Woodworking is fun, but let me tell you: if you work for hours at something and then test fit it and find that it doesn't fit... cue the sad trombone music.
In the interest of answering my own questions - specifically "Can the joinery like I suggested work? and "Is plywood or MDF preferred?" I decided to to see if I could build a mock-up of the speaker cabinet design I posted the other day.
I had both MDF and plywood scraps sufficient to make a small box, so I went for the plywood - figuring I'd be happier without the MDF dust. Having worked through most of the problems in constructing this unusual box, I feel confident that plywood is a very good choice.
So, the first thing to figure out was how to make the notch I wanted to join the slanted sides with the baffle. I've developed a couple tricks to doing it, but the heart of the technique is cutting partway through the panel with the piece laying flat on the table, then turning the piece vertical and running it over the same blade. To do that you should use a jig to hold the piece vertical and keep it tight against the fence. I built one last night and used it, but then it broke before I figured out the tricks to getting the notch sized right. It is all about where you put the fence, and there's no good way to measure. Anyway. I was able to get them cut pretty nicely, eventually:
The second thing to work through is the compound miter needed to fit the slanted panels to fit together. This is where I think I ran into trouble, but I think I can come up with a way to deal with it. I may need a tip or two.
I built this mock up without a back - I probably would have added one if it had come out better, and I will need to try again to figure out how to do it right, but I am not there yet.
Here's the pieces I made. The small square is the baffle in this scenario. You can see the rabbets I cut in the two slanted panels to make up for the fact that the baffle is recessed within the slanted panels, but mounts on top of the flat panels (you'll see better how they fit together in a later picture I think). I have to do that one kerf at a time - this is a simple operation I run here. I made all of these with only a table saw, using the built in bevel gauge, the miter gauge to feed through the pieces, and the fence - that's it. Ultimately, I didn't even measure anything (though I did on a few of the first iterations). Turns out patience and careful use of geometry are better allies than measurements, at least so far for me.
As I began to fit the pieces together, it seemed like they were going to fit nicely (of course I had been test fitting everything as I went, but not all of it together). Three sides and the baffle look pretty good.
But when you go to include the fourth side, the compound miter shows itself to be wrong...
So I ask you - is this reasonable proof of concept?