Generator embedding optimization: Difference between revisions

Dave Keenan (talk | contribs)
Convert to generators: Changed "generator preimage transversal" to "generator detempering".
Cmloegcmluin (talk | contribs)
target set → target-interval set
Line 8,456: Line 8,456:
When computing an all-interval tuning where the dual norm power is <math>∞</math>, we use a variation on the method we used for ordinary tunings when the optimization power was <math>∞</math>.  
When computing an all-interval tuning where the dual norm power is <math>∞</math>, we use a variation on the method we used for ordinary tunings when the optimization power was <math>∞</math>.  


In this case, our optimization power is also still <math>∞</math>. That is to say, that in this case we're doing the same computation we would have been doing if we had a finite target set, but now we're doing it as if the primes alone were our target set.  
In this case, our optimization power is also still <math>∞</math>. That is to say, that in this case we're doing the same computation we would have been doing if we had a finite target-interval set, but now we're doing it as if the primes alone were our target-interval set.  


Let's get the minimax-S tuning of meantone. With three proxy target-intervals and two generators, we end up with four constraint matrices:
Let's get the minimax-S tuning of meantone. With three proxy target-intervals and two generators, we end up with four constraint matrices: