As this is the fourth and final year of my degree (and my Masters year, no less), I am finally getting to do a research-project-type-thing. I add the ‘type-thing’ there because I’m not sure that it really qualifies as research as such – it’s only two modules, after all, in a subject where research necessitates a pretty comprehensive knowledge of everything that has gone before and requires a certain confidence/ genius, delete as appropriate. I have neither of these things, and have anything but a comprehensive knowledge of electrical engineering.
Nonetheless, it is wonderful to finally be doing coursework again! To have a sustained project, with a long term focus and outcome is something that I’ve really missed, and one of the reasons why I enjoyed my Teaching module so much last year.
My topic, for those who are interested, is ‘The Mathematics Inside Digital Synthesisers’. I didn’t choose it myself exactly – well, I did, but I picked it out from a list of about thirty such topics rather than picking something from thin air in the manner that arts people particularly seem to choose their dissertations. Still, this is good fun. It’s basically computational Maths with a very specific application – and is not an area that huge numbers of mathematicians know anything about. If you read around the textbooks or articles on t’net, the vast majority have been written by and for engineers – namely, here is the Maths, use it. I, on the other hand, am trying (with the help of my rather lovely albeit slightly time-warped supervisor) to understand the Maths, and hopefully next term get on to constructing some of my own.
If I ever manage to get on top of the written report (haha, 50 pages in LaTeX), I’d really like to have a play-around with the synthesiser that my supervisor coded himself from scratch and include some sound samples. Nothing too complicated, mind. Getting synthesised sounds to sound realistic is incredibly difficult, much more so than one might imagine, and even those who understand the Maths basically play around with it by ear. But what I’d like to do is the simple stuff, directly related to the calculations I’ve been doing. Here is a sine wave at 440kHz. Here is what it sounds like with a low pass filter. Here is what is sounds like with different envelopes, and the introduction of a resonance paramater. Here is what it sounds like with time-variant resonance (- that’s my new, never before been done angle to the project. Even part way through, my instinctive reaction is “Help I can’t do this The chances of me screwing up are so huge that maybe I shouldn’t even try Why am I doing this” etc. etc. etc.).
Daunting it may be, but it is nonetheless incredibly interesting, or at least I think so :-)
On a not entirely unrelated note, have a graph! Needless to say, it’s not quite the one I intended to draw (but is very cool nonetheless, si?)

Recent Comments