quantization

The quantization is lossy but for large enough K I do not think it would be noticeable.

One option (albeit a performance-degrading one) would be to implement the same quantization in a local proxy.

The solution there was quantization of energy packets and a statistical treatment of the fewer amount of packets that come through.

With the dithering you have a statistical guarantee that the quantization errors average out.

However, that does not invalidate my points about the other two aspects of quantum physics -- fluctuations and quantization -- does it?

If you were talking about EE DSP stuff, you'd call it quantization noise.

I'd expect that there'd be a fairly large number of tokens so that quantization wouldn't come into play too much.

The median cut quantization with a few tweaks (using the the variance typically - see the post for more) actually gives very good results.

I've always seen equal temperament similarly to digitizing or any other form quantization.

The biggest wins I've found (and implemented in libimagequant) were: • Voronoi iteration after generating palette (like a final step in LBG quantization).

There's no good reason this couldn't be the case, and quantization of fundamental units like Energy-time/Momentum-distance is certainly something that a programmer might reasonably do for a simulation.

So it's using a simple Median Cut quantization...

[1] Xiaolin Wu, "Color Quantization by Dynamic Programming and Principal Analysis", ACM Transactions on Graphics 11(4): 348-372, 1992.

That's a lot of perceptible quantization error when, it turns out, our pitch, timbre and spacial-placement sensory gear is sensitive down to the microsecond level (google "microsecond phase sensitivity" on that).