Recently I came across a description of Nvidia's work environment. Powerpoint is forbidden --- everything must be described on a whiteboard. This prompted me to think about teaching together with whiteboards vs. "slides". But let me begin by describing the Audio Engineering course I teach every spring.
Audio Engineering is a very interdisciplinary engineering course. It involves linear systems from EE, electroacoustics from Acoustics (part of physics), hearing (from Psychoacoustics, part biology, psychology and anatomy), circuits (again, part of EE) and then Digital Signal Processing (EE again).
Because I inherited this course, I can't change the prerequisites without a lot of trouble. So, the students are both juniors and seniors. Some have taken the filter design class, others not. Some have taken DSP, others not. So my approach to this course is to motivate the students to take more advanced classes after completion of this course.
For the most part (there are always exceptions...) the students are interested and motivated by music production and listening. So, I try and harness that interest. The text is "Acoustics, 2nd Edition" by Beranek and Mellow. It's a reasonable text but it biases my presentation because of its equation dominant nature --- more on that soon. It only covers Acoustics, specifically Electroacoustics, which means the rest of the course is without a text.
Since it's an engineering course, I want them to build stuff. This past year they built a binaural head (a.k.a. "Dummy head") and a loudspeaker. I want to expand this further. They also create simple DSP effects using matlab functions I provide. Using "canned" matlab functions means they don't get down in the weeds trying to figure out how to implement filtering --- it's not a DSP class and so I don't want to get them bogged down. But this has its problems since they don't really get an idea of various DSP problems including latency and quantization. They do learn about using the DFT, but I have to explain equal tempered scales every year.
So, onto lectures. I begin every lecture with music. I take a historical approach starting with the beginning of the 20th century. So, I like to use Stravinsky as a starting point. Some students know The Rite, others do not. But when I introduce Ionization, that's when the ears really begin to notice. The early tape pieces like Gesang and Les Paul's How High the Moon also make them think. And then we start doing pop music in the 1960s and they are at home. So, why have them listen to music? Two reasons: (1) Many don't know the history of recording (2) To get them to listen for nuance.
Now, here's where it gets problematic: I like to illustrate many of the devices invented over the years. A picture of the internals of a microphone says much more than a simplified diagram. So it would seem that showing "slides" would be best. And that leads into just displaying equations instead of deriving them. I do have lots to say about each equation, of course, but it's not the same as writing it down a line at a a time.
I once took a class poll for a different class (computer architecture): Did they prefer slides or whiteboard? Overwhelmingly whiteboard. And again, there are very nice pictures of old machines. Maybe the pictures aren't relevant, or maybe they're best for home appreciation.
Finally, I am always trying to build intuition. It's shocking how many students can look at an equation and yet not be able to explain how to manipulate it.
In conclusion, every time I teach this class, I am redoing the notes, trying to make the arcane bits clearer and emphasizing (and de-emphasizing) different parts. But maybe this is the year I'll wean myself off the PDF display...
