My work deals with videos. I take them from DV tapes to viewable, downloadable video on a website. In fantasy land I can just copy video from a camcorder, upload it, and I’m done right? Nope! HTLM 5 attempts to solve the problem with the new video tag. But some people (Apple) want to use various MPEG files like H.264 for HTML 5 Video. Well that isn’t practical. Yes H.264 is a good format. H.264/MPEG-4 have patents and users must pay a fee to use them. Distributing them with say Firefox, isn’t possible. Second lets say someone paid a lot of money to allow it’s distribution. I can’t redistribute software that encodes MPEG4 for free (legally). So my client’s can’t transcode their videos. Do I tell them oh just send a check too…yea right.
There is a solution hack
Ogg Theora It handles video great and has a quality to disk space ratio competitive to H.264. It also handles shaky camcorders better which is huge for what I do. I can easily make a script with ffmpeg2theora that converts dv to ogg theora. I use virtualdub for extracting the DV tape video and avisynth for editing if anyone is curious. Now it’s easy to convert a large avi into a much smaller ogv file ready for uploading.
Next the server runs cron jobs to encode it as flv flash and MPEG1. MPEG1 is sometimes requested by the same companies that run Internet Explorer 6 with no plans up updating it ever. Flash if for backup since most people won’t have a browser than can play HTML 5 video. Since the flash video in my case is only for previews really, I can drop it’s quality big time so the flv file doesn’t chew through disk space. MPEG1 quality is dropped too because MPEG is horrible. Only a few clients need it so tough they get low quality. I play the flash with FlowPlayer and play the video with something like this
<video src="file.ogv" >
Play flash with flowplayer if html5 video fails.
</video>
Now if fully digital camcorders would get cheap, we could stop using those intensely annoying DV tapes.