Re: [PD] Face tracking
Great ! i love the video, and the patch seems to be kind of what I need, from within PD. Will try to get it work on my mac Thanks ! JM Le 28 janv. 2014 à 23:48, pured...@11h11.com a écrit : i don't know if it's exactly what you need: Make me (opencv / gem): http://puredata.info/Members/patrick ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] cosinesum vs sinesum in arrays
Dear List, not sure if I am getting something wrong here, but it seems that sending the message ; cosinesum array999 1 does something different (besides an obvious phase shift between sine and cosine functions) to the resulting contents of that table: the first partial on cosinesum seems to be the DC component, while the first partial in sinesum is the, well, first partial. Accordingly, the second partials have double frequency in sinesum as in cosinesum. Perhaps this is intentional, but was just wondering... best, Peter ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Face tracking
Is [pix_opencv_facetracker] threaded? If not, I'd suggest using FaceOSC to avoid the 30fps it's locked too when tracking and lower framerate when searching. On Jan 29, 2014, at 6:00 AM, pd-list-requ...@iem.at wrote: From: Antoine Villeret antoine.ville...@gmail.com Subject: Re: [PD] Face tracking Date: January 29, 2014 at 5:17:20 AM EST To: Jean-Marie Adrien j...@jeanmarie-adrien.net Cc: pd-list@iem.at List pd-list@iem.at Hello, there is [pix_opencv] around there : https://github.com/avilleret/pix_opencv This is a fork (a modified copy) of the main sourceforge external repository. This is not included in pd-extended but I've planned to release binaries for MacOSX, Windows and Linux in a not too far future... Concerning FaceOSC, it wroks great and it's now included in [pix_opencv_facetracker], it's pretty experimental since it has not been tested widely. Depending on what you need exactly, you can also use the [pix_opencv_haarcascade] external with some free data set or with your own trained model. Hope this helps. Antoine Dan Wilcox @danomatika danomatika.com robotcowboy.com ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Face tracking
Is [pix_opencv_facetracker] threaded? no it's not If not, I'd suggest using FaceOSC to avoid the 30fps it's locked too when tracking and lower framerate when searching. On Jan 29, 2014, at 6:00 AM, pd-list-requ...@iem.at wrote: *From: *Antoine Villeret antoine.ville...@gmail.com *Subject: **Re: [PD] Face tracking* *Date: *January 29, 2014 at 5:17:20 AM EST *To: *Jean-Marie Adrien j...@jeanmarie-adrien.net *Cc: *pd-list@iem.at List pd-list@iem.at Hello, there is [pix_opencv] around there : https://github.com/avilleret/pix_opencv This is a fork (a modified copy) of the main sourceforge external repository. This is not included in pd-extended but I've planned to release binaries for MacOSX, Windows and Linux in a not too far future... Concerning FaceOSC, it wroks great and it's now included in [pix_opencv_facetracker], it's pretty experimental since it has not been tested widely. Depending on what you need exactly, you can also use the [pix_opencv_haarcascade] external with some free data set or with your own trained model. Hope this helps. Antoine Dan Wilcox @danomatika danomatika.com robotcowboy.com ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] MobMuPlat update
Hi all, MobMuPlat (open-source tools for running Pd patches with custom interfaces on iOS, via libpd) has a significant update with some fixes and new features. Major additions include a 2D graphics widget, an implementation of the LANdini networking protocol (allowing direct network connections, with guaranteed delivery, in addition to multicasting), multichannel audio and selectable audio devices (bluetooth/airplay/etc), compass, device vibration, and more. www.mobmuplat.com new intro vid: https://vimeo.com/85295522 As always, thoughts/bug reports/etc are welcome, feel free to email me off-list. Best, Dan === ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Pd GUI freeze with error message(Tcl)
