We took 2 files with the same 10m30s duration:
- ProRes 4444 video file (file size: 20.36GB)
- MP3 audio file (file size: 20.2MB).
Each file was imported into SpeedScriber and we timed how long it took from clicking the Transcribe button until the file appeared in the main file list with “Processing” status.
The test was repeated 10 times and here are the results:
Yes that’s correct, both files took 18 seconds to start transcribing.
How is that even possible?
Let’s go back… back to the time when human-based transcription first appeared as a service on the web. Media files were (and still are) large but internet connections were very slow. So to upload a file to one of these services you’d first have to convert it to a compressed audio format such as MP3.
Fast forward 15 years or so, and while internet connections have got much faster, there are no benefits to uploading a large video file because automated transcription only requires audio. Even today, all web-based transcription services suggest uploading a compressed audio file such as MP3.
So what’s the secret?
Whether 20MB or 20GB, the original file size makes no difference to SpeedScriber because the imported file is never uploaded. The magic is done behind the scenes — any imported file is converted to a mono audio-only compressed file which is uploaded and used for the transcription.
With SpeedScriber, there’s no need to spend extra time making a special ‘export for transcribing’ because the app does it for you. Automatically.