soundfonts and looping (Granular or 'Boulder' synthesis)

I guess the technique I'm describing might be more accurately called
"boulder" synthesis .. .i.e., granular synthesis with large grains (> .1
seconds long each).

In other words, lets say you have a 3 second violin sample: you might at
random chooose the following bits of it:
0.5" - 1.4"
1.2" - 2.2"
0.83" - 1.99"
1.3" - 2.25"
and so on.

Now imagine each of these sound "boulders", one after the other, overlapped,
with appropriate envelopes---the result is a continuous sound, with
life-like randomness instead of dull looping.

For an example, check out my piece *Aftermath* (which I've posed before):
http://works.music.columbia.edu/~chris/aftermath/aftermath.mp3
(Incidentally, there are some musique concrete sounds in this piece 2, not
relevant to this discussion. Also . . the piece is in JI, as elucidated in
previous MMM messages, so this is not OT for the list :)

As far as implementation, if you took a typical synth engine, and replaced
it with this kind of engine, I think you could basically use the same
soundfonts that typical synth engines use . . .the difference would be in
the machine. It might be a little more CPU-intensive, but I think today's
computers can handle it :)

(Actually, I did specifically build this for* Aftermath* . . .what you're
hearing there is essentially a live performance on MAX/MSP. It's not a
general synth engine, it's specifically made for those string sounds, and
those sounds themselves are not a general soundfont, they are my own
"soundfont" made up of long string sounds from symphonies of Mahler,
Bruckner, (I love the sound of long, tortured, PostRomantic string notes)
and I think Beethoven, Prokofiev.).

--
--------------------------------------------
Christopher Bailey
chris@...
zipzappoozoo@...
http://music.columbia.edu/~chris

[Non-text portions of this message have been removed]

Christopher Bailey wrote:

> As far as implementation, if you took a typical synth engine, and replaced
> it with this kind of engine, I think you could basically use the same
> soundfonts that typical synth engines use . . .the difference would be in
> the machine. It might be a little more CPU-intensive, but I think today's
> computers can handle it :)

Maybe Csound could do this. It uses standard soundfonts but only as sample sources. So it tells you which sample(s) to play, the pitch correction, and the loop points. I haven't checked the library so I don't know if it's possible to plug in a new looping method.

Unless you can find a way of adding a flag to say you want the granular looping, you'll break soundfonts that depend on the loop sounding like a loop. That'll be true for anything that goes beyond the simplistic synth model the soundfonts are based on.

Graham

> As far as implementation, if you took a typical synth engine, and replaced
> it with this kind of engine, I think you could basically use the same
> soundfonts that typical synth engines use . . .the difference would be in
> the machine. It might be a little more CPU-intensive, but I think today's
> computers can handle it :)

By "machine", I mean, the [software] engine that plays the sounds (not the
ahrdware, specifically).

My only real point was, I was surprised that no synth engines out there
worked like this.

[Non-text portions of this message have been removed]

Hi Christopher.

I enjoyed your music; thanks for sharing.
I guess you must be aware that Curtis Roads did exactly
this kind of thing in the late 70's in his initial experiments
implementing Xenakis' theories of granular synthesis.
Roads says this in a lecture which is on the CD/DVD Point
Line Cloud:
<http://www.amazon.com/Point-Line-Cloud-Curtis-
Roads/dp/B0008G2EXM/ref=pd_bbs_sr_1?ie=UTF8&s=music&qid=1197570614&sr=1-
1>

For those interested in the basics:
<http://music.calarts.edu/~eric/gs.html>

... or a definitive text on the subject, "Microsounds":
<http://www.amazon.com/Microsound-Curtis-Roads/dp/0262681544/ref=pd_bbs_sr_1?
ie=UTF8&s=books&qid=1197400536&sr=8-1>

Yours,
Aaron Hunt
H-Pi Instruments

--- In MakeMicroMusic@yahoogroups.com, "Christopher Bailey" <chris@...> wrote:
>
> I guess the technique I'm describing might be more accurately called
> "boulder" synthesis .. .i.e., granular synthesis with large grains (> .1
> seconds long each).
>
> In other words, lets say you have a 3 second violin sample: you might at
> random chooose the following bits of it:
> 0.5" - 1.4"
> 1.2" - 2.2"
> 0.83" - 1.99"
> 1.3" - 2.25"
> and so on.
>
> Now imagine each of these sound "boulders", one after the other, overlapped,
> with appropriate envelopes---the result is a continuous sound, with
> life-like randomness instead of dull looping.
>
> For an example, check out my piece *Aftermath* (which I've posed before):
> http://works.music.columbia.edu/~chris/aftermath/aftermath.mp3
> (Incidentally, there are some musique concrete sounds in this piece 2, not
> relevant to this discussion. Also . . the piece is in JI, as elucidated in
> previous MMM messages, so this is not OT for the list :)
>
>
> As far as implementation, if you took a typical synth engine, and replaced
> it with this kind of engine, I think you could basically use the same
> soundfonts that typical synth engines use . . .the difference would be in
> the machine. It might be a little more CPU-intensive, but I think today's
> computers can handle it :)
>
> (Actually, I did specifically build this for* Aftermath* . . .what you're
> hearing there is essentially a live performance on MAX/MSP. It's not a
> general synth engine, it's specifically made for those string sounds, and
> those sounds themselves are not a general soundfont, they are my own
> "soundfont" made up of long string sounds from symphonies of Mahler,
> Bruckner, (I love the sound of long, tortured, PostRomantic string notes)
> and I think Beethoven, Prokofiev.).
>
>
>
> --
> --------------------------------------------
> Christopher Bailey
> chris@...
> zipzappoozoo@...
> http://music.columbia.edu/~chris
>
>
> [Non-text portions of this message have been removed]
>

Well, I shouldn't have said _exactly. I mean, not Mahler
string sections the way you are doing it... I meant only the
basic idea you've outlined using acoustic samples.