I've seen these also happen when something tries to address a non-existent widget (e.g. when it is being closed or something similar). There is an easy way to prevent this from ever stopping GUI from working by encapsulating all commands streaming from pd-gui with a simple catch{command}. This is what in part pd-l2ork does and I cannot remember when was the last time I had a problem like this. -Original Message- From: pd-list-boun...@iem.at [mailto:pd-list-boun...@iem.at] On Behalf Of Miller Puckette Sent: Wednesday, January 29, 2014 12:29 AM To: Jonghyun Kim Cc: pd-list@iem.at Subject: Re: [PD] Pd GUI freeze with error message(Tcl) Hi Jong - I think the *(Tcl) INVALID COMMAND NAME' stuff isn't related to Pd freezing, but I'd like to find out how it happens and fix it. The GUI can freeze if Pd is overloaded with audio computation on Mcintosh compuers - the only thing I can suggest is reduce the amount of audio computation (I know - nobody would ever want to do that :) cheers Miller On Wed, Jan 29, 2014 at 04:03:17AM +0100, Jonghyun Kim wrote: Hi list, I don't know why this cause, but sometimes it appears and GUI freeze, but still Audio alive. Only GUI die... Number Box, Vu meter, Sliders, and so on, these GUIs all stop, and don't show the actual numbers... Anyone knows this issue? -- *(Tcl) INVALID COMMAND NAME: invalid command name .x4de9a0.c* *while executing* *$tkcanvas itemconfig $tag -text $text* *(procedure pdtk_text_set line 2)* *invoked from within* *pdtk_text_set .x4de9a0.c .x4de9a0.t395fcb0{98.13}* *(uplevel body line 19)* *invoked from within* *uplevel #0 $docmds* -- Pd-0.45-4 (32bit) Mac OS X Mavericks Screenshot attached. Thanks, Jong ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Pd GUI freeze with error message(Tcl)
On 01/29/2014 01:08 PM, Ivica Ico Bukvic wrote: I've seen these also happen when something tries to address a non-existent widget (e.g. when it is being closed or something similar). There are a few things: 1) Weird treatment of the tk error window in OSX. It looks like sometimes it causes an infinite loop, but I can't predictably generate the error. 2) If I understand it correctly, the tk event loop can run into the same problem the Pd audio dsp scheduler runs into with when you have too many calculations per dsp tick. But it's worse with tk because it isn't optimized for doing timely graphics updates in the first place. If you are sending a constant stream of just enough data to the gui that tk no longer has any idle time to update graphics, then you may get yourself into a situation where tk never updates and you see a freeze. Moreover, tk seems not to use its idle time very efficiently. I built a recursive loop for the search plugin specifically to let the user interact, cancel operations while building a search index. But you can see how inefficient its drawing routine is. (Maybe this has to do with X11 as well, but I think the sluggish redrawing happens on OSX too.) This all is particularly problematic when combined with Pd's architecture, because it's often sending a barrage of messages to the gui. For example: * an error message to the console. The console is fairly snappy when dealing with an email's worth of text, not so snappy when dealing with 100 error messages being sent every 10ms. * moving a large number of objects in Pd-extended or Vanilla * moving an array in Pd-ext or Vanilla * resizing an array where each element of the array is an svg tiger -Jonathan There is an easy way to prevent this from ever stopping GUI from working by encapsulating all commands streaming from pd-gui with a simple catch{command}. This is what in part pd-l2ork does and I cannot remember when was the last time I had a problem like this. -Original Message- From: pd-list-boun...@iem.at [mailto:pd-list-boun...@iem.at] On Behalf Of Miller Puckette Sent: Wednesday, January 29, 2014 12:29 AM To: Jonghyun Kim Cc: pd-list@iem.at Subject: Re: [PD] Pd GUI freeze with error message(Tcl) Hi Jong - I think the *(Tcl) INVALID COMMAND NAME' stuff isn't related to Pd freezing, but I'd like to find out how it happens and fix it. The GUI can freeze if Pd is overloaded with audio computation on Mcintosh compuers - the only thing I can suggest is reduce the amount of audio computation (I know - nobody would ever want to do that :) cheers Miller On Wed, Jan 29, 2014 at 04:03:17AM +0100, Jonghyun Kim wrote: Hi list, I don't know why this cause, but sometimes it appears and GUI freeze, but still Audio alive. Only GUI die... Number Box, Vu meter, Sliders, and so on, these GUIs all stop, and don't show the actual numbers... Anyone knows this issue? -- *(Tcl) INVALID COMMAND NAME: invalid command name .x4de9a0.c* *while executing* *$tkcanvas itemconfig $tag -text $text* *(procedure pdtk_text_set line 2)* *invoked from within* *pdtk_text_set .x4de9a0.c .x4de9a0.t395fcb0{98.13}* *(uplevel body line 19)* *invoked from within* *uplevel #0 $docmds* -- Pd-0.45-4 (32bit) Mac OS X Mavericks Screenshot attached. Thanks, Jong ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] cosinesum vs sinesum in arrays
