source file: mills2.txt Date: Fri, 20 Dec 1996 11:30:22 -0800 Subject: fwd: RE: practical vs. theoretical implementations of micro... From: JLoffink@bangate.compaq.com ------------- Original Text From: John Loffink@HW Stor@Sys Hou, on 12/20/96 9:24 AM: To: smtp[tuning@eartha.mills.edu.] Re: practical vs. theoretical implementations of microtonality This discussion is based upon several false assumptions. Some corrections.... 1. There is no such thing as a Kurzweil K3000. You must be referring to the K2500. 2. It is very simple to do octave shifts of instruments. Multiple pitch tables are not necessary to do more than 17 notes per octave for a piano range instrument. If you need instruments with 250 discrete pitches, assuming that they repeat by octaves, you can create the pitch table once, copy the instrument and then merely shift the pitch by octaves to create the full scale pitch range. This is not possible with the Ensoniq series of samplers however (see next item). 3. The author seems to be mistaking the architecture of the Ensoniq series with other manufacturers instruments as well. The Ensoniq samplers cannot store the program parameters (tunings, filter settings, etc.) separate from the sample RAM. This is one of the great weaknesses of the ASR-10. The other is the limitation of loading 8 instruments. Kurzweil and I believe E-mu and Akai can and do allow multiple program settings using a common set of samples. So all these calculations starting with a 32M sample bank and then multiplying it by X times are wrong except for those using an Ensoniq. To do the octave shifts one merely needs to copy a few hundred bytes of program RAM and readjust the relevent parameter. 4. The assumption is made that tuning tables are the only means of creating a non-12 note scale on instruments without full scale tuning tables. This is correct only when talking about non-equally spaced scales. For equal tempered scales on the Kurzweil K2000 and K2500, it is very easy to adjust the keymap tracking parameter to a rough value so that the individual samples track the pitch range they were intended for. One then adjusts one of the two available pitch modulation sources to fine tune the pitch tracking to get a tuning accurate to +-1 cent over the entire MIDI range. Since the Kurzweil gives the pitch tracking modulation depth in real units of cents, it is very easy to calculate the proper values for any equal tempered scale, including non-octave repeating scales such as the Carlos alpha, beta and gamma scales. I recently gave an example of this to a tuning list member for 19TET and I believe he forwarded those parameters to this list. On the Kurzweil instruments this is stored as a program parameter in program RAM, not in sample RAM. The K2000 and K2500 are limited to 999 programs per instrument. I believe this to be an adequate number for most users even though you would have to use separate programs for the same sample set in 19TET and 31TET scales (but again, no extra sample RAM). Any one of these 999 programs are instantly available by a MIDI extended program change message. Multiple ET scales could even be mixed within the same composition and sequence. It takes a few minutes to calculate the parameters for a new ET scale and about 30 seconds to program each instrument. 5. The K2500 and Emu EIV both allow 128M of sample memory. The lowest cost Kurzweil option is to get a K2500R (Rack) without the sampling option for about $2500. You don't need the sampling option to use samples, only to do your own sampling from microphone, line or digital inputs! The SIMM slots and SCSI interface are on the basic unit. 30 pin 16Mx8 RAM is currently running $95-$120, so the full 128Mbyte cost is less than $1000. Of course, the assumption that you need this much is entirely wrong. Add 64M RAM, the Program RAM option, a SCSI CD-ROM drive, a 1 or 2 Gig hard drive and several sample CD-ROMs and the total price ticket is $4000-$4500. Yes, this is expensive, no doubt about it, but it is also one hellaciously powerful system. 6. Software updates do occur, especially for the more expensive instruments, so the situation is not entirely hopeless. I urge everyone to contact the manufacturers and request microtonal tuning capabilities. Make sure they know that your purchase decision is dependent upon this feature. I am personally working with Kurzweil on these matters and would love for other manufacturers to be responsive as well. One problem is that there is no consensus as to what microtonal users want. One user may want only a few full keyboard tuning scales while another may want dozens of octave based scales that can be selected on the fly for complex just intonation modulations. And let's be realistic, no mass market manufacturer will support resolutions to .01 cents or better because the VLSI chips simply aren't that accurate. Users needing this accuracy will have to rely on specialized systems like KYMA or the Justonics synthesizer. For Marcus Hobbs: The literate I've seen on the Ensoniq MR Rack says it has only one RAM tuning table location. Is this incorrect? John Loffink Received: from ns.ezh.nl [137.174.112.59] by vbv40.ezh.nl with SMTP-OpenVMS via TCP/IP; Fri, 20 Dec 1996 21:23 +0100 Received: by ns.ezh.nl; (5.65v3.2/1.3/10May95) id AA16486; Fri, 20 Dec 1996 21:25:30 +0100 Received: from eartha.mills.edu by ns (smtpxd); id XA16479 Received: from by eartha.mills.edu via SMTP (940816.SGI.8.6.9/930416.SGI) for id MAA27173; Fri, 20 Dec 1996 12:25:28 -0800 Date: Fri, 20 Dec 1996 12:25:28 -0800 Message-Id: <199612202024.PAA27765@synapse.gsfc.nasa.gov> Errors-To: madole@ella.mills.edu Reply-To: tuning@eartha.mills.edu Originator: tuning@eartha.mills.edu Sender: tuning@eartha.mills.edu