To do with teaching, somewhat

Morning all.

It’s almost the end of the first “half” of Semester One and rather than working flat out on my thesis I find myself working flat out turning a course designed for hundreds of students into a course that functions for six; this is very difficult to do and I wish I had been advised just how small the enrolment was going to be before taking it on.

Still, one must make do.

As some of you are doubtless aware, I record my lectures and provide the narrated slides to the students. It’s a good idea for two reasons; I have to keep on my toes and make sure I’m actually being rigorous in my pedagogy… and the students benefit because they can revisit my explanations at a latter date (as well as catch-up should they miss a class).

The long-term plan is, once the thesis is finished, I will put, online, a series of videos on Critical Thinking; a sort of online primer made up on ten minute chunks of argument detection, extraction and analysis.

Which leads me to my related point; at the moment everything I put up online for the students is H264 video with either MP3 or AAC audio. Some of you might be aware that H264 is currently not the darling of internet nerds and geeks because it has patents attached to it and someone, somewhere, has to pay a license fee for you to watch material encoded with such a codec. There is a good argument against the use of such patented video codecs; programmers in developing nations are not usually in the position to pay the MPEG-LA group the necessary monies to license H264 and there are issues even in the developed world, with Firefox, famously, refusing to pay that license (and thus needing a little ledgermain to view H264 outside of a Flash container).

Theora is the darling of people who argue this way; it appears to be patent-free and Wikipedia, for one, is pushing it hard. Theora is based on a much older encoding algorithm than H264, but it’s free and, the argument goes, we should prefer to use free software whenever possible.

It also produces much larger, uglier files than H264, and this is what concerns me. I need my recordings to be small and tight; I don’t want my students to waste their precious bandwidth downloading these files (for foreign readers: most citizens of Aotearoa/Te Wai Pounamu do not have unlimited bandwidth in re their internet access) so size matters. Visual quality; not so much; it is mostly text and I can happily reduce the framerate down to about 8 frames per second, with infrequent keyframes, to little visual detriment.

The other issue is that Theora encoding is slow; in the time it took to write this post I could have outputted my H264, remuxed the audio and had coffee; I’m only 42% through the Theora encoding at the moment, which, to use the vernacular, sucks.


The resulting file is, well, bigger and slightly blurrier than the H264 encode. 20% larger is a significant difference in size and it seems pretty consistent; I’ve tried re-encoding several other lecture recordings and the results are bigger and blurrier every time.

I’m keen to use patent-free video codecs and when I do the video primer I’ll look into this in more depth to see how things stand, but at the moment, it seems patent-laden codecs really are more user-friendly and easier to work with.

Sad but true.


bi -- IJI says:

I’ve heard that the Adobe Flash video format supports some sort of lossless encoding of the visuals — can’t remember what it’s called — but I’m still not sure how to get the tools at hand to output this encoding.

In any case, I think a lossless encoding along the lines of GIF or PNG will be the preferred format for recording slides, in contrast to lossy codecs such as H264 (which are probably more suited for recording video from general webcams and stuff).


Ah, but Flash is the devil; it’s less open than H264 and often encapsulates H264 video anyway.

No, my issue is that, given the animated transitions and the like that feature in my slides, and the need to sync those animations to what is being said, I have to use video rather than a static slide and a recording per slide.

Also, H264 is a pretty impressive video format; unlike MPEG and DivX, it’s a frame by frame encoder, so the fidelity is quite impressive, as long as your source material is clean and clear.