Real Life

>However when classically trained singers or players of fretless string
>instruments perform music that is allegedly in some form of just
intonation, >I wonder how often a performer's realisation deviates from the
theoretically >ideal tuning prescribed by a composer.

In the case of the classical strings and woodwinds, and voice, just about
100% of the time. This goes for "standard" tuning as much (and probably
more) than alternate tunings.

>How would one go about analysing a performance to determine if the tuning
>was more like one tuning than another?

If it is a solo melody, you can use pitch tracking software. Otherwise,
you sit down with the recording and transcribe it in 12tET. Or, if you
already have a score, so much the better. Then, you work out several
plausible versions in Just Intonation and have a computer perform them.
Listen and tweak, listen and tweak, listen and tweak, listen and tweak...
and remember that fixed pitch theory can't cope with melismatic intonation,
and all that goes along with it.

>In other words while we can know what they are supposed to sing or play, how
>do we demonstrate when they are actually doing it?

I would be careful with that word "supposed".

Carl

>>However when classically trained singers or players of fretless string
>>instruments perform music that is allegedly in some form of just intonation,
>>I wonder how often a performer's realisation deviates from the theoretically
>>ideal tuning prescribed by a composer.

>In the case of the classical strings and woodwinds, and voice, just about
>100% of the time. This goes for "standard" tuning as much (and probably
>more) than alternate tunings.

>>How would one go about analysing a performance to determine if the tuning
>>was more like one tuning than another?

> If it is a solo melody, you can use pitch tracking software. .................

Thanks Carl. I can see from your response what I omitted. I was trying to keep it brief.

I've written a score using standard notation for four a capella voices. I want them to sing the pitches slightly differently to those implied by the notation. With the scale based on a 7-limit diamond I made the following pitches correspond
1/1 C
8/7 D
7/6 Eb
6/5 Eb
5/4 E
4/3 F
7/5 Gb
10/7 F#
3/2 G
8/5 Ab
5/3 A
12/7 A
7/4 Bb

I created a tuning crib for each of the singers using Finale to create the MIDI file, MIDI2CS to convert this to Csound, Manuel's Scala to work out the tuning ratios and a bit of editing to tune all pitches using John Fitch's Csound opcode cps2pch following some examples given by Bill Alves. (Thanks, guys, you've probably forgotten by now, but don't worry it was a couple of years ago).

A pitch tracker to verify whether the singers got it right in melodic sections, well yes maybe .... but it's in the harmonic sections that I'd really like to know. Because I wonder how far I must go in creating such tuning cribs in future pieces. It would be nice to have an objective way of demonstrating to singers, or players of fretless instruments that they have succeeded in achieving a particular type of intonation. eg like put a recording of the (ensemble) performance through some sort of pitch analysis to show them when they hit the bulls-eye and when they were a little off the ideal pitch. I asked the singers not to sing vibrato when they sing in ensemble only in solo sections. I guess I'm more concerned about JI in the harmonic sense. Does anyone know of suitable analysis software for this purpose ie. working backwards to extract tuning information from the real world.

Greg S

After years of working with analog recordings, filters, oscillators and
frequency counters, I have recently found the Windows-platform freeware
program

Spectrogram (www.monumental.com/rshorne/gram.html)

and the shareware programs

AcidWAV and Tune!It

to be useful in analysis of pitch and timbre in real, existing
performances.

Typically, I will make 48k samples under AcidWAV, run an FFT under AcidWAV
or Spectrogram to see if there is any noise that can be removed, and
carefully -- albeit subjectively -- select ranges and periods to sample.

I cannot overemphasize the importance of subjective musical judgement in
carrying out such analyses. From my experience, any attempt to perform the
analysis independently from such judgement will be buried in data that is,
at best, peripheral to the listening experience.