EMUSIC-L Digest Volume 38, Issue 01 This issue's topics: Computer synthesis and computer sound manipulation csound (11 messages) Representations of music signals 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: Fri, 20 Mar 1992 08:30:00 LCL From: DOWRJ@VAX1.COMPUTER-CENTRE.BIRMINGHAM.AC.UK Subject: Computer synthesis and computer sound manipulation Hi - although I have just subscribed to the emusic-l list, and have not been a subscriber of comp.music that long either, I have not heard that much mention of computer synthesis or computer sound manipulation. My problem is as follows: we wish to mount a variety of software of the type mentioned on our IBM3090 computer. Already we have the MIT freebees such as csound (which runs nicely) the phase vocoder, the resampler and so on. Many of these things are still in a stage of porting as they were primarily written for UNIX, NeXT or Mac, and the IBM is a bizzare and difficult beast. I have also had a quick try of porting CMIX from princeton, but this failed, no doubt (I think) because the parser expects ASCII but the IBM is EBDIC(!) What I really want to know is, is there any interesting or even fairly boring software floating about out there that we could have, and mount on our main- frame? What ever happened to the CARL package, for example? Even youir dusty old FORTRAN programs would be useful, as long as they read SHORT files and don't expect anything fancy in the way of headers (IRCAM we can manage ...). Actually FORTRAN is better because they are easier to optimize for the vector processor. Any hints, help, etc. warmly appreciated to: DOWRJ@uk.ac.bham.vax1 (JANET - i.e. this is UK address ordering!) Robert Dow (University of Birmingham, Department of Music) ------------------------------ Date: Mon, 2 Mar 1992 20:09:05 GMT From: jmobberley@VAX1.UMKC.EDU Subject: CSOUND problems w/the NeXT SOS - CSOUND users -- we have been using CSOUND with an IBM compatible using MicroTechnology Unlimited's DS16 converter and software...in order to use the SOUNDIN unit generator and to use GEN 01 for waveform generation it was necessary to create a file called CSOUND.ini which allowed the user to define the variable SFDIR. This declaration allows the CSOUND program to locate pre-existing soundfiles for use by SOUNDIN and Gen 01. We now have a NeXT machine and are running CSOUND on it as well, but have had no success in using a CSOUND.ini file for variable declaration. SFDIR has been identified in both the .cshrc file and in the .profile file which has allowed other functions to take place properly, but we still can't use SOUNDIN. Any help? Respond via E-mail at JMOBBERLEY@VAX1.UMKC.EDU or call 816-235-2968. Thanks; Jim Mobberley, Univ. of Missouri-Kansas City, Conservatory of Music. ------------------------------ Date: Mon, 16 Mar 1992 22:34:00 EDT From: MORGANS@MCMASTER.BITNET Subject: Re: Equipment recommendations This has nothing to do with equipment recommendations but: WHAT THE HECK IS CSOUND??? Thanks, Shawn Morgan MORGANS@sscvax.cis.mcmaster.ca ------------------------------ Date: Thu, 19 Mar 1992 10:41:36 EST From: "Joseph D. McMahon" Subject: CSOUND 92 10:34:00 pm > WHAT THE HECK IS CSOUND??? Csound is a fascinating software package. I'm sure others can give you more details, but the general idea is that you define your sound-generation capabilities by writing programs to do the sonic transformations. These programs are executed, and the resulting stream of numbers is fed to a digital-to-analog converter. Since everything is completely under your control, you can do *anything* that you can manage to program. We've dubbed this a "total control" system. This is not necessarily fast; generally the ratio of compute time to real time is 10 to 1 or greater. --- Joe M. ------------------------------ Date: Thu, 19 Mar 1992 16:01:33 EST From: "Joseph D. McMahon" Subject: Re: CSOUND > > Is this on a Macintosh? Do you then download the generated "samples" to your > sampleplayer to actually make the sound? More info please.. this sounds > interesting. Thanks! > It can be done on a Macintosh, but it is very, very slow there. I believe that the version of csound available for the Mac uses the Mac sounds drivers, with the result that the sound is not exceedingly high quality. However, I think it would be possible to work some thing out so that one of the hard-disk recording boards could be used for playback. A number of people run csound on the NeXT, since it has a very goo built-in digital sound processor. Paul Lansky, among others, has used it - his latest album, "Smalltalk", has at least one piece which uses csound on the NeXT. If I wre going to try csound, I'd