Quote:
Originally Posted by samiam  One idea that comes to mind, after reading this thread, is to see if it is possible, with today's computational power, to use something fancier than an all-pass filter for diffusing the sound, such as using a convoluter with a very short impulse response convolution (mainly as an EQ and as a way to smear transients). However, while computing power has gotten a lot better, we may still not be at the point where it's practical to have a few concoluters, albeit short ones, in the core of a reverb algorithm. |
You're describing something similar to the 'Room' algorithm in the PCM96. There are 48 real-world impulse responses that can optionally be dumped into a standard reverb. The impulse responses don't use a full convolver. Instead, they're mathematically reduced into a form that allows some realtime modification. Unlike a convolver, the impulses can easily be rubberbanded in time, EQ'ed and reversed. While this algorithm was originally done with film post in mind, it has a lot of flexibility and potential for future growth. In a very real sense, it's the grandchild of the ambience algorithm that you like in the PCM90.
I have no wish to get into the convolution vs. algorithmic debate (well only a small wish). The best convolved reverbs sound excellent. They're just a little hard to match up with material sometimes. Algorithmic reverbs offer greater flexibility in terms of EQ and other important characteristics. For that reason I don't see them going anywhere.
Ray-tracing is similar. Your example sounded very good, but it didn't sound necessarily superior to a well-done algorithmic or convolved reverb. Even though Moore's law will indeed make processing power a moot point, I still think that --with the exception of early reflections -- ray-tracing may represent serious overkill. But stay with it and prove me wrong, eh?
N.S.