The slides presented on the PDE's website were made using Beamer with PdfLatex. The files used for generating Lecture 16 with pictures and .sty file included can downloaded using here. Lecture 16 Source (8MB)
At the top of the Lec16.tex file there are two possible headers: \documentclass[handout]{beamer} which produces the handout file Lec16H.pdf if pdflatex Lec16.tex is run with this option. Uncommenting the header \documentclass{beamer} and commenting out the previous header will produce the .pdf file use in the talk recording Lec16.pdf.
To combine three slides of the Lec16H.pdf onto one page for easy printing and creating a compact handout.
pdfnup --a4paper --keepinfo --nup 1x3 --no-tidy --frame true --scale 0.92 --no-landscape --outfile PDEsLec16.pdf Lec16H.pdf
The key point here is to make the video as compact as possible. An estimate for the size of a good audio file only is roughtly 1MB per minute. The size of the .pdf slides obtained in step 1 is about 2.5 MB. The video which keeps track of stepping though the slides and syncs with the voice over averages at about 50MB per 30 min. (30MB audio file and 20MB to keep track of the changing video of pdf slides and mouse motion. ) The video resolution is 1440x1080 pixels. Before recording I made the mouse pointer big and red using the Unity Tweak Tool -> Redglass Cursor -> Large Cursors in ubuntu, so that it would be easy to follow.
The software I used for screen capture is Vokoscreen (Vokoscreen website Link) A windows version does exist. Vokoscreen is free. It is a very tiny program that simply records the screen and audio either from the computers internal microphone or one plugged in. I was initially surprised by the small file sizes but it is due to the encoding that the program uses. The videos it produces are .mkv files (Matroska Video files as in the Russian dolls). Also very importantly the videocodec is libx264 with libmp3lame as the audiocodec which offers state of the art compression in both the audio and video. There is also a setting called frames, the default is (25) this is the number of frames per second that the video records. This is a major spacesaver. For ordinary videos the setting lies between 30 to 60 fps for HD. For slides with mouse movement 20-25 fps should work. As you open up the vokoscreen program, the first tab is a computer screen. There are several options Fullscreen Window and Area. I chose area then a blue square with green dots appear which allows you to adjust the size of the area you are going to record. There is also a Countdown which you can set to give you time to get the pdf viewer or powerpoint into presentation mode after pressing record. Give yourself about 20 seconds. The spanner and cog tab allows you to specify the videopath, ie where you want the videos stored.
Note you should be able to use other screen captures programs to record the talk. Powerpoint may have the functionality built in. The key to getting the small files lies in the original presentation not changing much from slide to slide, low frame rate and the choice of compression algorithm libx264, libmp3lame and .mkv video format.
After I made the videos I wanted to crop the beginning and the end parts which shows me opening the program and putting it into full screen mode. I also wanted to remove the chat when somebody phoned me while recording. I used the ffmpeg commandline library for this. The commands are detailed below.
Because Youtube are very effective at streaming and permits you to download videos. I decided to upload the videos there and only provide links to the videos on the course website. I tested that the total data streaming costs when streaming these videos at HD were no more than the uploaded video screen captures ie 50MB for 30 min. How to Upload to YouTube instructions Link. As an example the youtube link to the first part of Lectures 16's talking slide made with the beamer file can be found here -> Link to Lecture 16 Part 1