source file: m1463.txt Date: Wed, 1 Jul 1998 15:57:10 +0000 Subject: Adaptive Tunings From: "Patrick Ozzard-Low" I wrote: >>One thought: personally, I think I >>would be more or less content if this could be implemented as a >>convenient 'post-production' technique. But people who are >>interested in using these machines for live performance would >>presumably want more than this? Bill Sethares wrote: >All of my adative tuning work has been "off-line". I make a standard >MIDI file, read it into my adaptive algorithm, wait a while, and >then check out how it did. This is a very awkward way to work, and >is essentially a kind of 'post-production' method as you describe. I >think adaptive tunings will only be really useful when you can sit >down and play as if it were a musical instrument. Such >'post->production' methods are fine for demonstrating the concepts, >but not >for exploring the technique fully. I agree with Bill that real-time adaptive tunings would be best. I guess quite a few people on the list will be familiar with John S. Allen's Key Concepts prototype Notebender keyboard. I only know of it from reading JSA's web pages. It allows variable intonation for a keyboard by 'undercutting' the black keys, so that both black and white keys can slide (with finger pressure) to and from the player - these movements are then tranmitted as pitch changes. JSA says this arrangment is less easy to control than a traditional keyboard, but easier than clavichordal intonations. However, a configurable, automatic system, with can be modified or overriden by physical input would be better. >I'm currently attempting a 'real-time' version using Max, though its >unclear at the present time how many simplifications are going to be >needed before it will actually run in real time. Of course, this is >for MIDI instruments only. A far more interesting thought is the >idea of operating on acoustic sounds... As regards acoustic instruments, I have been writing up a few ideas on this in the 21C Orch Instrs paper - particularly my idea of creating an instrument (woodwind or brass) which may switch at will between adaptive (fixed-but-variable, its normal condition) or fixed intonation; and conceivably, between a number of ATS. Firstly, I've mentioned before Giles Brindley's Logical Bassoon - a purely acoustic instrument in which the 18/19th century keywork is replaced by a system of solenoid driven tone-hole pads. Professor Brindley is still playing this instrument every week in an amateur orchestra in London. (Brindley is a famous neurophysicist, not an instrument maker). I've spoken to him on the phone, but have yet to see the instrument(s) - I think he built 3 of them. Lets take this a step further. Here's one of my crazy ideas - and a short dump from the paper. I'd certainly appreciate any feedback on this idea. Would it work???!!! EXTRACT STARTS ********* Concieve a 'logical' woodwind in which an internal electronic pitch sensor provides feedback (via software) to a auto-corrective venting mechanism in real-time. The instrument would be 'self-correcting' within certain configurable tuning limits; further, this might be applied in a number of ways, for example, to achieve_ strict_ temperament, or to achieve _strict_ just intonation in any key (at will). For large n-division systems the system could ensure that the instrument is always at one pitch or another of a temperament, making performance more reliable; for JI systems, this would function similarly, but an on-board switch would tell the software which pitch is 1/1 at any given time. Real-time (instantaneous) digital pitch correction for audio has been available for some years, for example, as part of the ProTools digital audio processing suite (recently released in rack-mount form as the Antares Auto-Tune). [Anyone have good/bad experience of this??] A performance by a singer or instrumentalist can be instantaneously 'corrected' so that all inflections adhere strictly to a given tuning system. The Antares allows custom scales, but the typical implementation is strict 12-ET. This runs counter to the norm of variable intonation (for singers, wind and brass etc.), but more sophisticated realisations are clearly possible - for example, in its current version the Antares is able to 'ignore' vibrato and glissandi. The suggestion here is not to use such a mechanism to translate an acoustically produced sound into a digitally modulated one, but to feedback corrections of the electronically monitored pitch signal to the tone-hole closure system of the acoustic instrument. The player would thus be in constant dialogue not only with the vibrations of the instrumental system, but with extremely fine adjustments of tone-hole closure being carried out independently by the corrective process. This would require the finest possible balance of sensitivity and stability in terms of software interactivity, since unless pitch adjustments were fine, unwanted glissandi and suchlike would result; on the other hand, extreme subtleties of pitch inflection should not be overridden, and the player would need to be able to vary the sensitivity of the device (from 0 -100% ) at the slightest touch of a key. EXTRACT ENDS ************ If this could be made to work on a woodwind, it is conceivable it could also work for brass too - this time the interactivity taking place through a slide mechanism (maybe driven by a tiny stepper motor??) The important aspect of this is that the physical interface is replaced by a virtual one. (By the way, in the case of woodwind and brass, the most important (personal) aspect of 'contact' with a *keyed* (or valve) instrument is the mouth: the performers of 'logical acoustic instruments' report no loss of 'feel' or sensitivity). In addition, an adaptive acoustic piano could be made by creating an instrument with a very large number of unison string groups, and the keyboard transmits via a Disklavier type piano action which is controlled by 'adaptive' software. Patrick O-L ------------------------------ End of TUNING Digest 1463 *************************