MIDI stretched piano sounds

OK, I did it: added a piano stretch option to my retuning program. It's
even possible to get a stretched 12-tET version. I'm using the curve
that Monz quoted in

/tuning/topicId_13830.html#13830

And that Allan Myhara cleverly turned into a simple expression in

/tuning/topicId_13830.html#13879

Right around middle C, there's very little stretch per octave, but at
both extremes the change gets steeply intense. From the lowest to the
highest note, there is a full 50 cents of stretch.

Two list members sent me piano sequences - thanks, guys!! I'll be
returning stretched and tuned versions, and maybe even a
double-stretched and negatively stretched version for listening. I'm
still doing sanity checks on the output, but so far it seems to be
working correctly.

What does it sound like? To my ear, using my rather rough piano voice,
about like it did before stretch! Even a file I made up with C's all
up and down the line sounds quite alike to me on first hearing. I'm
going to make a file with the notes arpeggiated and see if that makes it
clearer. And/or, try another piano voice.

One very strange thing is happening to the files I've run through so
far: the overall tuning range tends to move downward rather than being
more extreme. In other words, high and low notes tend to be the ones
that are tuned downward from 12-tET; notes in the middle of the keyboard
tend to be tuned upward from 12-tET. This could be a bug, though my
spot checks look ok. Or it could be an anomaly in the fairly small
sample set. It DOESN'T seem to be that the pieces have only low notes
in them.

Method note: when the piece is tuned, it is tuned as before, without
regard to stretch (or even the knowledge of what octave each pitch
degree resides in). Then stretch is added in at the last moment, with
knowledge of actual pitch locations. I did have to modify the channel
assignment logic so that notes of the same pitch class can't share a
channel unless the pitch is IDENTICAL (same octave). That makes
channels scarcer than ever! But for solo piano, it should usually work.

JdL

--- In tuning@y..., "John A. deLaubenfels" <jdl@a...> wrote:

> One very strange thing is happening to the files I've run through so
> far: the overall tuning range tends to move downward rather than
being
> more extreme. In other words, high and low notes tend to be the
ones
> that are tuned downward from 12-tET; notes in the middle of the
keyboard
> tend to be tuned upward from 12-tET.

That's bizarre.

> Method note: when the piece is tuned, it is tuned as before, without
> regard to stretch (or even the knowledge of what octave each pitch
> degree resides in). Then stretch is added in at the last moment,
with
> knowledge of actual pitch locations.

I don't see how that could result in the behavior you're reporting
above.