Pd-list April 2010

pd-list@lists.iem.at

143 participants
284 discussions

removing objects causes recomposing of the dsp graph
by Mathieu Bouchard 09 Apr '10

09 Apr '10

> There seems to be a (linear?) relationship between the number of > objects(connections?) in a patch and the time it takes for the dsp graph > to be redrawn. At least I think the drop outs are related to the dsp > graph redrawing because of the following observation: This is because Pd doesn't wait for the next DSP tick to recompile... It doesn't even wait for the end of the current logical instant (imagine [delay 0]). Therefore it has to redo the job for each deletion. Also, nothing is done in the form of a diff, so, Pd has to recompile the whole graph each time, nothing like incremental compilation. This can often lead to O(n^2) running-times. > Is there a way to 'manually' control the redrawing of the dsp graph and > also to suppress it, when not necessary? you are already doing it, using «dsp 0» and «dsp 1» messages. just make sure you are not turning on a dsp that was already off before you deleted. You do that by using [r pd] with [route dsp] and storing that value in a [f]. then canvasdelete.c could also be making a clock_delay(self,0) to resume the dsp a bit later, but I wonder whether that's actually tricky to do. I can say, at least, that it's dangerous to make this delay any longer than 0, because the dsp could be using objects that don't exist anymore. _ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801

2 2

Displaying as two digits
by Mathieu Bouchard 09 Apr '10

09 Apr '10

Roman Häfeli wrote : > Now you made it worse than it really is. Actually, it will end up as: > 1.23457e+06 or 1234570. It's true though, the reformatting truncates > significant digits. Actually, the only way I can think of to store > 32bit-float numbers with full precision in Pd is to write to > 32-bit-float soundfiles. How about using [makefilename %.8g] ? Together with the current bugs in atom_string, those symbols will be saved as full-precision floats that Pd will understand. %.6g gives 6 decimals of precision = max 19.93 bits %.7g gives 7 decimals of precision = max 23.25 bits %.8g gives 8 decimals of precision = max 26.57 bits to get full precision you need %.8g, because in float32, you have 24 bits of precision, which is why 16777215 is an exact float32 number, for example, while 16777217 is not. _ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801

3 3

[PD-announce] Thirty-fourth meeting of the PureData users group of Montréal (pdmtl#34)
by Mathieu Bouchard 09 Apr '10

09 Apr '10

la trente-quatrième rencontre du club d'utilisateurs de PureData de Montréal aura lieu à Vidéographe Production, 4550, rue Garnier (coin Mont-Royal) le mercredi 14 avril 2010, à compter de 18h30. au programme : | | .-o----------------------o-. | Alexandre~ T Porres~ 440 | `-o------------------------' | ---o--- 3ª Convenção Internacional de Puredata em São Paulo The 3rd international Pd Convention was held from 19th to 26th of july 2009 in São Paulo, venues took place at 4 SESC units, the MIS-SP (museum of Image and sound of Sao Paulo), and PUC University. We had the support of MIS, Sesc, and the CNPq agency as the major funders. We had around 50 participants from outside of Brazil, plus 20 brazilians, and the format was closely related to the previous edition: Performances, Installations, Paper Sessions, Demo Sessions, workshops, and a Hack Lab... I had the will and idea to organise this convention when I came to PdCon07 in Montreal, in 2008 Mathieu Bouchard and Alexandre Castonguay went to Brazil, and we had a small gathering that served as a way to check the local interest and join forces. It was organized and carried with no funds at the place I was living (Ibrasotope), which also works as an artist run center. That event was organized along with ricardo Palmieri, who was my partner in the organization of PdCon09 along with Paloma Oliveira, who later joined and helped with the production team. Many other people helped, of course. I will show some stats about the convention and show some videos and discuss its impact and further consequences in the local scenario, and worldwide. ---o--- | | | .-o---------o--------o-. | pix_Marc Fournel 180 | `-o--------------------' | ---o--- VOID 0.2 Présentation de la documentation et survol de la programmation de VOID 0.2, une installation interactive qui utilise la rétroaction vidéo pour créer un jeu de déconstruction et reconstruction visuelle en temps réel d'une source vidéo provenant d'une caméra miniature mise à la disposition des utilisateurs. VOID 0.2 intègre aussi trois interfaces sans fil qui, lorsqu'elles sont manipulées, modifient différents éléments de la composition visuelle finale. VOID 0.2 est une expérience sensorielle créée pour être présentée dans un dôme immersif (écran de 360 X 180 degrés). La première diffusion de cette installation a eu lieu en mars 2010 à Casablanca (Maroc). VOID 0.2 est issue d'une résidence de recherche effectuée au centre i-DAT situé à Plymouth (Angleterre). ---o--- | | .-o--------------------------o-. | #Mathieu #Bouchard (255 0 0) | `-o--------------------------o-' | | ---o--- `------- Je présenterai mes nouvelles patchs réalisées à l'aide de GridFlow (et d'un peu de GEM). Je montrerai dither.pd et rgb_delay.pd, qui sont deux exemples qui sont nouveaux dans GridFlow 9.8, ainsi que supercycloid.pd, un exemple qui sera nouveau dans GridFlow 9.9. Je montrerai plusieurs variations sur chacune de ces patchs, je parlerai un peu de leur fonctionnement, et surtout, je vous montrerai ce qui en résulte :) Je montrerai aussi comment générer de gros polygones GEM à partir de GridFlow. j'utiliserai un exemple à 480 sommets, et ça n'impliquera PAS un [polygon] à 481 inlets (!) : on pourra même varier le nombre de sommets dynamiquement. ---o--- | | . . . à bientôt ! _ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801 _______________________________________________ Pdmtl mailing list Pdmtl(a)lists.artengine.ca http://lists.artengine.ca/cgi-bin/mailman/listinfo/pdmtl _______________________________________________ Pd-announce mailing list Pd-announce(a)iem.at http://lists.puredata.info/listinfo/pd-announce

