EMUSIC-L Digest Volume 7, Number 1 Today's Topics: EMUSIC-L Tape Progress (3 messages) Roland D-50 (2 messages) MIDIMan vs. SMPTE (2 messages) Roland JX-Series Equivalent? ICMC Paper on Timbre Perception and Orchestration Summer Courses at the University of Padua 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: Sat, 1 Jul 89 11:23:05 EDT From: Cecilia Tan Subject: Tape progress It's Cecilia here, on her last day with her own mainframe account. Administrivia re: the tape-- if you need to mail me, please mail my friend Nick at ST501220 at BROWNVM. I'll also continue monitoring e-music-l through him. Progress report: The first submission is here! Joe McMahon has sent me two pieces. They arrived yesterday and I haven't had a chance to hear them. WHERE ARE ALL THE OTHERS?!?!?! Get your stuff in, people. You have one month. That's not very long. Refer to my previous message about mailing details! Also, has anyone got any suggestions for a title? "The E-music Tape" doesn't really ring. -ctan ------------------------------ Date: Sun, 2 Jul 89 17:57:50 EDT From: Cecilia Tan Subject: More tape stuff A second contribution has come in. I haven't had a chance to listen to it either. This one from John McIntyre in Michigan. Now I'm getting really excited and look forward to some more pieces to arrive soon. It's all up to you folks. (By the way John, the ten minute "limit" isn't strict, really just a guideline) -cecilia ------------------------------ Date: Fri, 7 Jul 89 11:59:22 EDT From: Cecilia Tan Subject: Deadline approaching Hello, it's me again. Just a reminder that I need all your submissions for the emusic-l compiliation tape by August 1st. I don't mean in the mail on that day, I mean in my mailbox. Thus far we have accumulated about 20- 25 minutes of music, some of which is mine. (I spent about four hours working on a piece last night which will be about 4 mn. long) So crank up your sequencers, your LFOs and fm tones and make some NOISE! It would be nice to get a note once in a while, too, about how things are going... -cecilia (ps. please cc: any mail to me to st501220 at brownvm also, since the death of my account is imminent) End of EMUSIC-L Digest *********************** Date: Thu, 6 Jul 89 15:37:09 BST From: Nick Rothwell Subject: Roland D-50 27 Jun 89 07:23:00 EST <1411.8906271238@lfcs.ed.ac.uk >Well, I don't like it much. I guess I'm a musical clod, but my tastes run >to the classical and new age, and the D-50 does not seem to have many >patches amenable to those kinds of music. The keyboard operates >beautifully, and the programmer is very nice, too. The bottom line is, >though, I like thick, rich sounds and the LA synthesis just does not >produce them as well as my Juno 2. I think you're judging it a little too quickly. The D-50 isn't quite as rich as a true analog, I agree, but I think the internal effects compensate for this, and it can do some beautiful orchestral and atmospheric timbres. I have some lovely Jupiter Strings patches which I use all the time. I'll try to get my D-50 patch banks (approx 20 of them) online as soon as I'm over my LA->UK Jetlag... BinHex or uuencode? Anybody want comments about the Oberheim Cyclone "MIDI Performance Arpeggiator"? Nick. -- Nick Rothwell, Laboratory for Foundations of Computer Science, Edinburgh. nick@lfcs.ed.ac.uk !mcvax!ukc!lfcs!nick ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ Fais que ton reve soit plus long que la nuit. ------------------------------ Date: Thu, 6 Jul 89 12:35:44 EDT From: Eric Harnden Subject: d50 upload uu'ed, and either arc'ed or zipped, please.i don't think that zoo is in quite wide enough use yet. don't forget the info file. when i get your stuff, i'll get it in the archive along with a bunch of patrick's d10 stuff that i've been hanging on to. Eric Harnden (Ronin) The American University Physics Dept. (202) 885-2758 ------------------------------ Date: Thu, 6 Jul 89 16:11:23 BST From: Nick Rothwell Subject: smpte etc. 28 Jun 89 10:14:00 EST <4826.8906281822@lfcs.ed.ac.uk >So what the heck does the Midiman do? Basically, it turns any ordinary >cassette player into a sequencer, or more acurately, a sequence player. >It records MIDI data from a real sequencer (your home computer) directly >onto tape. This is a very useful facility. It means that, for example, you can use a computer-based sequencer, dump a song onto tape in MIDI form, and then run a voice editor on the computer to alter some of the sounds while the song is playing back. Of course, if MIDI systems on computers were TRULY integrated there would be no need for this... I don't trust cassette tape for anything really critical like live