>>Agreed! I did the same sort of spectral analysis thing and it took a long time! On the other hand, I was so pleased with my quasi- imitative sound, not just as a good brass sound, but as a great synthesizer sound. And this is really my point. There is a lot of untapped potential in the modular analog synth. I think the module I'm talking about could be very useful.<< Yeah, there's plenty of untapped potential in a modular synth even beyond filtering. Unfortunately today's synth climate isn't one where people want to put in the time or to learn to tap into that potential. The pioneering spirit died out decades ago. Having gotten part of the way into a string synth I am writing, I've learned how to do some things I didn't know before, like program a multi-band filter bank with the bands shaped anyway I want. I thought about a product that would be the be-all and end-all of filter banks, with something like a 56 bands and all kinds of advanced features. I don't want to list them all to give others ideas, but one would be to analyze incoming sounds for formant info and set the filterbank to a good match. That would greatly help with a starting point. It could also pop into EQ mode or into narrow bandpass comb filter mode in addition to other modes of operation. >>How about sweeping it with an EG?<< The 128 band formant filter in the K5000 additive synth can sweep around like a phaser. So there are special effects uses like that. >>Or maybe some interpid folks could come up with settings to share with the user community that would be also useful for imitative sounds, based on their hours of spectral analysis? Or maybe a setting that imitates the MoogCE String Filter?<< Actually below is a demo I did where you can hear the difference between different generic string filter imitations (similar to the Moog) done by placing bandpass filters in various spacings and patterns followed by a couple of demos where I started to modifiy one of them to match a real violin. I'm still not finished with the violin ones, but the confined boxy sound is gone when you start to randomonize the freqs and amplitudes of the bands to more closely match a real instrument. Note I have never posted this following demo before as it was slated for the trash bin, but it's kind of intesting hearing some subtlely different timbres using different filterbank configurations on the Nord modular. It starts off with no filter bank at the beginning. All demos below are played in realtime with no multitracking and just me improvising random music. http://home.att.net/~elhardt3/String_Filter_Experiments.mp3 Here are a couple Nord synthesized violin demos which I've posted before, but I can't remember if I posted to this list so they may be new to you. The second one is Nord violin over Roland rompler strings. http://home.att.net/~elhardt4/Nord_Violin_Maneuvers.mp3 http://home.att.net/~elhardt4/Nord_Violin_Romantic.mp3 -Elhardt
Message
Re: [motm] Re: Imitative Synthesis and Implications for Hardware
2007-04-25 by Kenneth Elhardt
Attachments
- No local attachments were found for this message.