Posted by Rick Denney on July 16, 2001 at 09:29:04:
In Reply to: Re: Re: Re: Re: Re: Re: Yet another change to The Tuba Sound posted by Carl on July 14, 2001 at 15:05:41:
And, of course, the danger is that in doing so they achieve optimality in those few parameters quantified by the software only by de-optimizing characteristics not measured.
A simple principle from operations research (the study of optimality): You must have an objective function to optimize. What constitutes a good tuba sound? We know it when we hear it, but as my simple tests have pointed out, we know very little of how to describe it objectively.
As with previous discussions that centered around intonation as a primary objective, we must consider that all the qualities important in a tuba may well present a non-linear model of enormous complexity, and that incremental trial and error (the method currently in use) actually achieves better results. Herr Meinl is wise not to abandon his traditions.
And I still wonder how a pickup at the mouthpiece can measure projection. Would a recording tuba sound the same to this system as an otherwise identical upright-bell tuba? We know they sound different to the audience.
Rick "who uses computer modeling too much to ever trust it" Denney