(A message [below] from my sponsor [well, supervisor])
Dear all
I think this might be of interest to some of you. The PD port is stalled, but now that the code is public, let's make it happen ;-)
pa
==
(and here's some info about the project)
The HISSTools Impulse Response Toolkit, a set of Max objects to capture and manipulate impulse responses, is now available in version 1.1 at http://eprints.hud.ac.uk/14897/ Apart from a new object and few bug fixes, the major news is the support of both 32 bit and 64 bit versions of Max, on both platforms (MacOS and Windows). The externals are still SSE optimised and all other goodies.
We are also releasing as promised all the source code under modified BSD license, available now at: https://github.com/HISSTools/HISSTools_Impulse_Response_Toolbox
There is also a second video tutorial available, this time on electronic acoustic room correction https://vimeo.com/79202287 http://www.youtube.com/watch?v=QXOJWv3mvrc
Finally, there was a presentation at IRCAM of the musical methodologies that were investigated so far within the project. These conclusions are available in a new paper available here: http://eprints.hud.ac.uk/18549/
We hope you will find good use of these resources.
Alex and P.A.
I was passed this info just before the new year but thought it would get lost amidst the general festivities and may also appear to be a bit, well, last year when we've all got our heads around getting back into stuff.
(some background info) About a year or so ago Katja and I put out a call to the list and forum to see if there was any interested parties willing to help port these from Max to Pd. There wasn't, and that's fine of course.
In the meantime Alex (Harker) & P.A. have done a fairly sterling effort in separating the source code out into general dsp stuff and the Max specific, aiding greatly in any port attempt.
My memory is that we had planned to at least port *something* from the library hoping that that may pique some interest in our community when people got some idea of the many creative possibilities opened up by having these objects available in Pd - but we haven't as of yet.
To give some idea of the task as I see it: it took the four of us that ported one object ([ipoke~]) from Max to Pd almost 4 months of occasional work so to do this on a similar timescale - a decade!
In fact this is a quote from Katja when I wrote to her after receiving the above email:
"HISSTools has all the elements a live performer can use to improve sound quality. As such it is a kind of 'standard work' that would ideally be available for every real time DSP framework. I am to an extent familiar with all topics covered (fft, fast convolution, filter inversion, minimum phase conversion). I know how to read and write C and had already expressed my interest to be involved in a Pd port. But because of the complexity I recoiled, and do not want to make commitments. It is way too much for a hobby programmer who incidentally writes a Pd external in free time. The port should probably be initiated by a professional, or an advanced DSP student, in the context of an institution like a university."
Before even starting on the various classes the job would probably start with designing a cross-platform build system for the stuff. It makes no sense to use Xcode or Visual Studio for a Pd project. HISSTools has quite a few dependencies and should probably be done with an autotools build system (configure script, libtool, etc).
It's a big job - no question. And if someone, who (for me) is one of the heaviest Pd coders I know, baulks at the sheer size of the task then I'm in no real position to be doing much more than cheering from the sidelines.
Certainly I'm way too far down the post-grad line to be thinking of taking something like this on as a PhD type project but perhaps there's someone who may be interested? I will certainly do all I can to try and instigate some institutional support from Huddersfield (I think there's some recent post-grad fee-waiver opportunities for example) and I'm sure Prof P.A. (who can actually do something about it) would be keen. Or perhaps there is a possibility of a joint institutional project or something - I don't know really.
Anyway - it's out there, it's probably doable, and there's nothing yet in Pd that does this kind of stuff in one nicely coded bundle.
Perhaps we can at least discuss it amongst ourselves?
Regards and best wishes to all for the New Year,
Julian