Video Processing

[kevin_standlee]

As time has permitted since the Fannish Inquisition on Friday, I have been processing the video from the various presentations and uploading it to a Westercon 68 Fannish Inquisition playlist on YouTube. This is a time-consuming process, involving converting the MXF files to MP4/MP3, then joining the video and sound in MovieMaker, then generating a combined file, then uploading the videos to YouTube. Not wanting to spend the entire rest of the weekend in my hotel room, I've done things in bursts. As of this writing, the seated Worldcons' and the 2017 bidders' (the three who showed up at Westercon, to wit: DC, Helsinki, Montreal) presentations are online, and the 2018 bids should be online by the time we leave tomorrow. If I'm lucky, I'll get the 2019 presentations online by the time we leave as well, but they might have to wait until later.

At some point I will upload one of the "quick" MP4 files generated by the proxy card in the camera so the people can get an idea what the difference between the camera's native format (albeit converted) is versus the "quick and dirty" recording. As I've said before, for the WSFS Business Meeting, our current plan is to upload the "quick" files as soon as we can. They have good sound but indifferent video, but they don't require all of that post-processing and can be simply uploaded straight to YouTube; thus people will be able to hear what happened at the Business Meeting the same day, rather than several days later.

Comments

[kevin-standlee.livejournal.com]

We think it has something to do with the transfer rate to the SD card. We're going to try with a faster SD card and see what happens.

[billroper]

My suspicion would be that it's less the speed of the SD card and more the speed of the processor in the camera that is dumping the data to the SD cards. If what's coming out on the SD cards is closer to the original resolution of the image, then the camera would have to do less processing. If it's trying to dump a high res image at low res, then it would have to do more processing and might be using an "optimized" algorithm to get it out, sacrificing quality for speed.

I am, of course, just speculating. :)