synth tuning

Hello Erich, Everyone,

I'm happy to anounce I'm now the owner
of a Yamaha PSR-530, XG compatible keyboard. It's limited
to twelve note scales (-63 to +64), 1200/768 (100cent/64) resolution.
The range and number of tweakable parameters is terrific, making it
possible to obtain very professional results. Comes with a built
in computer interface (Mac & PC). Also, a 5-octave keyboard (with one
split point), pitch-wheel, & decent sounding 12 Watt stereo speaker
system.

You do however have to use a computer to do any tweaking, and turn
off local control, run the keyboard through a MIDI thru (hardware or
software,
I use the little freeware "OMS thru" on my PowerMac) and back into the
PSR-530.
A bargain at the equivalent of $500.

>yamaha and roland synths both have sysex messages for
>scale tuning, which allows you to change the tuning of all
>tones in one pitch class (all c's, all c#'s ..) in increments of 1 cent.

Actually, I believe Yamaha has fudged the terminology in the
manual for the PSR-530. The MIDI Master tuning can in fact be tuned
-100 cent to -100 cent. But for microtonal tuning ("Scale Tuning" in
Yamaha speak), they could potentially confuse users with little
microtonal technical knowledge, because this is what they say :

"...alows each individual note of the octave to be tuned over the
range from -64 to +64 cents in 1 -cent increments (1 cent = 1/100th
of a semitone)."

These are actually 64ths of a semitone, aren't they ?

>the sysex messages are vendor specific and therefore different.
Yup, from what I've seen, Roland GS sysex microtuning messages
consist of a single 22-byte message, while the Yamaha PSR's XG
apparently requires 12 seperate 9-byte messages (one for each note).
This is how I do it. I'm waiting for clarification on some things
from Yamaha.

>general midi has master channel tune in increments of
>100/8192 cents (much finer)
>so it is possible to tune each channel separately.
Thanks for mentioning this possibility, Erich, it completely escaped my
attention. Could someone clarify : There is a MIDI MULTIPART "Detune"
parameter. It says : "-12.8..+12.7[Hz]" . So, what is the size of the
increments ? And does anyone here know how to write the Sysex message ?
The manual gives two "High-Mid-Low" Addresses, and says :
"1st bit3..0-> bit7..4" & "2nd bit3..0-> bit3..0"

There is also the RPN Parameter : "Fine Tune", which says
"[00H,00H] - [40H,00H] - [7FH,7FH]",
"(-8192*100/8192) - 0 - (+8192*100/8192)"
Same questions, what size the increments, and how goes the Sysex ?
If it's 0 to 8192, then I can figure out the Hex.

Just using a keyboard with atleast one split point, it is possible to
fool around with many equal temperaments. I've been giving it a bit of
thought and it should handle equal temps of upto 24 notes per octave.
All it entails is transposing both left & right sides to the same
"octave" range,
then tuning the both parts to the odd numbered pitches of an ET, and
shifting
the channel master tune so the pitches on the right hand become the even
numbered pitches of the scale.

I guess some of you do this already. Anyway, it's safer than
impersonating Penn & Teller.It's more flexible than tuning
individual channels, since you retain more polyphony.

>since a midi keyboard (or any other midi instrument)
>will produce notes only on one channel, you need a tool for
>rechannelizing. there is a freeware program
>midicable which allows you to do exactly that.

Here we go again ! Is this a Power Mac application or Windows ?
If it's Windows, can someone point to a Power Mac thingy for
rechannelizing ?

Ciao,
Drew

> if X , Y are two tone frequencies,
> how to tune the value Z=sqrt(X*Y) acousticaly

I'm not completely certain what in particular you mean by "acoustically"
in this particular case, but I suppose it's worth pointing out that, to our
logarithmic tuning perception, a geometric mean sounds like an arithmetic
mean.

So, if by "acoustically" you mean "by ear", you can tune that
intermediate pitch so that the melodic interval size from the 1:1 to the
pitch in question sounds the same as the interval from the pitch in
question to the 5:4.

On Mon, 24 Aug 1998, Jan Haluska wrote:
> if X , Y are two tone frequencies,
> how to tune the value Z=sqrt(X*Y) acousticaly
> ( sqrt(X*Y) denotes the geometrical mean of the tones X and Y )?
> Demonstrate the method on X = 1, Y = 5/4 .

In the general case, this is a Hard Problem. Which is why widespread
experimentation with random ETs has had to wait until the computer
revolution.

In the specific case requested, that of determining the mean tone (used
in 1/4-comma meantone tuning), there are various methods. If I were
tuning my virginal or something, I would probably do something like
this: tune up a (just) 3/2 and a 6/5 to get 9/5 above; tune down a 3/2
and then up a 4/3 (again just) to get 8/9 below, then tune the target
string (halfway between 10/9 and 9/8, we hope) to beat at the same rate
with each of those.

--pH http://library.wustl.edu/~manynote
O
/\ "Churchill? Can he run a hundred balls?"
-\-\-- o
NOTE: dehyphenate node to remove spamblock. <*>

On Tue, 25 Aug 1998, Paul Hahn wrote:
> tune up a (just) 3/2 and a 6/5 to get 9/5 above; tune down a 3/2
> and then up a 4/3 (again just) to get 8/9 below, then tune the target
> string (halfway between 10/9 and 9/8, we hope) to beat at the same rate
> with each of those.

Gack, I messed up. To get the desired result the upper interval would
have to beat twice as fast as the lower. Hmm--I think to get equal
beating you'd have to do something like this: up to the 5/3, down a
chain of three 3/2s-4/3s to 27/32. The desired mean tone would beat
equally with the 5/3 (a narrow 3/2) and the 27/32 (a narrow 4/3).

--pH http://library.wustl.edu/~manynote
O
/\ "Churchill? Can he run a hundred balls?"
-\-\-- o
NOTE: dehyphenate node to remove spamblock. <*>

On Tue, 25 Aug 1998, Paul Hahn wrote:
> On Tue, 25 Aug 1998, Paul Hahn wrote:
>> tune up a (just) 3/2 and a 6/5 to get 9/5 above; tune down a 3/2
>> and then up a 4/3 (again just) to get 8/9 below, then tune the target
>> string (halfway between 10/9 and 9/8, we hope) to beat at the same rate
>> with each of those.
>
> Gack, I messed up. To get the desired result the upper interval would
> have to beat twice as fast as the lower.

You can also get it to work by turning it upside down--tune down two
4/3s from the 5/4 to get 45/64, up a 4/3 from the 5/4 (or a 5/3 from the
1/1) and then down a 6/5 to get 25/18; the desired mean tone would then
beat equally narrow of a 8/5 with 45/64 and narrow of a 5/4 with 25/18.

--pH http://library.wustl.edu/~manynote
O
/\ "Churchill? Can he run a hundred balls?"
-\-\-- o
NOTE: dehyphenate node to remove spamblock. <*>