EMUSIC-L Digest Volume 67, Issue 05 This issue's topics: Physical modeling discussions Physical Modeling (16 messages) Your EMUSIC-L Digest moderator is Joe McMahon . You may subscribe to EMUSIC-L by sending mail to listserv@american.edu with the line "SUB EMUSIC-L your name" as the text. The EMUSIC-L archive is a service of SunSite (sunsite.unc.edu) at the University of North Carolina. ------------------------------------------------------------------------ Date: Mon, 8 Aug 1994 22:46:13 EDT From: Charlie Williamson Subject: Physical Modeling We've been reading a lot about physical modeling lately and I can't help but make the comparison between that and fractals, physical modeling defining organic sounds and fractals describing organic shapes. If the comparison is valid, I wonder if there is any experimental software that lets you fiddle around with physical models like there is for fractals, even if it doesn't work in real time or to very high resolution. Has anybody heard of any that will run on a Mac? Charlie Williamson ------------------------------ Date: Tue, 9 Aug 1994 09:06:34 -0700 From: "Shawn T. Rutledge" Subject: Re: Physical Modeling > If the comparison is valid, I wonder if there is any experimental software > that lets you fiddle around with physical models like there is for fractals, > even if it doesn't work in real time or to very high resolution. Has anybody > heard of any that will run on a Mac? Well, if the physical models can be represented with normal mathematics then I don't see why Csound wouldn't work although it would be rather labor-intensive to write up all the equations. I don't know about doing fractals with it though; it either has a random number generator function or it doesn't. Csound can be gotten as PC binaries or as source code. You could probably compile it for the mac. -- _______ KB7PWD shawn.rutledge@asu.edu (_ | |_) html: http://enuxsa.eas.asu.edu/~rutledge/home.html __) | | \__________________________________________________________________ * fusion * sci fi * rave on * alternative energy * C++ * ham radio * packet * ------------------------------ Date: Tue, 9 Aug 1994 11:45:10 -0500 From: Siberian Khatru Subject: Re: Physical Modeling Excuse the dumb question, but what is "Csound"? Where do I find it? Is it shareware? So many questions.... Jon Southwood jsouthwo@keller.clarke.edu ****************************************************************************** "I once played poker with a deck of tarot cards. I got a full house and four people died."-Steven Wright "If I had only known, I would have been a locksmith."-Albert Einstein "Truly great madness cannot be achieved without significant intelligence." -Henrik Tikkanen ****************************************************************************** ------------------------------ Date: Tue, 9 Aug 1994 12:05:50 -0700 From: "Shawn T. Rutledge" Subject: Re: Physical Modeling > Excuse the dumb question, but what is "Csound"? A package that lets you write code to generate music. Instrument descriptions are stored in .orc files, as sets of equations, and the songs are stored in .sco files which simply say when to turn the instruments on and off, at what pitch, etc. After you have written both files (with a text editor) you use the csound executable to compile them, and the result (minutes to hours later depending on the speed of your machine and the complexity of the music) is a humongous .wav file. There is also a windows shell so you can just point to a file and edit or compile it with a mouse click. > Where do I find it? Is it shareware? ftp.maths.bath.ac.uk, directory is pub/dream. I think it's freeware actually. > So many questions.... There is a Csound mailing list. Mail jpff@maths.bath.ac.uk and ask about getting on it. I'm on it but I forget what I had to do... -- _______ KB7PWD shawn.rutledge@asu.edu (_ | |_) html: http://enuxsa.eas.asu.edu/~rutledge/home.html __) | | \__________________________________________________________________ * RISC * C++ * IEEE * PIC * robotics * quantize the universe * packet * ------------------------------ Date: Tue, 9 Aug 1994 10:50:02 +0000 From: Jan Magnus Barber Subject: physical modelling Is it possible to "hear" a fractal by for example assigning one coordinate in the fractal to time in sound, the other to pitch and the various values of a point to duration/timbre or otherwise? I've been wondering for a while, and bring the question forth in view of the ongoing discussion! (Another question for the fractal-knower: (not related to music) is a 3d fractal possible? How could this be displayed?) Magnus ------------------------------ Date: Tue, 9 Aug 1994 10:59:11 +0000 From: Pete Thorpe Subject: physical modelling Reply to ChasJesus@AOL.COM... Thoughts on