navigate to the Media Library by clicking on the button in the top-left corner of the application and locate your file. Select the ellipsis (...) button and choose Separate into mono files. This will create two new files based on channel 1 and channel 2 of your file.
Click on the ellipsis button next to one of the files and select Transcribe file. Before doing so, you may click on each of the two files and select the Play button in the preview window below to audition it and see if one is preferable over the other.
Your file should now transcribe normally.
Understanding Phase
Phasing is an effect in audio where the crests/troughs of two waveforms are out-of-sync with each other, causing an audible delay or warping effect (especially noticeable in headphones).
When two combined audio signals are in-phase, their waveforms are perfectly aligned. This results in a perceived boost of the volume of the signal:
When the two combined signals are out-of-phase, or offset from each other, the signal will be lowered. If they are completely out of phase, the resulting signal may even be canceled completely.
In a real-world recording, it may be a little harder to see the difference.
Phasing issues in recorded speech are generally created when two microphones have been used to record a single speaker but one of the microphones was placed at a particular distance away from the source/other microphone.
To avoid an out-of-phase signal, the second microphone should either be placed as close as possible to the first microphone, or the more distant microphone should follow the 3-to-1 rule of microphone placement:
When placing multiple microphones for a given sound source, the microphones should be at least 3 times the distance apart from each other as the closest mic is from the source.
In other words, if the closest microphone is 10 inches from the source, the second microphone should be 30 inches from the first microphone.