What is this?

I took all the URLs from HTTP Archive that ended ".mp3" or ".mp4", deduped by host, then made a range request to them with the following:

To run these tests yourself, see the repo.


Warning: These URLs are media files from the web. As such many will be unsafe for work.