This is a multi-part message in MIME format. ---------------------- multipart/mixed attachment --------------D0D6E93DCF950FD5F3050426 adkpiano wrote: > Hi ListI'm having a late night browse tonight and found a page perhaps > worth a look to all those considering putting together the ultimate > ETD. I guess any info is at least worth a look.Try "Musical Sound > Modelling With Sinusoids Plus Noise" - WOW !!, Pass the sleeping pills > darling....they're in my anorak.A prize will be awarded to those able > to summarise in 10 words or less. Isn't "Phase spectra Computation" > something to do with pregnancy > testing. http://www.iua.upf.es/~sms/docs/msm/ * The Deterministic plus Stochastic Model * General Diagram of the Analysis/Synthesis Process * Magnitude and Phase Spectra Computation * Peak Detection * Pitch Detection * Peak Continuation * Stochastic Analysis * Deterministic Subtraction * Stochastic Approximation * Representation of the Analysis Data * Modifications of the Analysis Data * Deterministic Synthesis * Stochastic Synthesis Alan King Mpta >Piano Tech>Scotland > > adkpiano@breathemail.net Hi Alan. Finnally got time to look at this a bit closer. The basic jist of all this is "how to synthesize sounds" so that they closely represent the origionals. This is accomplished by picking apart what a musical sound is, sampling an instrument and subjecting it to analysis based on the components arrived at, and then using this information to re-create the sound synthetically. Lots of err.. shall we say "beyond high school math" here. But you dont need that to understand the article in general terms. I am not sure how this all applies to the posts origional questions, but the article does point out the need for a long sustained and clear "sound" for the purposes of sampling in regards to accurate measuring of partials. This would support what John Coleman and others have written to me about the drawbacks of my (and obvioiusly others before me) idea for a "beat reader" based on measureing of partials. I still am in doubt as to how much this problem is a matter of techology / techniques. Another approach to "beat reading" might be to amplify a say 20 cent frequency range (as exclusively as possible) and then electronically measure and count the amplitude variations beats create. For example say one wanted to measure the 4:2 of c3 c4. Since one can expect this to be around 522 hz we could amplify the range of 510 to 530. This would bring "forward" the beating in this range between c3 and c4. Once thats done we could then measure the difference between the two strings, or the amount of beating by counting the amount of "peaks" per second in the "volume" of the sound. Might be worth a try to see if it could be made functional, and usefull. Richard Brekne --------------D0D6E93DCF950FD5F3050426 An HTML attachment was scrubbed... URL: https://www.moypiano.com/ptg/pianotech.php/attachments/78/0d/a9/37/attachment.htm --------------D0D6E93DCF950FD5F3050426-- ---------------------- multipart/mixed attachment A non-text attachment was scrubbed... Name: richardb.vcf Type: text/x-vcard Size: 206 bytes Desc: Card for Richard Brekne Url : https://www.moypiano.com/ptg/pianotech.php/attachments/b3/b2/fe/ad/richardb.vcf ---------------------- multipart/mixed attachment--
This PTG archive page provided courtesy of Moy Piano Service, LLC