physical modelling... Structures incorporating small random irregularities may well be a way to increase the intimate unpredictability of a sound arising from vibrations in such a structure. However, physical modelling, if you're thinking of the new Yamaha equipment (VL1) is not quite what's going on. They use an approach based on good ol' LTR filter technology known as digital waveguides - a ROUGH approximation to an actual physical model (because current hardware is not fast enough for real time simulation of that). As a result, you have quite a predictable output (although some would dispute this, namely the performers!) listen to the strings on the VL1 and you'll hear what I mean. What I'm trying to say is that yes, fractal techniques may be good for simulating the irregularities in an acoustic resonator but that's no good unless we can get real time sound out of such a simulated structure. I'm not sure that there is an analog in sound to the beautiful pictures created by fractal/chaotic algorithms. Totally different experience. You can't see a sound as a whole, and how would you know that the wonderful picture you had on the screen sounded good? Would be interested in further thoughts on this matter, as I believe representation is one of the central problems of computer music. I could go on for ages, but I won't! Hope this is of interest. Read J.Smith III's waveguide material in computer music journal if you're into current practical approaches to physical modelling. I'd be interested to hear any comments. Cheers, Pete Thorpe, University of York, UK. ------------------------------ Date: Tue, 9 Aug 1994 11:11:56 -0700 From: Nate Waddoups Subject: Re: physical modelling On Tue, 9 Aug 1994, Pete Thorpe wrote: > I'm not sure that there is an analog in sound to the beautiful pictures > created by fractal/chaotic algorithms. Totally different experience. You > can't see a sound as a whole, and how would you know that the wonderful > picture you had on the screen sounded good? Would be interested in > further thoughts on this matter, as I believe representation is one of > the central problems of computer music. I could go on for ages, but I won't! The algorithmic-composition mailing list just had a brief, but interesting, flurry of email about fractals and music... the dicussion there was more about the idea of note-level fractal algorithms than waveform-level, though. I have a suspicion that fractal algorithms might be useful in generating synthy sounds, sort of along the lines of older analog devices, but it seems to me that if the goal is to model real acoustic instruments, fractals might well be more trouble (in terms of coming up with the right parameters, as well as the computing power) than they would be worth (in terms of producing realistic sound). That's just a gut-level intuition, though... If anyone's curious about how the algo-comp discussion went... pretty much the only thing everyone agreed on at the end of the thread was "well, that could be pretty interesting." as an aside: I keep getting duplicate (or near-duplicate) messages in my mailbox from synth-l and emusic-l. It seems silly to post the same question to both lists, since a) the topics of the lists don't have THAT much overlap, and b) it seems like there is a large amount of overlap in the subscribers - there's no point in asking the same people twice. Is it just me, or does anyone else feel this way too? +-------------------+------------------------------------------------------- | natew@halcyon.com | You write Windows code? Mac code? OS/2 PM code? +-------------------+ Want a great job in the pacific northwest? Send me email, we'll talk... ------------------------------ Date: Tue, 9 Aug 1994 11:37:37 -0700 From: Nate Waddoups Subject: Re: physical modelling On Tue, 9 Aug 1994, Joe McMahon wrote: > Anyway, the idea is that you seed a linear automaton with a bit map, run it > for a bunch of generations, and then convert the bitmaps into numbers and > feed them to your handy D-to-A converter. The article simply indicated that > this could be done -- I was hoping I could get a more visceral description > of how it sounded from the author, but was unsuccessful. 