sequences (although, perhaps Midiman has error correction...?), so I'm going to investigate the Elka CR-99. Essentially the same functionality, but using 3.5" floppies, which gives you random access between songs (although not within songs, although you can "fast-wind"). It has tempo control and overdubbing, all in the MIDI domain, of course. It only has a single MIDI in, with the two MIDI-outs being copies; this is a shame, because I'm currently using about 20 MIDI channels from my Mac. I'm still finding my way around the Cyclone, but I expect that it will sync to MIDI clock messages without too much hassle. I've already had it running the other way, with Performer (version 3.0 - anybody want a review?) slaved to the Cyclone. Our live set-up will probably involve sequences and clock info playing back from a CR-99 or something like it, and live play from keyboards, possibly into the Cyclone which is sync'ed to the sequencer clock. has anybody heard of the Alesis Datadisk? Sounds like another of the same to me...? Nick. ------------------------------ Date: Thu, 6 Jul 89 11:33:00 EST From: JWK%OPUS@MCOIARC Subject: RE: smpte etc. Sure Nick, I'd like a review of Performer 3.0. I haven't upgraded yet... should I? -j. ------------------------------ Date: Thu, 6 Jul 89 17:00:23 EDT From: Dan Bornstein Subject: Module replacement for a Roland JX3P? Does anyone know if there is a tone module (i.e. tone generator, no keyboard) that has the same patch scheme as a Roland JX3P (or JX8P too, I believe they're the same). I have a JX3P and want to get a better keyboard, but I'm attached to the sounds of the synth. So, I'm looking for a module I can buy that will generate 3P-style sounds, freeing me to get a new keyboard. Any suggestions? Thanks! -dan Bitnet: danfuzz@brownvm Internet: danfuzz@brownvm.brown.edu AppleTalk: Find me a long enough cable and I'll see what we can do. ------------------------------ Date: Sun, 2 Jul 89 00:42:54 CDT From: Greg Sandell Subject: ICMC Paper on Timbre Perception and Orchestration Enclosed is a proposal for a paper I will be reading at the 1989 meeting of the International Computer Music Conference in Columbus, Ohio (Nov. 2-5). It's on some work I have been doing for the past year in timbre perception and which will form the subject of my disseration in Music Theory at Northwestern University. I have presented the proposal exactly as the ICMC got it, except for the examples, which were impossible to transmit over the net. This may make reading this a bit odd (or perhaps an inducment to come to the meeting and see them there!), but I think that my main points can be understood anyway without the enhancements of the visual examples. There is one exception: figure 1 has been transcribed and included, at the end. Also, there is a brief description of all the other examples (at the end). Greg Sandell greg@music.nwu.edu Northwestern Computer Music SANDELL@NUACC Francis Searle Bldg uunet!oddjob!nucsrl!numusic!greg 2299 Sheridan Rd Evanston, IL 60208 (312) 491-3895 Timbral Dimensions in Orchestration ----------------------------------- Gregory J. Sandell (Northwestern University) International Computer Music Conference November 1989, Columbus, Ohio The study of musical timbre perception has been a complicated area of study because of the difficulty maintaining control over timbre's multidimensionality in genuinely musical contexts. Most of the meaningful knowledge we possess about the fundamentals of timbre and the way it is perceived have been through the methods of psychophysics, with its bias for sounds in isolation; in the traditional division of musical disciplines however, timbre falls into the department of orchestration. In fact, orchestration is frequently the musical context in which we actually confront timbre; even for many musicians today (especially those who create their music in MIDI environments), greater creative efforts are directed towards expressive combinations of fixed resources (presets) than the composition of sounds from scratch. The need for a systematic approach to orchestration is no less great today than it was in 1820. The current psychoacoustic literature provides some hints for an initial inquiry into orchestration and perception. John Grey's (1975) three-dimensional paradigm (spectrum, harmonic synchrony and attack), a successful model for explaining timbral similarity, is a useful template for approaching orchestration. How, for example, do the three dimensions govern whther two timbres are perceived as *blending* together? Blend is an appealing point of departure, not only because it can be explored as a psychological relation between two