how about the "tanhauser" (heavy audio tools now) compiler, that can compile patches to C code, and then, on another step on this chain, we could make it an external?
looks like a window into a gen~ like idea, but I might be far off
cheers
2016-02-24 21:59 GMT-03:00 Jonathan Wilkes via Pd-list <pd-list@lists.iem.at
:
It sounds very difficult, but I imagine gen~ does something like that.
I don't think the payoff is big enough to justify the development, unless what gets compiled are good old signal and/or control object chains that everybody is already familiar with.
That's what happened with Javascript engines. It's quite extraordinary to see what can be done in the browser be people who have no idea what V8 is doing under the hood. At the same time the people who are experts are building high-quality frameworks in a few years that would have previously taken a decade to develop. And I can play old arcade games inside my browser. Everybody wins.
One the other hand, that's pretty damned complex and expensive development. Outside of trivial cases like a chain of unary op signal objects with single connections, I have no idea how one would optimize Pd patches, much less on the fly. But surely the first step is better realtime analysis tools, so we can quickly know where the bulk of the CPU time is spent.
-Jonathan
On Wednesday, February 24, 2016 6:18 PM, Brian Fay < ovaltinevortex@gmail.com> wrote:
The issue with blocking is that you don't have fine-grained control of an audio, process, right? If you want that level of fine-grained control, you either need to explicitly set the blocksize to 1 in your patch/sub-patch, or you need to actually dig into the C code for the externals and change the logic there (which requires you write C code, recompile things, restart Pd...).
Theoretically, there could be a way to swap out the code for an external while Pd is running, rather than restarting Pd. Still, you'd have to write C and pray that you didn't introduce some terrible bug.
But theoretically, instead of writing externals in C, couldn't we come up with a high-level representation of a Pd external in a visual programming environment similar to Pd? Then we could compile that down to an external, and actually use it in our Pd patch, without actually reloading Pd. It sounds very difficult, but I imagine gen~ does something like that.
For reference, I believe Extempore provides the ability to edit and replace a low-level audio process while the program is running (I'll have to rewatch some conference videos to confirm that).
On Wed, Feb 24, 2016 at 3:41 PM, Matt Barber brbrofsvl@gmail.com wrote:
OK, now I'm having trouble even imagining how an unblocked audio model could possibly behave (at least, as David points out, in a real-time context).
On Wed, Feb 24, 2016 at 2:58 PM, David Medine dmedine@ucsd.edu wrote:
This doesn't answer Matt's question at all (apologies), but just as a clarification, ChucK *does *block audio. It's just that ChucK always blocks at the minimum size of 1 sample per block. 1 is still a block size though, and it still implies the same problems associated with order of operations, feedback, interpolating control input, and parallelization that a block size of 64 does.
Also, maybe this has already been pointed out on this thread, but block 1 is super slow because it means that you have to load all your DSP functions onto the CPU cache every 1/SR seconds instead of 64/SR seconds. Blocking by 64 buys a lot. Having a locally adjustable block size is a great feature (that ChucK lacks) because you can do it for special needs cases (like variable delay patches, for example).
Anyway, in my opinion, the block thing isn't a limit to Pd, but a limit to real-time digital signal processing.
On 2/24/2016 11:27 AM, Matt Barber wrote:
Are there any other DSP environments besides ChucK that don't block audio? Last time I tried ChucK (2012?) its efficiency was still abysmal. [block~ 1] definitely takes a hit, but it's usually possible to minimize how much of the DSP chain is actually blocked at 1. I guess with Csound you can specify a k-rate equal to the sample rate which is also effectively a single sample block. I haven't ever used Csound in a real-time context, and most of what I do with it compiles much more slowly than real time in any case.
On Wed, Feb 24, 2016 at 1:44 PM, peiman khosravi <peimankhosravi@gmail.com
wrote:
You can do this with MSP's poly~ too but I'm guessing that the CPU costs are quite heavy. Moreover, there are operators in gen that are designed for low-level operations.
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk/ *
On 24 February 2016 at 16:15, cyrille henry < ch@chnry.netch@chnry.net> wrote:
Le 24/02/2016 16:50, peiman khosravi a écrit :
One great advantage of maxmsp is gen, which gives you sample-level patching with the possibility of a one-sample delay.
you can use [block~ 1 1 1] in a pd subpatch.
cheers c
P
On Tuesday, February 23, 2016, Samuel Burt <composer.samuel.burt@gmail.com mailto:composer.samuel.burt@gmail.com> wrote:
David, One thing I attempted and couldn't find a solution for was thefollowing, mostly owing to the limitation of interfacing with a 64 sample block size.
I wanted to have a directory of hundreds of audio recordings. Each onewould be a single wavelength from an interesting sound, like a bass clarinet, marimba, harpsichord, tambourine, etc. Each would begin and end at a zero crossing so you could chain them together to make complex timbres. They could be chained in sequence, randomized, or loaded in meta-data-matched chunks. I ran into a problem figuring out how to trigger the next sound based on the ending of the last sound in a sample accurate way. Sound file loading or even buffer playback triggering waits until the start of the next block size before it updates. If you have a waveform that lasts 205 samples (64+64+64+13), you have a gap of 51 silent samples before the next waveform would start. Not only do you not get the continuous sound you want, this winds up creating a periodic pattern with a frequency of 689 Hz (44100/64).
David, I like your idea "what (if anything) someone tried to do in Pd,but couldn't given its limitations". I think this could be a wonderful challenge if we could have a monthly thread like this where the best minds among us come up with solutions to some of the hardest conceptual challenges in Pd.
I'm still struggling with loading dozens of files, audio dropouts, andother similar problems. Someone else expressed frustration about Pd's single-threaded status. I too have feared upgrading my computer based on the limitations of current multicore processors (although realistically I think we can all look at the "turbo-boost" level or whatever Intel calls it to determine where our processor might run with a demanding patch. I understand the fact that you can't run your audio process on multiple cores, because it is a linear process. It would be great if the GUI could run on a second core, a process that loads audio into memory could run on third core, while GEM could automatically run on a fourth core. I don't have any concept of how feasible that would be, though. Does the GUI in pd-l2orc run on a separate core?
Sam Message: 4 Date: Tue, 23 Feb 2016 09:01:06 -0800 From: david medine < <dmedine@ucsd.edu>dmedine@ucsd.edu <javascript:_e(%7B%7D,'cvml',' dmedine@ucsd.edudmedine@ucsd.edu');>>
One thing I'd be interested in knowing about is what (if anything) someone tried to do in Pd, but couldn't given its limitations(apart from look/feel/convenience issues).
--
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk < http://peimankhosravi.co.uk/miscposts.rss%3E< http://spectralkimia.wordpress.com/http://spectralkimia.wordpress.com/%3E*
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list http://lists.puredata.info/listinfo/pd-list
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list http://lists.puredata.info/listinfo/pd-list
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list http://lists.puredata.info/listinfo/pd-list
_______________________________________________Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list