Pd-list December 2002

pd-list@lists.iem.at

63 participants
88 discussions

Re: [PD] Gem -0.87 pix_mask & pix_film - was Performance Tuning (disk IO)
by bbogart＠ryerson.ca 08 Dec '02

08 Dec '02

Hello All, I was thinking about using movies rather than jpegs to store my 800 odd image files, but I've just remembered I'm using pix_mask to make the text nice and antialiased. So I did a few trials using pix_film and pix_mask. It does not seem to work at all, is it supposed to? This is on the standard gem-0.87 distro. I can load movies into pix_film attached to pix_mask but the masking is not rendered, and when I change frame I get the message: error: Gem: pix object cannot handle gray image Its not clean which pix object is creating this message, and neither movie is grey!!! (I tried with my own movies and the mov that comes with gem) I'm using the very same patch I used to set up the alpha using jpegs so alpha and all the other stuff one needs is there. This functionality is in the cvs gem? Thanks Ben ----- Original Message ----- From: guenter geiger <geiger(a)xdv.org> Date: Saturday, December 7, 2002 11:08 am Subject: Re: [PD] Gem -0.87 Performance Tuning (disk IO) > > > Hi Ben, > > you might try to convert the image into a movie. > This could be done on Linux with ffmpeg for example. > > On the Gem side image loading could be put into a additional > thread to prevent such problems. > > Guenter > > -------------------------------------------------------------------- > ---- > Music Technology Group - UPF, Barcelona <" > target="l">http://www.iua.upf.es/mtg/eng/>AGNULA - A GNU Linux > Audio distribution <www.agnula.org> > DeMuDi - Debian Multimedia Distribution <www.demudi.org> > > > On Sat, 7 Dec 2002 bbogart(a)ryerson.ca wrote: > > > Hello All, > > > > I'm working on the implimentation of my thesis > > project in pd/Gem and have run into some performance > > issues (in linux using -rt). > > > > I have 4 video clips (Photo JPEG 256x256 quicktimes > > from 3000-11000 frames) and about 800 128x512 jpeg > > stills. > > > > any two of the 4 video clips are playing at any one > > time, and being mixed using alpha. > > Three of the jpegs are being used at any one time. > > The jpegs are chosen by the output of some chaos math. > > > > The video stuff playes great on its own, and the > > stills come up great too, animation is smooth and > > nice (Geforce2 MX-400) > > > > The problem is that when I load a new set of jpegs > > the video playback freezes for about 500ms and then > > continues. It looks like its the disk access causing > > the issue, since the videos are being played off > > disk. I tried putting the videos on a different HD > > than the images, but there is still that pause. > > > > Anyone have any suggestions on what I could do to > > improve the performance? This is an installation and > > as such needs to run for days on end without a > > performance drop. > > > > In my tests using the images without the video I > > could open a new jpeg every 80ms or so, very > > impressive. But why the video render pause? I > > thought about using pix_multiimage, but 800 128x512 > > tiffs is probably too much for my 640MB of RAM. > > > > Any suggestions would be appreciated. > > > > Thanks > > Ben > > > > > > _______________________________________________ > > PD-list mailing list > > PD-list(a)iem.kug.ac.at > > http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list > > > >

1 0

crossfade of complete sample buffer
by Orm Finnendahl 07 Dec '02

07 Dec '02

Hi everyone, before hacking it up myself I thought, I'd ask if anything like that already is realized or could be implemented in an abstraction using standard pd objects (I could imagine it's somehow possible, but couldn't find a way). I need to gradually transform the contents of a sample buffer from one sound to another, by crossfading all of the samples from sound A to sound B in a given time (say 2000 msecs). What I look for is an in place change of the sample values in the array. The lengths of the two sample buffers of course are considered equal. Not to get me wrong: the cross fade doesn't work from left to write in the sample buffer, but all values of the buffer are crossfaded by crossfading the samples from sound A at all the indexes with the respective values of sound B "at once", using subsequent "steps", which are completely independent of the sample frequency to perform the cross fade, much like a line object. Any ideas? -- Orm