sounds, but also because it has been a central topic of traditional orchestration. Rimsky-Korsakov's treatise, for example, contains many statements such as "strings do not blend so well with the brass, and when the two groups are placed side-by-side, each is heard too distinctly," or (on the French Horn), "the middle notes resemble those of the bassoon and the two instruments blend well together." How, then, will spectrum, attack and harmonic synchrony relate to these phenomena? Fifteen of the instruments used in Grey(1975) (all except one of the oboes) were played in simultaneous presentation to a group of eight listeners, all musicians, in a study of blend. The tones were generated with CSOUND software, at 44.1kHz, from the original *line-segment approximations*. The original durations (ca. 330 msec) and relative loudnesses of the tones (which had been adjusted for perceptual equality by Grey's subjects) were retained. Retaining Grey's pitch levels, however, produced undesirable results in simultaneous presentation (beating), as did starting tones at an objective time-zero (splattered attacks). Accordingly the tones were adjusted to a more objective unitary Eb4 by using a pitch detection program from CARL, and the relative start-times of every pair of tones was adjusted to attain equal *Perceptual Attack Time* (by using values given in Gordon, 1984). Subjects used a ten-point scale to rate blend, defined as a continuum between "fused" (10) and "separated" (1). Each subject made judgments for all 120 possible pairings (which included same- instrument pairings) exactly four times. Because all within-subject and between-subject correlations of data were significant (p < .05), the results were averaged into one pool of data. Perhaps the best way to explore the results is to step through a selection of trials and observe their ratings. (Fig. 1) This chart shows the pairing which received the highest average blend judgment at the top, the lowest average blend judgment at the bottom, and selections of pairs which received ratings along a continuum in between these two extremes. (Play tape of 13 trials.) A small subset of this group of 13 is shown in various represenations in figures 2-7. Two facts may be immediately notes: 1) there are many "dark" instruments on the blended end, and "bright" instruments on the other, and 2) instruments with quick, clean attacks are contraposed, in the same way, to those with longer, noisier attacks. The 3-D amplitude plots (figs. 2, 4 and 6) show lower blend judgments with widened spectral content, while the waveforms (figs. 3 and 7) show lower blend judgments for trials with attack noise present. Fig. 5, a time- frequency plot, shows the attack noise, as well as attack inhamonicity, from a different perspective. A general characterization of what happened is shown in figs. 8 and 9. In fig. 8, the brightness of instruments is shown on the horizontal dimension, with dark of the left, and bright on the right; despite a few outliers, the two dimensions are correlated (r = -.655, p < .05). A single point on this graph, say, the muted cello, represents the average rating for all pairs of instruments which contained the muted cello. More specificallly, the role spectrum plays in this data is that if two instruments are dark, they get a very high blend rating; if one of the two is replaced by one which is slightly brighter, then the blend judgment goes down. Stated as a rule of thumb, if one of the instruments in a pair is dark, the pair will blend better than if both instruments are bright. Similarly, if one of the instruments in the pair has a quick, clean attack, the pair will blend better than cases where both have a slow, noisy attack; fig. 9 shows the affect of attack characteristics in a way analogous to fig. 8. Here the correlation is not as strong but still significant (r = -.549, p < .05). Grey's other dimension, harmonic synchrony, is notably absent as a determining factor: figs. 4 and 6 show that pairs with an asynchronous member can have very different blend ratings. The correlation of blend judgments with two dimensions that were also determinants in Grey's study (spectrum and attack) raises the possibility that similarity may be used as a first approximation for blend. Figures 10 and 11 show Multidimensional Scaling solutions from Grey(1975)[1] and for the current data, respectively. In the latter spatial proximity indicates perceived blendedness. Grey's solution was used as a starting configuration for the blend solution so as to maximize the apparent similarities between the two.