14 Nov
2008
14 Nov
'08
2:36 p.m.
hello again.
while patching some examples in puredata, i tried to compare the datastream of [pd]
and analog modular synthesizers for a better understanding of modular synthesis.
the result is now this one question of the basic data stream desing of pd.
why is there this difference of audio and data streams, because it isnt any
in the analog domain. is it the domain question, digital/analog, or a datastream
management need?
best
maex
14 Nov
14 Nov
5:20 p.m.
On Fri, 2008-11-14 at 19:45 +0100, maex(a)firstfloor.org wrote:
hello again.
while patching some examples in puredata, i tried to compare the datastream of [pd]
and analog modular synthesizers for a better understanding of modular synthesis.
the result is now this one question of the basic data stream desing of pd.
why is there this difference of audio and data streams, because it isnt any
in the analog domain. is it the domain question, digital/analog, or a datastream
management need?
the short answer, in the terms your question was posed in, is datastream
management.
the slightly longer answer is that pure data (and max) are derived from
the original Music N language's conception of how to manage this kind of
thing (the same ideas still found in CSound and SuperCollider). even
though they have gone far beyond it, they continue to distinguish
between audio & control datastreams for reasons that are mostly related
to efficiency. there are many, many cases where control data being
delivered at a bits-per-second rate significantly below that of audio
(or even as a stream of events rather than a constant flow) is more than
adequate, and saves a lot of CPU cycles.
15 Nov
15 Nov
4:38 a.m.
the short answer, in the terms your question was posed in, is datastream
management.
the slightly longer answer is that pure data (and max) are derived from
the original Music N language's conception of how to manage this kind of
thing (the same ideas still found in CSound and SuperCollider). even
though they have gone far beyond it, they continue to distinguish
between audio & control datastreams for reasons that are mostly related
to efficiency. there are many, many cases where control data being
delivered at a bits-per-second rate significantly below that of audio
(or even as a stream of events rather than a constant flow) is more than
adequate, and saves a lot of CPU cycles.
hello.
thanks for the answer, so in a way it is a rational construction
to save cpu time. exactly this i figured out for instance with the snapshot conversion
and a fast metro, and i thought it might be better to have just one
stream of data instead of converting the different streams back and forth, but
this thinking was wrong. i am amused that this still is a issue at these times.
m
6:23 a.m.
Hallo,
maex(a)firstfloor.org hat gesagt: // maex(a)firstfloor.org wrote:
thanks for the answer, so in a way it is a rational
construction
to save cpu time. exactly this i figured out for instance with the snapshot conversion
and a fast metro, and i thought it might be better to have just one
stream of data instead of converting the different streams back and forth, but
this thinking was wrong. i am amused that this still is a issue at these times.
I don't think it has to be seen as an amusing limitation, as in music
you just have both: sporadic events and flows. Maybe it's the way, the
human mind (was trained to) work(s).
Think of a light switch: Although there constantly is power applied to
it (a flow) it's quite handy to switch between on and off. You're not
thinking "I'd like to multiply the electric power by a very small value
now" but you think: "It should be dark now." Same for keyboards: A
keyboard is just a bunch of switches.
Also when programming (and using Pd *is* programming) it often is
crucial to specify an execution order in your actions which can be
easier with sporadic events than with signal flows, as you can order
events easily, even when they happen at the same time, but one after
another.
So in general the control/signal thing is not only there for CPU
performance reasons, it's also there because it's a useful device for
the brain - which is amusing, now that I think of it. ;)
Ciao
--
Frank
5:34 a.m.
Hallo,
Paul Davis hat gesagt: // Paul Davis wrote:
the slightly longer answer is that pure data (and max)
are derived from
the original Music N language's conception of how to manage this kind of
thing (the same ideas still found in CSound and SuperCollider). even
though they have gone far beyond it, they continue to distinguish
between audio & control datastreams for reasons that are mostly related
to efficiency. there are many, many cases where control data being
delivered at a bits-per-second rate significantly below that of audio
(or even as a stream of events rather than a constant flow) is more than
adequate, and saves a lot of CPU cycles.
I think, an important distinction to make is that between constantly
flowing signals and sporadic events. In computer music a typical
example would be the different time scales of audio signals and midi
events or between an audio recodding and a midi file or between a real
performance and a written score: These are fundamentally different
ways to desribe music, but both have a place in music making.
Ciao
--
Frank Barknecht Do You RjDj.me? _ ______footils.org__
14 Nov
14 Nov
7:17 p.m.
Perhaps CV is the analog of a data stream in the analog domain?
maex(a)firstfloor.org wrote:
hello again.
while patching some examples in puredata, i tried to compare the datastream of [pd]
and analog modular synthesizers for a better understanding of modular synthesis.
the result is now this one question of the basic data stream desing of pd.
why is there this difference of audio and data streams, because it isnt any
in the analog domain. is it the domain question, digital/analog, or a datastream
management need?
best
maex
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user
15 Nov
15 Nov
5:24 a.m.
Hallo,
maex(a)firstfloor.org hat gesagt: // maex(a)firstfloor.org wrote:
while patching some examples in puredata, i tried to
compare the datastream of [pd]
and analog modular synthesizers for a better understanding of modular synthesis.
the result is now this one question of the basic data stream desing of pd.
why is there this difference of audio and data streams, because it isnt any
in the analog domain. is it the domain question, digital/analog, or a datastream
management need?
Miller Puckette's book has a whole chapter devoted to the general
topic of control and audio streams:
http://crca.ucsd.edu/~msp/techniques/latest/book-html/node40.html
Note that it would be a misunderstanding of Pd if you view it just as
a software implementation of a modular synthesizer. In fact, Pd's
predecessor (IRCAM Max) didn't have audio capabilities at all.
Ciao
--
Frank Barknecht Do You RjDj.me? _ ______footils.org__
5:32 a.m.
Miller Puckette's book has a whole chapter devoted to the general
topic of control and audio streams:
http://crca.ucsd.edu/~msp/techniques/latest/book-html/node40.html
Note that it would be a misunderstanding of Pd if you view it just as
a software implementation of a modular synthesizer. In fact, Pd's
predecessor (IRCAM Max) didn't have audio capabilities at all.
yes of course.
m
16 Nov
16 Nov
10:03 p.m.
Synthesizer Pioneer Buchla (http://buchla.com/) has the destinction
between audio and control in his hardware too. Sometimes it is refered
as west coast approach, being the opposit of Moog Modulars where you can
connect everything to everything. Buchlas destinction is even in the
cables, common jacks for audio and banana for control voltages.
Cheers,
Malte
--
----
media art + development
http://www.block4.com
current events:
exhibition spame-moi La Motte-Servolex, France 17.10.-20.12.2008
5855
days inactive
5857
days old
linux-audio-user@lists.linuxaudio.org
8 comments
5 participants
participants (5)
-
Frank Barknecht -
Garett Michael Shulman -
maex＠firstfloor.org -
Malte Steiner -
Paul Davis