Aaron Hunt
H-Pi Instruments

--- In MakeMicroMusic@yahoogroups.com, "Aaron Andrew Hunt" <aahunt@...> wrote:
>
> Hi Christopher.
>
> I enjoyed your music; thanks for sharing.
> I guess you must be aware that Curtis Roads did exactly
> this kind of thing in the late 70's in his initial experiments
> implementing Xenakis' theories of granular synthesis.
> Roads says this in a lecture which is on the CD/DVD Point
> Line Cloud:
> <http://www.amazon.com/Point-Line-Cloud-Curtis-
> Roads/dp/B0008G2EXM/ref=pd_bbs_sr_1?ie=UTF8&s=music&qid=1197570614&sr=1-
> 1>
>
> For those interested in the basics:
> <http://music.calarts.edu/~eric/gs.html>
>
> ... or a definitive text on the subject, "Microsounds":
> <http://www.amazon.com/Microsound-Curtis-
Roads/dp/0262681544/ref=pd_bbs_sr_1?
> ie=UTF8&s=books&qid=1197400536&sr=8-1>
>
> Yours,
> Aaron Hunt
> H-Pi Instruments
>
>
> --- In MakeMicroMusic@yahoogroups.com, "Christopher Bailey" <chris@> wrote:
> >
> > I guess the technique I'm describing might be more accurately called
> > "boulder" synthesis .. .i.e., granular synthesis with large grains (> .1
> > seconds long each).
> >
> > In other words, lets say you have a 3 second violin sample: you might at
> > random chooose the following bits of it:
> > 0.5" - 1.4"
> > 1.2" - 2.2"
> > 0.83" - 1.99"
> > 1.3" - 2.25"
> > and so on.
> >
> > Now imagine each of these sound "boulders", one after the other, overlapped,
> > with appropriate envelopes---the result is a continuous sound, with
> > life-like randomness instead of dull looping.
> >
> > For an example, check out my piece *Aftermath* (which I've posed before):
> > http://works.music.columbia.edu/~chris/aftermath/aftermath.mp3
> > (Incidentally, there are some musique concrete sounds in this piece 2, not
> > relevant to this discussion. Also . . the piece is in JI, as elucidated in
> > previous MMM messages, so this is not OT for the list :)
> >
> >
> > As far as implementation, if you took a typical synth engine, and replaced
> > it with this kind of engine, I think you could basically use the same
> > soundfonts that typical synth engines use . . .the difference would be in
> > the machine. It might be a little more CPU-intensive, but I think today's
> > computers can handle it :)
> >
> > (Actually, I did specifically build this for* Aftermath* . . .what you're
> > hearing there is essentially a live performance on MAX/MSP. It's not a
> > general synth engine, it's specifically made for those string sounds, and
> > those sounds themselves are not a general soundfont, they are my own
> > "soundfont" made up of long string sounds from symphonies of Mahler,
> > Bruckner, (I love the sound of long, tortured, PostRomantic string notes)
> > and I think Beethoven, Prokofiev.).
> >
> >
> >
> > --
> > --------------------------------------------
> > Christopher Bailey
> > chris@
> > zipzappoozoo@
> > http://music.columbia.edu/~chris
> >
> >
> > [Non-text portions of this message have been removed]
> >
>

> Hi Christopher.
>
> I enjoyed your music; thanks for sharing.
> I guess you must be aware that Curtis Roads did exactly
> this kind of thing in the late 70's in his initial experiments
> implementing Xenakis' theories of granular synthesis.

Yes . . . it would be wierd if someone came up with the idea independently,
and magically thought to call it "granular synthesis" just as Xenakis did.
(A better term might be "granular processing", since the procedure, as I use
it, makes use of found sounds).

My point was not that granular synthesis is new, but that, as far as I know,
it's not been used in a commercial synth engine, and I find that odd.

CBailey

[Non-text portions of this message have been removed]

At 07:08 AM 12/14/2007, you wrote:
>> Hi Christopher.
>>
>> I enjoyed your music; thanks for sharing.
>> I guess you must be aware that Curtis Roads did exactly
>> this kind of thing in the late 70's in his initial experiments
>> implementing Xenakis' theories of granular synthesis.
>
>Yes . . . it would be wierd if someone came up with the idea independently,
>and magically thought to call it "granular synthesis" just as Xenakis did.
>(A better term might be "granular processing", since the procedure, as I use
>it, makes use of found sounds).
>
>My point was not that granular synthesis is new, but that, as far as I know,
>it's not been used in a commercial synth engine, and I find that odd.
>
>CBailey

Granular synthesis is used in many commercial products for
time stretching, and is also used as a synthesis technique in
some, such as Reason (in Malstrom). -C.

--- In MakeMicroMusic@yahoogroups.com, "Christopher Bailey" <chris@...> wrote:
> > I guess you must be aware that Curtis Roads did exactly
> > this kind of thing in the late 70's in his initial experiments
> > implementing Xenakis' theories of granular synthesis.
>
> Yes . . . it would be wierd if someone came up with the idea independently,
> and magically thought to call it "granular synthesis" just as Xenakis did.
> (A better term might be "granular processing", since the procedure, as I use
> it, makes use of found sounds).
>
> My point was not that granular synthesis is new, but that, as far as I know,
> it's not been used in a commercial synth engine, and I find that odd.

Ah, I see. I think it is an option now in many software synths. Does anyone know of a
hardware synth using it?

Yours,
Aaron Hunt
H-Pi Instruments