Hi,
As a student who is *very* interested in taking part in the PD project under Google Summer of Code, I thought I'd throw a few ideas out and see what other people thought of them.
These are all things I'd be interested in pursuing.
- A set of externals designed for real-time, high-level analysis of control
data. This could include things like histograms, rhythm analysis, rate-limiting and data-filtering objects, all designed for use on a constant stream of input.
- A set of externals designed to make string assembly and parsing much
easier, particularly geared towards the creation of OSC messages.
- Finally, a set of band-limited oscillators, to allow for spectrally rich
waveforms without aliasing.
Any thoughts?
Thanks everyone,
Greg
These are all cool proposals Greg.
On Wed, 12 Mar 2008 20:24:33 -0500 "Greg Surges" surgesg@gmail.com wrote:
- A set of externals designed for real-time, high-level analysis of control
data. This could include things like histograms, rhythm analysis, rate-limiting and data-filtering objects, all designed for use on a constant stream of input.
There is some overlap with "mapping" library, and most if not all have been attempted as abstractions rather than externals. Though there's plenty of room for new data analysis objects and this proposal seems to have nice bite sized chunks (contains many independent objects that could stand alone) which makes it attractive for this GSoC requirements (discourage over-ambitious monolithic projects)
- A set of externals designed to make string assembly and parsing much
easier, particularly geared towards the creation of OSC messages.
Strings have always been a problem in Pd. Puredata has a deep traumatic wound since its childhood AFAICS, because Millers position has been that it didn't need strings and they could always be "implemented on top". Several others like Bryan Jurish have made string libraries, and so you would find a good mentor. But beware this isn't as simple as it seems on the surface because of the lack of an intrinsic string type. This proposal has merit in that it might challenge entrenched and stuck thinking, because strings have run us into cul-de-sacs before and maybe a fresh vision would push that forward.
- Finally, a set of band-limited oscillators, to allow for spectrally rich
waveforms without aliasing.
Again, much has been done before and there's no shortage of BLOSC abstractions and some externals. Unifying these into an efficient and coherent set would be a good project, as would porting existing optimal implementations from Csound or other open sources.
a.
On Mar 12, 2008, at 10:00 PM, Andy Farnell wrote:
These are all cool proposals Greg.
On Wed, 12 Mar 2008 20:24:33 -0500 "Greg Surges" surgesg@gmail.com wrote:
- A set of externals designed for real-time, high-level analysis
of control data. This could include things like histograms, rhythm analysis, rate-limiting and data-filtering objects, all designed for use on a constant stream of input.
There is some overlap with "mapping" library, and most if not all have been attempted as abstractions rather than externals. Though
there's plenty of room for new data analysis objects and this proposal seems to
have nice bite sized chunks (contains many independent objects that could
stand alone) which makes it attractive for this GSoC requirements (discourage
over-ambitious monolithic projects)
It might makes sense to split up some of the mapping objects into
more specific libraries. It's getting pretty big these days. I
suppose this could be a GSoC project as well.
- A set of externals designed to make string assembly and parsing
much easier, particularly geared towards the creation of OSC messages.
Strings have always been a problem in Pd. Puredata has a deep
traumatic wound since its childhood AFAICS, because Millers position has been
that it didn't need strings and they could always be "implemented on top". Several others like Bryan Jurish have made string libraries, and so
you would find a good mentor. But beware this isn't as simple as it seems on
the surface because of the lack of an intrinsic string type. This proposal has
merit in that it might challenge entrenched and stuck thinking, because strings
have run us into cul-de-sacs before and maybe a fresh vision would push that forward.
At the Pd meeting at LAC in Cologne, IOhannes talked about
incorporating a mechanism for registering any generic type, like
Martin Peach's strings, or data for Gem, PDP, or gridflow. This
could make a good GSoC project for someone who wants to get into the
depths of Pd.
- Finally, a set of band-limited oscillators, to allow for
spectrally rich waveforms without aliasing.
Again, much has been done before and there's no shortage of BLOSC
abstractions and some externals. Unifying these into an efficient and coherent
set would be a good project, as would porting existing optimal
implementations from Csound or other open sources.
I second this, it would be great to have a complete, clean library of
bandwidth-limited oscillators for Pd. It would be a good project too.
.hc
a.
-- Use the source
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
Computer science is no more related to the computer than astronomy is
related to the telescope. -Edsger Dykstra
On Wed, 2008-03-12 at 20:24 -0500, Greg Surges wrote:
Hi,
As a student who is *very* interested in taking part in the PD project under Google Summer of Code, I thought I'd throw a few ideas out and see what other people thought of them.
[...]
- Finally, a set of band-limited oscillators, to allow for spectrally
rich waveforms without aliasing.
there are already some externals for band-limited oscillators around. search the svn repository for blosc. there is [sqosc~] from mrpeach, which is a bandlimited pulse generrator and there is also a small set of absrtactions in pdmtl:
[synth.osc.saw.bl~] bandlimited saw [synth.osc.square.bl~] bandlimited square [synth.osc.triangle.bl~] bandlimited triangle
(this last set is built almost purely with internals. they use the tof external [destroysend]. they will work even without it, as long as you don't delete one of those abstractinos.)
roman
Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de
Hallo, Roman Haefeli hat gesagt: // Roman Haefeli wrote:
there are already some externals for band-limited oscillators around. search the svn repository for blosc. there is [sqosc~] from mrpeach, which is a bandlimited pulse generrator and there is also a small set of absrtactions in pdmtl:
[synth.osc.saw.bl~] bandlimited saw [synth.osc.square.bl~] bandlimited square [synth.osc.triangle.bl~] bandlimited triangle
(this last set is built almost purely with internals. they use the tof external [destroysend]. they will work even without it, as long as you don't delete one of those abstractinos.)
There also is [trapez~] for any kind of waveform with jumps in it, like saw, square, multisawquare, etc.: http://lists.puredata.info/pipermail/pd-list/2007-12/057134.html Foldover protection is realised according to The Book, cornered waves are not handled, only dependency: Miller Vanilla Pd.
Ciao
Frank Barknecht _ ______footils.org__
-
Andy Farnell -
Frank Barknecht -
Greg Surges -
Hans-Christoph Steiner -
Roman Haefeli