4 16

Gem -0.87 Performance Tuning (disk IO)
by bbogart＠ryerson.ca 07 Dec '02

07 Dec '02

Hello All, I'm working on the implimentation of my thesis project in pd/Gem and have run into some performance issues (in linux using -rt). I have 4 video clips (Photo JPEG 256x256 quicktimes from 3000-11000 frames) and about 800 128x512 jpeg stills. any two of the 4 video clips are playing at any one time, and being mixed using alpha. Three of the jpegs are being used at any one time. The jpegs are chosen by the output of some chaos math. The video stuff playes great on its own, and the stills come up great too, animation is smooth and nice (Geforce2 MX-400) The problem is that when I load a new set of jpegs the video playback freezes for about 500ms and then continues. It looks like its the disk access causing the issue, since the videos are being played off disk. I tried putting the videos on a different HD than the images, but there is still that pause. Anyone have any suggestions on what I could do to improve the performance? This is an installation and as such needs to run for days on end without a performance drop. In my tests using the images without the video I could open a new jpeg every 80ms or so, very impressive. But why the video render pause? I thought about using pix_multiimage, but 800 128x512 tiffs is probably too much for my 640MB of RAM. Any suggestions would be appreciated. Thanks Ben

2 1

Re: [PD-dev] CVS now handles binaries
by Adam Lindsay 07 Dec '02

07 Dec '02

Hi Chris, I'm glad other Mac people are thinking about this. My concern was primarily with packaging binaries of *externals*, which, looking back, wasn't clear in my last email. That changes things, but you bring up a lot of points worth talking about anyway. I hope people don't mind that I've cc'd this message to the general pd-list, since it's a bit more, uh, general. On Friday, December 6, 2002, at 06:31 PM, chris clepper wrote: >> I've been giving some thought to that on the OSX side, mostly with >> regards to the packaging system. I'm not sure what's the most >> convenient >> way to make installable packages on the MacOSX side of things. Apple's >> .pkg is attractive, but the current package format doesn't seem to be >> very manipulable by the command-line. > > why even bother with a packaging system for the binary? how about > this for an install on OSX: > > step one: download binary > step two: un-tar/zip/stuff downloaded file > step three: open pd folder and double click on CLICK_ME_TO_RUN_PD > smiley face icon > for step three to happen, write a pdstart.command that has ./pd > -options and put it in the pd/bin dir. make an alias to it with that > big smiley face icon at the root of the pd folder. Can you believe I hadn't considered that? Yes, that would be a nice, friendly, easy way of approaching things. > i don't think you can assume that the typical Mac user uses the > command line at all, so OSX apps should take that into account. the > above approach seems to be in line with most free/shareware > applications distributed for OSX, which will make Mac users feel right > at home. > > one thing to work out is explaining how the runtime options are used. > this could be a point of confusion for people when they first try to > setup their midi/audio and add various libs to pd. But so my interpretation of the "Mac OSX way" is slightly different. I see it as a real UNIX, but with all the tricky bits hidden out of sight. Users interact with the nice smiley bits exposed for them, but don't worry about the stuff hidden out of sight. Power users can use the command line all they like. So that's where I was coming from with a standard /usr/local/pd folder. You can put a smiley "Pd start.command" in /Applications, and no one will be the wiser. So yes, the externals, and in particular managing paths, are the tricky bit, which is where I tried to start the discussion. If we can control help paths in a manner similar to the lib/extern path, then we are most of the way there. That way, people can look at something like a /Library/Pd hierarchy for externals, and come up with some automagic way of linking help files and binaries. ( I spent an afternoon this week looking at simple heuristics for figuring out libraries, how help files are linked, and which .pd files are help files and which are not. My conclusion: the last bit is really hard, across the external libraries that exist. So the grand plan of a magic drop Folder with an attached AppleScript and shell script is on hold. bailing wire and glue.... ) >> Also--where do they go? The /usr/local tree (where I personally >> prefer to >> put pd) is typically hidden from the user. It would be nice to be >> able to >> put files in a place like /Library/Pd, where an ordinary user has the >> capability to scre^H^H^H^Hmanipulate their installation in the finder. > > i think you have answered your own question. the OSX version of pd > should go where the user wants it to go. if you install to /usr/local > that would make pd vanish to the majority of mac users! Well, none of the stuff in the main distribution folder is of any use to a normal mac user... almost nothing to view, and nothing to run from the finder. But a friendly pd.command in a prominent position? That does everything. I also advocate a semi-standard place for the main distribution binaries because Pd isn't necessarily to be run directly by the user. :) Externals are a different matter. A user probably wants to at least see her "plug-ins" in a user-visible folder. I'd like to accommodate that while putting links to help and libraries in the "right" places automatically. >> We could emulate your debian approach--how does that lay out files for >> externals (and their associated source, help, and other files)? > > why emulate debian at all? it's not debian it's OSX. if anything it > should more closely resemble the win32 version. That's a point. But I was asking more about the debian packages in terms of how the externals are approached: is everything from an external kept in a single folder, with symbolic links spread everywhere? Are the actual binaries and helpfiles moved? Best regards, adam

