back to list

Numerology (Mac OSX sequencer) having tuning features added

🔗paolovalladolid <phv40@...>

1/31/2006 8:37:09 AM

A very recent post by Jim Coker, developer of Numerology, in response
to a customer's question about microtuning:

"There are 2 ways to transmit micro-tuning to a MIDI instrument:

Best:
- if the instrument responds to the MIDI Tuning Spec, then the
ideal approach is to generate a tuning dump and have the instrument
use that tuning. Some virtual instruments I know of that actually support
this are FM7 and the TimeWARP 2600.

Decent:
- if the instrument DOES NOT respond to the MIDI Tuning Spec,
then the app generating the notes must:
- generate each note w/ necessary pitch bend to tune the note
- generate each note on a separate MIDI channel
This usually works if:
- the target synth doesn't try to re-order MIDI messages
(I hear that Logic sometimes does this), which would screw up the
pitch bend messages
- the target synth can run in multi-timbral mode w/ the necessary
number of parts,
one part per voice.

FYI:
The AU spec has direct support for per-note alternate tunings
via an 'extended Note API'. All synths are supposed to support
this, but I assume that most of them throw away any fractional
pitch information.

Alternate tuning support in Numerology would definitely support
the latter, and probably also the first two, though I don't know what
the implementation priority would be.

In Num 2.0, *all* pitch values are 32-bit float, and are only converted
to 7-bit MIDI values at the last minute.

Cheers,
Jim"

I learn something new every day. What I just learned:

- Audio Unit specificiation (Mac OSX) supports per-note alternate tunigns
- Logic already implements Hermode tuning (yes, I really didn't know)
- Jim hung out with Robert Rich (champion of JI in synth music world))
at NAMM

🔗Phi <phi@...>

1/31/2006 11:44:54 AM

Hi Paolo!

> I learn something new every day. What I just learned:
> - Audio Unit specification (Mac OSX) supports per-note alternate tunings

Where did you find this AU specification?
I'm really interested about it!

TIA ;-)
Philippe

🔗paolovalladolid <phv40@...>

1/31/2006 12:28:11 PM

--- In MakeMicroMusic@yahoogroups.com, "Phi" <phi@o...> wrote:
>
> Hi Paolo!
>
> > I learn something new every day. What I just learned:
> > - Audio Unit specification (Mac OSX) supports per-note alternate
tunings
>
> Where did you find this AU specification?
> I'm really interested about it!

The info is in Jim Coker's post that I quoted, which you in turn
quoted from. ;)

Remember, Jim develops software for OSX. He knows his Audio Units...

🔗paolovalladolid <phv40@...>

1/31/2006 12:30:23 PM

I forgot to post the Numerology URL:

http://www.five12.com/

Jim Coker, the developer, is a nice guy and very responsive to customers.

🔗Phi <phi@...>

2/1/2006 12:18:07 AM

>> Audio Unit specification (Mac OSX) supports per-note alternate tunings
>> Where did you find this AU specification?
>
> The info is in Jim Coker's post that I quoted, which you in turn
> quoted from. ;)
> Remember, Jim develops software for OSX. He knows his Audio Units...

Thanks Paolo! Found at the ADC website about "Extended Note Event":
<http://developer.apple.com/audio/>

Example page 19 of the "QuickTime Music Architecture Guide":

"0-127 standard pitch, 60 = middle C 0x01.00 ... 0x7F.00 allowing
256 microtonal divisions between each notes in the traditional equal
tempered scale"

Cheers ;-)
Philippe

🔗Gene Ward Smith <gwsmith@...>

2/1/2006 7:25:02 PM

--- In MakeMicroMusic@yahoogroups.com, "Phi" <phi@...> wrote:

> "0-127 standard pitch, 60 = middle C 0x01.00 ... 0x7F.00 allowing
> 256 microtonal divisions between each notes in the traditional equal
> tempered scale"

This seems to be saying that Macintosh uses 3072-equal for tuning
purposes. Correct?

🔗Jon Szanto <jszanto@...>

2/1/2006 8:01:07 PM

Gene,

{you wrote...}
>--- In MakeMicroMusic@yahoogroups.com, "Phi" <phi@...> wrote:
>
>> "0-127 standard pitch, 60 = middle C 0x01.00 ... 0x7F.00 allowing
>> 256 microtonal divisions between each notes in the traditional equal
>> tempered scale"
>
>This seems to be saying that Macintosh uses 3072-equal for tuning
>purposes. Correct?

Yeah, but only by virtue of binary coding, not because someone looked for an EDO to support.

Cheers,
Jon