3 ms·

This is interesting. Having been part of a startup which focused on video upload, encoding and streaming I think there is a market for this service. Past experi

This is interesting. Having been part of a startup which focused on video upload, encoding and streaming I think there is a market for this service. Past experience has shown me that one of the unplanned significant overheads of a service like this is the myriad of esoteric codecs out there and what they do to ffmpeg and related encoding tools. There is always a source or device that produces a video file which causes ffmpeg to hang or segfault on transcoding, or to output a flawed transcoded file with poor video or no sound, etc. I'm curious how you deal with these 2 issues. The first issue being an infrastructure issue, where incoming media can cause problems for your tools, the second is a quality issue, where the output your tools produce may not be what the customer expects. What is your plan to handle these 2 problems?

felixge 16y ago
Thanks for the feedback. This is exactly the reason we are doing this, video uploading and encoding is one hell of a nasty problem - something people would be glad to knock of their todo list. In terms of dealing with the 2 problems: We've been working on this startup for 1.5 years by now, so we learned a lot of the things that aren't working by now : ). Basically we fix problems as they come in, try to get advice from the right people, and be smart about the way we deal with the command line tools. Thanks to node.js we can easily "time out" on hanging FFMPEGs and detect other weird error conditions. Anyway, I think our competition is ahead of us in terms of the encoding, but we'll catch up to them! We got the lead on uploading for the meantime : )