source file: mills2.txt Date: Tue, 11 Mar 1997 21:07:01 -0800 Subject: crystal sounds From: William Sethares In yesterdays digest, Ray Tomes asked me to clarify some about the techniques used in the "Sound of Crystals" article... ..can you please explain how you determined what frequencies to play at what time. I imagine that the crystal output (slowed down) is a bit like sitting on the keyboard and staying there so you need a way to seperate the notes. There are many ways that one can utilize spectra (such as that of the morphine cyrstal) to generate sound. The two techniques that I have explored are: (1) using the spectrum to define a single "sound" -- in this approach I generated about 2 seconds of sound by mapping the x-ray frequencies into audio frequencies. This was then transferred to a sampler, and played like any other sampled sound, in any desired tuning. (2) using the spectrum to define a "scale" -- in this case, each x-ray frequency of the crystal is mapped to a separate key of the keyboard (and of course, also mapped down into the audio range). Using a sine wave sample then allows the "complete" sound to be played by pressing all the keys simultaneously. Thus, in both cases, I "composed" the pieces by improvising on the keyboard. In the first case the crystal provides the basic "sound" while in the second case it provides the "tuning". Greg Taylor then said some nice things (blush) and continued... Just out of curiousity, what other XRD stuff sounded interesting? Can one generalize? I wouldn't mind trying to map an essential oil like Melissa, which I could then collide with the Allman Brothers tune of the same name.... Well, our original idea was to try and develop a kind of "auditory crystallography" in which you might use the ear to hear patterns in the x-ray diffraction data that the eye doesn't readily see. What we ended up with was a way of generating some interesting sounds (and scales, I suppose). The problem was that we were never really able to convince ourselves that we could reliably distinguish the sounds of various materials. Of course, every substance sounded somewhat different, but the differences weren't something that was easy to describe or remember. So, sorry Greg. We were never able to discern any real patterns in how various materials sounded, nor were we able to generalize enough to predict if something would sound "interesting" (whatever that means) or not. Thus, it would be impossible to predict if linguistic similarities (like your oil and an Allman Brothers song) would be acoustically compatible. Nor even if the sound of oil would be slick. [The other, perhaps larger issue that still remains largely unexplored is the issue of the effect of details of the mapping. I once started a paper on this, but never finished]. Received: from ns.ezh.nl [137.174.112.59] by vbv40.ezh.nl with SMTP-OpenVMS via TCP/IP; Wed, 12 Mar 1997 13:53 +0100 Received: by ns.ezh.nl; (5.65v3.2/1.3/10May95) id AA03054; Wed, 12 Mar 1997 13:53:36 +0100 Received: from ella.mills.edu by ns (smtpxd); id XA03050 Received: from by ella.mills.edu via SMTP (940816.SGI.8.6.9/930416.SGI) id EAA03231; Wed, 12 Mar 1997 04:51:56 -0800 Date: Wed, 12 Mar 1997 04:51:56 -0800 Message-Id: <3326A70B.1AD2@cavehill.dnet.co.uk> Errors-To: madole@mills.edu Reply-To: tuning@ella.mills.edu Originator: tuning@eartha.mills.edu Sender: tuning@ella.mills.edu