1 0

default font for pdp_text and pdp_qtext
by Hans-Christoph Steiner 09 Apr '10

09 Apr '10

So now that Bitstream Vera seems to have been purged from Debian and Ubuntu, I think that the PiDiP text objects should use a different font as the default. The GNOME default font is now DejaVu, so that probably makes sense to use. Otherwise, unless people manually install Bitstream Vera from outside of Debian/Ubuntu, they'll get this: PiDiP : additional video processing objects for PDP version 0.12.24 ( ydegoyon(a)free.fr ) error: [pdp_text] error: could not load default font, no text rendering! install Bitstream Vera, it's free! (http://www.gnome.org/fonts/) error: [pdp_qtext] error: could not load default font, no text rendering! install Bitstream Vera, it's free! (http://www.gnome.org/fonts/) .hc

1 0

framp~
by Matteo Sisti Sette 09 Apr '10

09 Apr '10

Hi, The intro-help.pd patch says: [framp~] - output a ramp for each block but that's not what it does. Is there an object that really does that? By the way I don't really understand what framp actually does. It says it "estimates" amplitude and frequency of "any component", but it also says that a sinusoidal component will appear in four components, so it seems like it simply computes the amplitude (i.e. modulus) of the spectrum on the right outlet, and on the left outlet: { the frequency corresponded to each bin if amplitude>0 0 otherwise Which is indeed _almost_ a frequency ramp, but with zeros here and there.... Or is there something more in the "estimation" of the components? If not, I think it would be much more useful (and even elegant I would say) to "split" [framp~] (which has a misleading name and a misleading description in intro-help.pd) into two objects: - one [???~] that would compute amplitude and _phase_ (instead of frequency that does not need to be computed at all), which can be seen as converting complex numbers to polar coordinates (modulus and angle?) - one [blockramp~] that would give a ramp of integers at each ramp (i.e. 0 to 63 in the default case of 64-size blocks) no??? The first could be done by abstractions, however it is a bit strange that you have an object (framp) that computes the amplitude and you don't have one that computes the phase...... I'm not sure the second can be done with abstraction: could you trust a [bang~] | [0, 64 1.451247( | [vline~] ? Or a: [loadbang] | [1( | [sig~] | | [bang~] | | | [clear( | / | / |/ [rpole~ 1] ? thanks m. -- Matteo Sisti Sette matteosistisette(a)gmail.com http://www.matteosistisette.com