'convert the bitmaps into numbers?' How so? Inverse Fourier? Interesting idea. > significantly different bodies of subscribers on the two lists. I think > that there may be more "lurkers" on both lists that are a much more > disjoint set than those of us who are perpetually flappin' our jowls. :-) You have a point there. +-------------------+------------------------------------------------------- | natew@halcyon.com | You write Windows code? Mac code? OS/2 PM code? +-------------------+ Want a great job in the pacific northwest? Send me email, we'll talk... ------------------------------ Date: Tue, 9 Aug 1994 11:52:57 -0700 From: "Shawn T. Rutledge" Subject: Re: physical modelling > The algorithmic-composition mailing list just had a brief, but > interesting, flurry of email about fractals and music... the dicussion > there was more about the idea of note-level fractal algorithms than > waveform-level, though. How do I get on that list? -- _______ KB7PWD shawn.rutledge@asu.edu (_ | |_) html: http://enuxsa.eas.asu.edu/~rutledge/home.html __) | | \__________________________________________________________________ * RISC * C++ * IEEE * PIC * robotics * quantize the universe * packet * ------------------------------ Date: Tue, 9 Aug 1994 14:27:54 -0500 From: Joe McMahon Subject: Re: physical modelling >I have a suspicion that fractal algorithms might be useful in generating >synthy sounds, sort of along the lines of older analog devices, but it >seems to me that if the goal is to model real acoustic instruments, >fractals might well be more trouble (in terms of coming up with the right >parameters, as well as the computing power) than they would be worth (in >terms of producing realistic sound). That's just a gut-level intuition, >though... There was a fascinating article in CMJ about 2 or 3 years ago about the use of linear cellular automata to synthesize timbres. I tried to get the author to post a synopsis of any further work, but never got a reply. Perhaps EMUSIC-L is not a prestigious enough forum... Anyway, the idea is that you seed a linear automaton with a bit map, run it for a bunch of generations, and then convert the bitmaps into numbers and feed them to your handy D-to-A converter. The article simply indicated that this could be done -- I was hoping I could get a more visceral description of how it sounded from the author, but was unsuccessful. --- Joe M. >as an aside: I keep getting duplicate (or near-duplicate) messages in my >mailbox from synth-l and emusic-l. It seems silly to post the same >question to both lists, since a) the topics of the lists don't have THAT >much overlap, and b) it seems like there is a large amount of overlap in >the subscribers - there's no point in asking the same people twice. > >Is it just me, or does anyone else feel this way too? The descriptions of the lists tend to make one think perhaps that there are significantly different bodies of subscribers on the two lists. I think that there may be more "lurkers" on both lists that are a much more disjoint set than those of us who are perpetually flappin' our jowls. :-) --- Joe M. -- "Now hold it down for an hour. Okay, now hold it down until you get a recording contract from Windham Hill." (MST3K) ------------------------------ Date: Tue, 9 Aug 1994 17:05:37 -0500 From: Joe McMahon Subject: Re: physical modelling > >'convert the bitmaps into numbers?' How so? Inverse Fourier? Directly and literally. A sixteen-bit segment of generation 1 becomes a sample, and so on until you get tired or run out of samples or sample memory. Because it's done this way, you can use some of the automata that yield chaotic configurations, or those which run for a number of generations then die out, or which land in a steady state. If anyone's interested, I can nutshell a nice document I've got here on the shelf about 1-dimensional cellular automata. --- Joe M. -- "Now hold it down for an hour. Okay, now hold it down until you get a recording contract from Windham Hill." (MST3K) ------------------------------ Date: Tue, 9 Aug 1994 22:36:25 -0700 From: Nate Waddoups Subject: Re: physical modelling On Tue, 9 Aug 1994, Shawn T. Rutledge wrote: > > The algorithmic-composition mailing list just had a brief, but > > interesting, flurry of email about fractals and music... the dicussion > > there was more about the idea of note-level fractal algorithms than > > waveform-level, though. > > How do I get on that list? It's either listserv@heinous.isca.uiowa.edu, or 'majordomo' at the same site. Send for HELP at that address... I wish I could be more specific, but I haven't bothered with it in months. +-------------------+------------------------------------------------------- | natew@halcyon.com | You write Windows code? Mac code? OS/2 PM code? +-------------------+ Want a great job in the pacific northwest? Send me email, we'll talk... ------------------------------ Date: Wed, 10 Aug 1994 23:08:06 +0200 From: Paul van Oss Subject: Re: physical modelling Reply to Pete Thorpe >Reply to ChasJesus@AOL.COM... > >Thoughts on physical modelling... > >... >... fractal techniques may be good for >simulating the irregularities in an acoustic resonator but that's no good >unless we can get real time sound out of such a simulated structure. > >I'm not sure that there is an analog in sound to the