Creating an audiobook with help from OpenAI Whisper (Episode #T2)
Embedded Files
Ten Minute Projects for the Curious: https://www.youtube.com/@10minute1liners
In this video, I show how to build a m4b audiobook file from a set of audio files using very basic commands from OpenAI Whisper, ffmpeg and AtomicParsley. The code can be found at https://www.innovationincrisis.com/10minute1liners/episode-T2 and downloaded here.
The steps to create a m4b audiobook from audio files are as follows:
Concatenate audio files into one m4b file using FFMPEG
Run whisper on m4b file to transcribe and time stamp text from the audio
Edit the tsv file to create correct chapter/section/part delimiters and times
Create a chapters tag file and tag the m4b file
Add title, author and narrator to m4b file
Design and add cover art to m4b file
The steps can be done as follows (most as bash files with a one line command).
STEP 1: create audio file list and concatenate audio files to create m4b file
STEP 1: create audio file list and concatenate audio files to create m4b file
STEP 2: Run Whisper on m4b file
STEP 2: Run Whisper on m4b file
STEP 3: Find chapter delimiters in tsv file and edit appropriately
STEP 3: Find chapter delimiters in tsv file and edit appropriately
STEP 4: Create chapters tag file and tag m4b file
STEP 4: Create chapters tag file and tag m4b file
STEP 5: Add title, author and narrator to m4b file
STEP 5: Add title, author and narrator to m4b file
STEP 6: Design and add cover art to m4b
STEP 6: Design and add cover art to m4b
Page updated
Google Sites
Report abuse