Late summer reflections on GSOC participation @Orcasound

As the days are getting shorter I’m approaching the end of this season with a great joy for the experience that I had participating in the Google Summer of Code (GSoC) this year. Last week we had our closing meeting with the Orcasound Team and sincerely I feel a little nostalgic for the end of the program as I really enjoyed the time spent with our mentors and my colleagues. However endings are always the beginning of something else, and from the advice given to us last Friday during the GSoC student summit, this could be the beginning of an open source life-long contribution journey. Now, I want to share with the aspiring GSOC students of 2022 and possibly future Orcasound members some advice for your future Success.

  • Start Early: Review the issues on the orcasound repos to make some first contributions, introduce yourself in the orcasound Slack, and discuss with the orcasound team the ideas that you have in mind.
  • Work independently without forgetting to ask for help: It is common to find yourself stuck during GSOC at some point. Try to find a solution by yourself first and if you don’t have success after good effort, come back to your mentor with the initial research that you did and ask for advice.
  • Take advantage of git: Git is the Swiss army knife of open source. Get used to the workflow and also learn the extra tools that Github has to effectively plan your time and work asynchronously with the team (eg. Projects, wikis, issues).
  • Learn cloud resources management: Orcasound has their databases hosted in the cloud (mainly AWS). It is good to know how to interact with the data and also learn best practices. There’s a primer on how to access the data if you want to start early.

This summer I also learned a lot from ffmpeg. It is a really efficient transcoder of media formats, but also has a ton of utilities for handling audio and performing complex filtering. I want to share with you this tutorial that I found very useful to understand the pipelines on ffmpeg. As an example, with the following snippet you can get statistics from a segment of audio data amazingly fast.

ffmpeg -i <input_file> -af "astats" -f null -

Finally, I want to thank all of the mentors: Valentina, Jesse, Kunal, Scott, Val, Paul, and the rest of the Orcasound gang. It was an amazing summer learning and creating with your guidance. Congratulations also to my GSOC partners Dmitry, Dhananjay, and Isabella for your excellent work. The Orcas made an appearance last week so checkout this amazing entry from Scott analyzing the calls.

Leave a Reply

Your email address will not be published.