beautiful pictures >created by fractal/chaotic algorithms. Totally different experience. You >can't see a sound as a whole, and how would you know that the wonderful >picture you had on the screen sounded good? Would be interested in >further thoughts on this matter, as I believe representation is one of >the central problems of computer music.... I report here of a TV-broadcast, that I thought was so interesting that I made some notes of it. It is about the visualisation of sounds. What stroke me was that the sounds we rather uninteresting, but the visual patterns were very complex and on another level as emusic people are thinking. The documentary was of a Swiss called Hans Jenny (died in 1972) who lived in Dornach, was a follower of Rudolf Steiner (Anthrosophie), a medical doctor, painter and researcher of what he called 'Klangfiguren' (figures of sound). He called his study Cymatica (sorry no greek available). For his research he used sand, pollen, a plate and a speaker... He discovered most interesting figures when he put the sand or pollen on the plate. Some figures looked like flowers (daisy-like), some like shells. By change of frequency it seems as if the material is reorganizing itself. It looked pretty fractal to me. The structure of the figure keeps constant as the sound stays constant, but the particles seem to move around. He also experimented with metal particles, sound and magnetic fields to get 3-D figures. A mixture of galactic Supernova's, Black Holes and skyscrapers in a process of building (or destruction) is my best possible description. There also is an instrument developed, a tonoscope, a kind of saxophone with no holes and at the end 2 membranes with some fine sand in between. They use it (in Dornach) in primary school to let the children see the structure of sound. You can see the structure of the vowel 'a' and the difference with the vowel 'e' for example. Hans Jenny also had philosophical thoughts with this, that I cannot quickly reproduce. Maybe best with a statement of Goethe: Who has art and science, Has religion; Who hasn't both of these, He has religion. Unfortunately I have no further information. The best try would be the Goetheanum in Dornach, Switserland. Groeten, Paul van Oss, Amsterdam ------------------------------ Date: Thu, 11 Aug 1994 08:50:33 -0500 From: Brian Good Subject: Re: physical modelling Joe M. write: > If anyone's interested, I can nutshell a nice document I've got here on the > shelf about 1-dimensional cellular automata. I'd be interested, unless it's Stephen Wolfram's article from Reviews of Modern Physics, ca 1983; got that one already. brian good bgood@sundagger.lerc.nasa.gov ------------------------------ Date: Thu, 11 Aug 1994 08:37:35 -0700 From: "Shawn T. Rutledge" Subject: Re: physical modelling > >> If anyone's interested, I can nutshell a nice document I've got here on the > >> shelf about 1-dimensional cellular automata. > > > >I'd be interested, unless it's Stephen Wolfram's article from > >Reviews of Modern Physics, ca 1983; got that one already. > > > >brian good > >bgood@sundagger.lerc.nasa.gov > > It's a series of things, actually, from Harlod V. McIntosh at the > Universidad Autonoma de Puebla. He emphasizes the use of de Bruijn matrices > to understand and predict behavior in an automaton. Includes Wolfram's work > among others. > > What I'll try to do is put together a good overview for the > non-mathematician who wants to know whether or not to be interested. Sounds interesting - I'm looking forward to it. -- _______ KB7PWD shawn.rutledge@asu.edu (_ | |_) html: http://enuxsa.eas.asu.edu/~rutledge/home.html __) | | \__________________________________________________________________ * virtual reality * sci fi * freedom of information * alternative energy * ------------------------------ Date: Thu, 11 Aug 1994 11:24:05 -0500 From: Joe McMahon Subject: Re: physical modelling >Joe M. write: > >> If anyone's interested, I can nutshell a nice document I've got here on the >> shelf about 1-dimensional cellular automata. > >I'd be interested, unless it's Stephen Wolfram's article from >Reviews of Modern Physics, ca 1983; got that one already. > >brian good >bgood@sundagger.lerc.nasa.gov It's a series of things, actually, from Harlod V. McIntosh at the Universidad Autonoma de Puebla. He emphasizes the use of de Bruijn matrices to understand and predict behavior in an automaton. Includes Wolfram's work among others. What I'll try to do is put together a good overview for the non-mathematician who wants to know whether or not to be interested. --- Joe M. -- "Somebody's got his finger on a DX7." (MST3K - "Being from Another Planet") ------------------------------ End of the EMUSIC-L Digest ******************************