1 0

pd 0.36 for IRIX o32/n32 and externals
by Soren Bovbjerg 06 Dec '02

06 Dec '02

Hi Miller I was running the n32 version. The reason it didn't crash was because I didn't compile with -DDL_OPEN so it never really defined 'makeout'. Anyway, i recompiled this morning, and it crashed like I guess it is 'supposed to' ;-) I also made a o32 binary which i have added to the 0.36 package at http://www.cvmt.dk/~sb/ Externals compile and run with this version. Is the n32 crash an old problem? Cheers Soren Miller Puckette wrote: > Hi Soren, > > Are you rnning the "o32" version of Pd? I believe the "n32" version was > never able to load externs, it just crashed. So now what else could be > going wrong, I wonder? If anyone has any ideas, I'd be happy to hear of > it. (I can't test this myself since all our SGI machines are broken.) >

2 1

ggee Windows Binaries
by Jordan J 06 Dec '02

06 Dec '02

Hello, After scouring google and of course pure-data.org all I ever came up with was an archive of the source code. Is there a compiled version out there? Thanks for any help. -Jordan

3 2

Filtering, hmmm
by matthew jones 06 Dec '02

06 Dec '02

thanks marius, dave. good ideas. just one extra thing (unrelated) how do you implement an FIR/IIR filter? do you need a 'delay-by-one-sample' patch or is there a more direct route? also, do you have to have a blocksize of one sample?? for an adaptive filter it would be nice to update the filter every new sample, and also use it every sample coming in (I'm guessing with a blocksize of, say, 64 that you only get access to the inputs every 64 samples??) (hope I'm not filling everyones inbox too rapidly) M

2 1

adc~ input (??)
by matthew jones 06 Dec '02

06 Dec '02

anyone got any clue what is (and how to use) the adc~ input. Any way of telling it what channel to use dynamically? cheers

3 2

Syntax (was "Pack Problems")
by Jordan J 06 Dec '02

06 Dec '02

Okay, I have finally solved my problem. The explanation about syntax didn't kick in until just a few minutes ago. Here's what I can summarize. I now realize that if [soundfiler] receives "$0sample" it will see that as a string and therefore it would be "0sample" and that is NOT a valid array in this case. Instead I have to tell it that $0sample is a symbol and not a string by sending "symbol" along with it. Cheers, and I *really* appreciate the help. Not many other people are awake at 3:30 am central time :) -Jordan

2 1

sending MTC messages over midi
by Eldad Gwetta 06 Dec '02

06 Dec '02

What is the right way of sending midi clock messages to an external hardware? Thanks Eldad

1 0

← Newer
1
2
3
4
5
6
7
8
9
Older →

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 December 2002