[2] Indeed, a number of similar features may be noted, but many significant departures as well. Figure 12 shows figure 10 again with arrows to show where the data points relocated in figure 11. There is a relationship to similarity judgments, but apparantly with critical differences. Results from other experiments involving non-unison interval presentations will also be discussed. ----------------------------------------------------------------------- Figures [1] Instrument Pair Rating --------------- ------ french horn and french horn 8.5 french horn and bassoon 7.5 muted cello and trumpet 6.7 cello and flute 6.5 bassoon and soprano sax 6.2 french horn and eb clarinet 6.0 english horn and eb clarinet 5.7 muted trombone and bassoon 5.3 muted trombone and bass clarinet 4.9 english horn and flute 4.5 cello (*sul ponticello*) and alto sax 4.2 soprano and alto saxes 3.5 alto sax and oboe 2.1 [2] 3-D plot of Harmonics, time and energy for two simultaneoulsy-played French Horns [3] Two waveforms, time-aligned; bassoon above, french horn below [4] 3-D plot of Harmonics, time and energy for simultaneoulsy-played flute and cello [5] 3-D plot of harmonics, time and frequency for simultanously-played english horn and eb clarinet, presented from two different eye- perspectives [6] 3-D plot of Harmonics, time and energy for simultaneoulsy-played flute and english horn [7] Waveform of simultaneously-played alto sax and cello (sul ponticello) [8] Plot of centroid (x-axis) agains blend judgments (y-axis) for all 15 instruments [9] Plot of atack duration (x-axis) agains blend judgments (y-axis) for all 15 instruments [10] MDS plot of similarity judgments from Grey(1975) [11] MDS plot of blend judgments [12] Grey's MDS plot with the blend MDS plot ---------------------------------------------------------------------- Footnotes [1] Only two of Grey's three dimensions are shown here. These data points were figured from Grey(1975) by eye (and a scanning device). The actual values for the solution are not given in his dissertation. [2] No MDS solution is unique, so each is fitted to some starting configuration, either a random one or, as in this case, a chosen one. Employing Grey's coordinates in this context does not amount to adding a bias to the results; it only shows which solution (among an infinite collection of possible solutions) has the greatest resemblance to Grey's. ----------------------------------------------------------------------- Citations Grey, J. M., *An exploration of musical timbre*, Report no. STAN-M-2, Department of Music, Stanford University, 1975. Gordon, J. W., *The perception of attack transients in musical tones*, Report No. STAN-M-17, Department of Music, Stanford University, 1984. ------------------------------ Date: Mon, 3 Jul 89 18:31:00 N From: MUSIC01@UNIPAD.INFN.IT Subject: Summer Courses at the University of Padua The Centro di Sonologia Computazionale of the University of Padua (Italy) organizes its summer courses of Computer Music. All courses are in italian, except Mr Dashow's (composition and synthesis of musical events using Music360 language) in english and italian. The price of subscription to the courses is in italian lire. Futhermore informations can be obtained contacting our Centre. The calendar follows: CENTRO DI SONOLOGIA COMPUTAZIONALE UNIVERSITA' DI PADOVA CORSI DI INFORMATICA MUSICALE 1989 Sede e segreteria operativa: C. S. C. Tel. 049/8750477 8753512 Via S. Francesco 11 35121 PADOVA 26 - 30 giugno TECNICHE E METODI DELLA MUSICA INFORMATICA Docenti: F. Cappello - G. De Poli - B. Fagarazzi - R. Doati - M. Graziani - G. Tisato -A. Vidolin - F. Villa Quota di iscrizione: 200.000 (soci AIMI) 300.000 (non soci) 3 - 7 luglio COMPOSIZIONE E SINTESI: LA PROGETTAZIONE DI EVENTI MUSICALI Docente: J. Dashow Quota di iscrizione: 300.000 (soci AIMI) 400.000 (non soci) 17 - 21 luglio PROGETTAZIONE E SINTESI: LABORATORIO MUSIC 5 Docenti: B. Fagarazzi - A. Provaglio Quota di iscrizione: 300.000 (soci AIMI) 400.000 (non soci) 24 - 28 luglio CAMPIONAMENTO E ELABORAZIONE NUMERICA DEI SUONI Docenti: M. Graziani - G. Tisato Quota di iscrizione: 300.000 (soci AIMI) 400.000 (non soci) 11 - 15 settembre ELABORAZIONE DELLA VOCE NELLA MUSICA INFORMATICA Docenti:G. Tisato-G.A. Mian- A. Provaglio Quota di iscrizione: 300.000 (soci AIMI) 400.000 (non soci) 18 - 22 settembre AMBIENTI ESECUTIVI E SISTEMI IN TEMPO REALE Docenti: F. Cappello - P. Balladore Quota di iscrizione: 300.000 (soci AIMI) 400.000 (non soci) 20 - 23 novembre CORSO DI AGGIORNAMENTO IN INFORMATICA MUSICALE (il personale direttivo e docente delle scuole e' esentato dall'insegnamento) Docenti:R. Doati- M. Graziani- G. Tisato F. Villa Quota di iscrizione: 200.000 ------------------------------