3 2

pd and multi-core processors
by Mathieu Bouchard 09 Apr '10

09 Apr '10

> What I have tried in the past is run one pd for audio and another one > for GEM stuff, which worked rather well. I wonder if it would make sense > to do the same with 2 pd instances doing audio, and exchange audio > between them. Maybe I could try that with Jack. But I think the latency > will be doubled, because the buffers of both pd's would add up...? might depend on whether you are chaining effects (série) or adding effects (parallèle). If what you have can be summarised as two subpatches both connected to the same [adc~] and the same [dac~], then you can put each subpatch in a different pd instance, and patch the rest with JACK, is that right ? _ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801

1 0

isosonic curves : level the perceptual loudness of different pitches
by Mathieu Bouchard 09 Apr '10

09 Apr '10

> This was first presented in the Second Puredata International > Convention, in which I thank Mathieu again as well as the other people > involved in the production of it - the event was basically a turn point > in my life for the better :) But the work I presented then seems to > have disappeared from the web... that one is in english, and is still a > good starting reference to the things I'm working with in my phd > research. I can send over request. do you mean http://artengine.ca/~catalogue-pd/28-Porres.pdf ? This paper has been there since 22 août 2007 and hasn't moved 2,54 centimetres. _ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801

2 2

removing objects causes recomposing of the dsp graph
by Roman Haefeli 08 Apr '10

08 Apr '10

Hi all I found that doing (officially unsupported) dynamic removal of objects causes a redraw of the dsp graph. Especially in big patches this can lead to audio drop outs. There seems to be a (linear?) relationship between the number of objects(connections?) in a patch and the time it takes for the dsp graph to be redrawn. At least I think the drop outs are related to the dsp graph redrawing because of the followig observation: * [7(-[until]-[delete(-[s canvasname] requires 700ms (on my computer on the patch I'm working) * [delete(-[s canvasname] requires 100ms * [dsp 0(, [7(-[until]-[delete(-[s canvasname], [dsp 1( requires also only 100ms It seems, that no time is used as long dsp is turned off, and only on turning it on again, the dsp graph is redrawn (cpu time is consumed). Interestingly it requires very little time to create message-only objects, which makes me think that after dynamically creating an object the dsp graph is not redrawn. Why does deleting an object triggers it, even it was not triggered on creation time of that very same object? Is there a way to 'manually' control the redrawing of the dsp graph and also to suppress it, when not necessary? Roman

1 0

[PD-announce] Workshop May 1, 2010
by Barry Moon 08 Apr '10

08 Apr '10

Workshop: Max/MSP and Pure Data for Interactive Music Arizona State University West Campus Saturday May 1, 10am-5pm Cost: $50 This is a one-day workshop led by Andrew May and Barry Moon. May is a composer specializing in real-time computer processing of acoustic sources. His regression analysis externals for Max/MSP and Pure Data can be found on his website: http://cemi.music.unt.edu/may/ May is an associate professor at University of North Texas. Moon is a sound/video artist creating works for a variety of audio and visual media. His Baz Tutorials can be found on YouTube: http://www.youtube.com/BazTutorials Moon is an assistant professor at Arizona State University. Topics covered in the workshop include: Audio analysis methods Audio processing, including effects, synthesis, and sampling MIDI, games controller, and other user inputs Video analysis for control of audio processing Setting up a live show (microphones, mixers, performers, etc.) A CD with May's and Moon's software will be handed out for the class, and participants can bring their own computers or use one of the computers in the lab where the workshop will take place. Participants are also encouraged to bring in their own software (patches) for help and feedback from May and Moon. Please contact Barry Moon with questions regarding the event: barry.moon at asu.edu Also, that evening at 6pm in Kiva Lecture Hall there will be a panel discussion and at 7:30pm a concert by the Tornado Project, featuring Elizabeth McNutt, flute, and Esther Lamneck, clarinet, performing live with computer. _______________________________________________ Pd-announce mailing list Pd-announce(a)iem.at http://lists.puredata.info/listinfo/pd-announce

1 0

memento issue
by David Schaffer 08 Apr '10

08 Apr '10

Hi there, I recently upgraded to the latest version of Pd extended and my memento setups don't seem to work anymore... it says it can't create "pool RRADICAL" ... What can I do to make it work again? I'm on Win XP. Thank you. D.S http://www.flickr.com/photos/schafferdavid/ http://audioblog.arteradio.com/David_Schaffer/ _________________________________________________________________ Hotmail: Trusted email with powerful SPAM protection. https://signup.live.com/signup.aspx?id=60969

3 2

← Newer
1
...
19
20
21
22
23
24
25
...
29
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

Pd-list April 2010