see if I could find a nice, fast machine which supported it for computation, and a personal machine of some kind for the output. --- Joe M. ------------------------------ Date: Thu, 19 Mar 1992 16:47:41 -0800 From: Tom Erbe Subject: Re: (Forwarded) Re: CSOUND There is a version of CSound for the Mac. It is available via anonymous ftp from ems.media.mit.edu. Get the .hqx files. It is slower than the NeXT version, but probably due more to processor speed than anything else. CSound produces 16-bit soundfiles on the Mac, suitable for playback through an Audiomedia or Sound Accelerator card or direct digital transfer to DAT. There is also a useful program at the MIT site that translates MIDI files to CSound score files. It runs only on the Mac. Forgot its name, but it is something obvious like "MIDItoCsound". Tom Erbe * Technical Director * Center for Contemporary Music * Mills College tom@mills.edu * Mills College, Oakland, CA 94613 * (510) 430-2191 , ------------------------------ Date: Thu, 19 Mar 1992 18:51:00 GMT From: Eric McCormick <0004775674@MCIMAIL.COM> Subject: CSOUND Is this on a Macintosh? Do you then download the generated "samples" to your sampleplayer to actually make the sound? More info please.. this sounds interesting. Thanks! ------------------------------ Date: Thu, 19 Mar 1992 22:02:31 -0800 From: Tom Erbe Subject: Re: (Forwarded) Re: CSOUND Micro Technology Unlimited sells a version of Csound for the PC. I don't know of anyone who has ported it and has it in the public domain though. Tom Erbe * Technical Director * Center for Contemporary Music * Mills College tom@mills.edu * Mills College, Oakland, CA 94613 * (510) 430-2191 ------------------------------ Date: Thu, 19 Mar 1992 22:27:22 -0700 From: Thunder-Thumbs Subject: Re: (Forwarded) Re: CSOUND Anything like the C-sound package for the IBM? Can anybody direct me to *any* good ibm shareware packages? Anything; sequencing, card programming, midi programming... (no music notation please, I already have that.) Thanks! Curt Siffert siffert@horton ------------------------------ Date: Thu, 19 Mar 1992 22:32:00 GMT From: Eric McCormick <0004775674@MCIMAIL.COM> Subject: (Forwarded) Re: CSOUND Is there source code available for CSOUND? I have access to mainframes where I could run the algorhythms, then transfer the file to the Mac. If source IS available, do you know what language it is currently in? Thanks, Eric McCormick 0004775674@MCIMail.com ------------------------------ Date: Sat, 21 Mar 1992 10:36:00 LCL From: DOWRJ@VAX1.COMPUTER-CENTRE.BIRMINGHAM.AC.UK Subject: csound csound is a comprehensive synthesis and processing package written by Barry Vercoe at the Media Lab, MIT. It is written in more or less portable C (although it does make use of one or two functions which are not ANSI such as open(), close()), and will certainly run well on any UNIX box i.e. SUNs, VAXen, etc. It also runs on the Macintosh under ThinkC 4.0. At the University of Birmingham we have it running on Apollo Workstations, our IBM3090 mainframe, and an old version also comes as part of York University's CDP (Composer's Desktop Project). The latest version of csound not only has the usual modules for processing and synthesis but also has those for doing FOF synthesis (cf. CHANT), and for phase vocoding. There is also the ability to take in standard MIDI files as score files, or (on speedy machines) the ability to run csound in real time, and trigger events from a MIDI stream. csound can output sound-files in AIFF format to be read by Mac type things such as Sound Designer, and also: 8-bit signed character samples, alaw samples, ulaw samples, short integers, long integers, floats, with or without IRCAM (1K) headers. It can output files for the NeXT to play too. The package is available via anonymous ftp to ems.media.mit.edu in the /pub directory. Look at the README's for details of what you need. The manaul is available as postscript. If anyone wants a copy of the IBM3090 version, they could contact me directly. Unfortunately the modified code is in a bit of a mess at the moment as I am re-porting it for another version of C (C370), but I have a running CSOUND MODULE. At Birmingham we run CSOUND on the IBM3090 for complex CPU intensive stuff, and transfer it to an ATARI TT (30MHz) via an ethernet connection. The ATARI runs CDP, and has a soundstreamer. We are getting some more ethernet boards which will allow us to bring the files into Sound Tools running on another ATARI (this will be a massive kludge!). The transfer takes some time, but since something which took a day to compute on a normal ATARI takes 20 minutes on the IBM we're not complaining. Robert