For part three, I decided to take on a different kind of issue - to improve transcript for a podcast episode. The process was similar to what I had done for my previous PRs. I forked the project and created a branch for the issue. However, since the work involved editing a markdown file, I decided to use GitHub's built-in editor to work on it instead.
The improvement process involved listening to the episode while reading through the existing transcript. I would pause as needed and fix any issues which I notice. These mostly involved a few missing words, missing pronunciation, incorrect pronunciation, and some incorrect words. For example, git pull was transcribed as git poll, O'Reilly as a Raleigh, abstraction as obstruction.
Though I was quite surprised by how accurate the auto generated transcript was. The only places where it would mess up was when there was a slight overlap in conversation, or when terms specific to software development, or names of publishers were being said.
Throughout this process, I learned a bunch about transcribing - what may be ignored (phrases such as um), when to be exact, how and when to transcribe laughs, background music, etc.
And once I was finished with the changes, I created a PR.
For further actions, you may consider blocking this person and/or reporting abuse
Top comments (0)