Post #10535

For months I've been using this pipeline with shit like jpegoptim for my email newsletters. Last night, for a laugh, I used ffmpeg to prepare images for the mailer. The size difference versus a naive jpegoptim run was within 0.2%, just with the defaults. This shows me two things: 1. We always make things harder than they need to be 2. ffmpeg is amazing But we already knew all that.