A bit ago, I was working on a video project with a tight deadline. The production plan was timed down to the minute for release, and it was still going to be close. As it goes with most projects, there are often still things that pop up that weren’t accounted for in that early planning process. In this case, the video was required to have a descriptive audio file accompanying it at release, and it was our team that was supposed to supply this. Having a fully accessible experience was important to our team, and, in this case unfortunately, there was a simple communication error for who was providing the various parts. Nevertheless, this was a blocker that needed to be resolved. The video was due in an hour. Everyone was busy doing their part. How were we going to come up with a descriptive audio file in time?

I had the script from the storyboard stage that described each scene, and I wondered if I could make some small tweaks and use the say command in MacOS to read it aloud and match the visuals. After a quick search on Stack Overflow, I found this solution and we were off to the races. Forty-five minutes remaining.

say -f script.txt -o descriptive-audio.aiff

The parameter -f is the input file that is spoken, and -o being the output file (try -v if you want to use a different voice than the one used in your system default).

I added some additions of [[slnc 100]] in various parts to delay between sections (this took a few tests to time correctly).

Lastly, I used lame to convert it to .mp3 and our team had our audio description file. The disaster was avoided and we had about thirty minutes remaining 🎉. If you have any one-liners that saved your day, we’d love it if you shared them with us on Twitter!