Dow, Department of Music, Barber Institute of Fine Arts, Edgbaston, Birmingham, B15 2TT, UK DOWRJ@uk.ac.bham.vax1 (JANET - address in uk order) ------------------------------ Date: Sun, 22 Mar 1992 15:57:00 EDT From: MORGANS@MCMASTER.BITNET Subject: Re: csound This is addressed to Rober J. "dowrj@uk.ac.bham.vax1" I would have written to you directly but for some reason or another, I can't. Anyways, I tried using FTP to get a copy of CSOUND for myself but I can't do that either. Do you think you or someone else reading this message would be able to send me a copy? Thanks, Shawn Morgan MORGANS@SSCVAX.CIS.MCMASTER.CA ------------------------------ Date: Fri, 6 Mar 1992 10:04:42 +0100 From: Giovanni De Poli Subject: Representations of music signals BOOK ANNONNCEMENT Representations of Musical Signals edited by Giovanni De Poli, Aldo Piccialli, and Curtis Roads Published by MIT Press, 1991 ISBN 0-262-04113-8, 478 pages, hardcover, US$ 50.00 Representations of Musical Signals describes a new generation of digital audio and computer music systems made possible by recent advances in digital signal processing theory, hardware design, and programming techniques. It explores new representations of musical signals that can have profund effects on the way musicians conceive of and realize musical ideas. In particular, the book focuses on models that combine time-domain and frequency-domain representations (grains, wavelets, and physical models), visual programming and advanced user interfaces, and that incorporate musical knowledge using artificial intelligence techniques and adaptive neural net- works. The 14 contributions take up issues of how musical signals should be displayed to musicians, engineers, and scientists who want to work with them, how professionals can work with the representations to accomplish musical tasks, how systems can be designed to permit working with multiple views of the same signal, and how representa- tions of musical signals should be organized to promote efficient communication between devices using these signals. Representations of Musical Signals is aimed at the expanding group of musicians, engineers, and scientists who are interested in innovative approache to digital audio and computer music. We expect that this book will be useful in undergraduate and graduate courses in computer music, musical acoustics, and digital audio signal processing. Giovanni De Poli is a member of the faculty of the Department of Electronics and Informatics at the University of Padua. Aldo Piccialli is a member of the faculty of the Department of Physics at the University of Naples. Curtis Roads is a composer and consulting editor of Computer Music Journal. INDEX Contributors xi Preface xiii I TIME-FREQUENCY REPRESENTATIONS OF MUSICAL SIGNALS 1 Overview 3 A. Piccialli 1 Timbre Analysis by Synthesis: Representations, limitations, and Variants for Musical Composition 7 J.C. Risset 2 Application of Time-Frequency and Time-Scale Methods (Wavelet Transforms) to the Analysis, Synthesis, and Transformation of Natural Sounds 45 R. Kronland-Martinet and A. Grossman 3 Analysis, Transformation, and Resynthesis of Musical Sounds with the Help of a Time-Frequency Representation 87 D. Arfib 4 Wavelet Transforms that We Can Play 119 G. Evangelista II GRANULAR REPRESENTATIONS OF MUSICAL SIGNALS 137 Overview 139 G. De Poli 5 Asynchronous Granular Synthesis 143 C. Roads 6 Pitch-Synchronous Granular Synthesis 187 G. De Poli and A. Piccialli III PHYSICAL MODEL REPRESENTATIONS OF MUSICAL SIGNALS 221 Overview 223 G. De Poli 7 The Physical Model: Modeling and Simulating the Instrumental Universe 227 J. Florens and C. Cadoz 8 The Missing Link: Modal Synthesis 269 J.-M. Adrien 9 Synthesizing Singing 299 J. Sundberg IV ARCHITECTURES AND OBJECT REPRESENTATIONS OF MUSICAL SIGNALS 321 Overview 323 C. Roads 10 Music, Signals, and Representations: A Survey 325 uy Garnett 11 An Object-based Representation for Digital Audio Signals 371 12 New Generation Architectures for Music and Sound Processing 391 S. Cavaliere V PARALLEL DISTRIBUTED PROCESSING REPRESENTATIONS OF MUSICAL SIGNALS 413 Overview 415 A. Piccialli 13 Understanding Music Cognition: A Connectionist View 417 C. Lischka 14 Qualitative Aspects of Signal Processing Through Dynamic Neural Networks 447 R. D'Autilia and F. Guerra Name Index 463 Subject Index 467 Available from: MIT Press, 55 Hayward St., Cambridge, MA 02142 USA Tel: (617) 625-8569 Fax: (617) 625-6660 In Europe from: The MIT Press, Ltd., 14 Bloomsbury Square, London WC1A 2LP, U.K. Tel: (071) 404 0712 Fax: (071) 404 0601 ------------------------------ End of the EMUSIC-L Digest ******************************