On 1/22/2012 10:29 PM, Julian Leviston wrote:
On 23/01/2012, at 4:17 PM, BGB wrote:
as opposed to either manually placing samples on a timeline (like in
Audacity or similar), or the stream of note-on/note-off pulses and
delays used by MIDI, an alternate idea comes up:
one has a number of delayed relative "events", which are in-turn
piped through any number of filters.
then one can procedurally issue commands of the form "in N seconds
from now, do this", with commands being relative to a base-time (and
the ability to adjust the base-time based either on a constant value
or how long it would take a certain "expression" to finish playing).
likewise, expressions/events can be piped through filters.
filters could either apply a given effect (add echo or reverb, ...),
or could be structural (such as to repeat or loop a sequence,
potentially indefinitely), or possibly sounds could be entirely
simulated (various waveform patterns, such as sine, box, and
triangle, ...).
Heya,
Yeah, I've had that idea for a while - although a more comprehensive
one (I write music). Take a look at what Apple did to their own
product Final Cut Pro... to turn it into Final Cut Pro X, and notice
that there are rumors surrounding Logic Pro X, and I'm pretty sure
you'll see that this idea is where Apple will most likely go when they
release Logic Pro X.
In Final Cut Pro, they call it their "magic timeline".
I am pretty sure this costs money though.
in my case, I was using Audacity, partly because it was free...
By the way, what you're describing CAN be done with Ableton Live
without much trouble... also Ableton Live has the ability to use Max
for Live, which is Cycling 74's excellent Max/MSP product inlined into
a Live instrument (what you're calling various waveform patterns).
It's sine, square/pulse and triangle by the way, not "box"... and we
also can use all sorts of other waveforms... generated or sampled...
I wasn't claiming originality though (although, if someone has free
software for this, I might be more interested).
I knew it was called square-wave, but for whatever reason (mental
slip...) wrote "box" instead. same idea though:
have a wave-generation function, stream things through filters, ...
although not nearly as advanced, this is sort of how the audio-mixer in
my 3D engine works (it can also use streams and apply filters through
streams, ...), as well as implementing effects like echoes and
doppler-shifting and delays (sounds will not start playing instantly,
but the mixer may instead calculate a delay based on how far away it is,
and start playing then).
[ sadly, I don't know of a good way to do "realistic" spatial echoes in
real-time, leaving such cheap tricks as "a box which enables echos"...
likely it would involve "some way to calculate a large FIR-filter based
on the sound-source, destination, and intermediate geometry"... ]
the main differences:
all mixer-streams mix directly to the output;
all streams are simple 1:1 filters (no composting or sequencing);
the mixer is mostly just used for playing in-game sound-effects (from
sound files);
the stream feature is thus far currently only really used as a means for
plugging in things like Ogg/Vorbis, MIDI, and Text-To-Speech playback
(feeding samples through streams or layering them isn't really done);
...
probably, if I were to implement something like my idea though, it would
be functionally independent of the 3D engine, and could probably be used
as a standalone mixer (capable of also producing output as WAV files, ...).
other features I would probably include (forgot to mention earlier):
ability to playback particular notes from particular instruments
(probably reusing code from my MIDI synth);
probable inclusion of spatial positioning and attenuation;
potentially: inclusion of or integration with MIDI (say, ability to pipe
streams in as instrument patches, and other MIDI extensions);
...
possible implementation routes (should I pursue this idea) include,
among other things:
(hackishly) augment my existing MIDI library with new features (namely,
stream-based mixing, most likely an API independent of the traditional
MIDI command-stream);
implement a new library, which may depend on MIDI library;
simply copy/paste any relevant code, but otherwise implement a
new/independent library.
the latter 2 options have the merit of reducing likely code-tangle (and
avoid polluting the MIDI synth) with what would otherwise be
functionality unrelated to MIDI, but the cost that they would add
another library to the project. (note: as-is, the MIDI synth does not
implement streams or delay-based mixing, it merely mixes directly and
operates according to time-slices. so most of this would be all new code
anyways).
or such...
_______________________________________________
fonc mailing list
[email protected]
http://vpri.org/mailman/listinfo/fonc