When sending the cosinesum message to an array, don't you have to specify the number of points too? And specifically a power of two? Plus, for a cosine you have to specify the first partial to have zero amplitude and the second an amplitude of 1 (or less). So the message should be something like [; array999 cosinesum 1024 0 1( First the name of the array, then the cosinesum function, table length and partial strengths. Hope I got your question right.. On Wed, Jan 29, 2014 at 6:01 PM, Peter P. p8...@aol.com wrote: Dear List, not sure if I am getting something wrong here, but it seems that sending the message ; cosinesum array999 1 does something different (besides an obvious phase shift between sine and cosine functions) to the resulting contents of that table: the first partial on cosinesum seems to be the DC component, while the first partial in sinesum is the, well, first partial. Accordingly, the second partials have double frequency in sinesum as in cosinesum. Perhaps this is intentional, but was just wondering... best, Peter ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] get sinusoid from a sine and a cosine oscillator
What you seem to be doing is creating a spectrum which has magnitude 1 everywhere, and the phase is varying at a constant rate vs frequency. That means it has a constant group delay. So... my guess is that you'd get an impulse in each block, whose timing depends on the rate of the phasor. When you vary the phasor frequency, it will coincide with the peak of the hann window at some point and be its loudest. Should be a periodic complex tone. I don't understand your goal: you've got sinusoids in the patch... to generate sinusoids? Chuck On Wed, Jan 29, 2014 at 9:12 AM, Alexandros Drymonitis adr...@gmail.comwrote: Say I have a full sine and a full cosine cycle stored in two tables. I'm trying the following to get a sinusoid from [rifft~] but it doesn't work. [phasor~] | [*~ sizeOfTable - 3] | [+~ 1] |\ | \ | [tabread4~ sine] | \ [tabread4~ cosine] | \ [rifft~] | |[tabreceive~ hann] || [*~ ] | [/~ 1536] I've set the block size to 1024 in this subpatch, and there's a hann window in the parent patch as well. The tables have three guard points, that's why I'm multiplying [phasor~] by the size of the table minus three and then add one. The output of this is a waveform with very low amplitude that kind of bounces up and down within a sine like mask. Don't know if I'm making my self clear. My main question is, how do you get a sinusoid out of a sine and a cosine? Also, what's wrong in my approach? ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] get sinusoid from a sine and a cosine oscillator
Yeah, well I'm trying to create shapes in Gem (say a circle) and create the sound they make. So, to make a circle, I'm making a ramp from 0 to 1, multiply it by 2pi and send it to [cos] and [sin] and store these values in two tables, which I then read for every instance of a [circle] (using [repeat] and [separator]). So, since for any shape, you need two coordinates, x and y, my thought was to use these two coordinates as the real and imaginary part of an FFT, merging the two dimensions in one. After the sinusoid, I'll try to make other shapes too, but I wanted to start from that to make sure that I hear exactly what I see. On Wed, Jan 29, 2014 at 10:30 PM, Charles Z Henry czhe...@gmail.com wrote: What you seem to be doing is creating a spectrum which has magnitude 1 everywhere, and the phase is varying at a constant rate vs frequency. That means it has a constant group delay. So... my guess is that you'd get an impulse in each block, whose timing depends on the rate of the phasor. When you vary the phasor frequency, it will coincide with the peak of the hann window at some point and be its loudest. Should be a periodic complex tone. I don't understand your goal: you've got sinusoids in the patch... to generate sinusoids? Chuck On Wed, Jan 29, 2014 at 9:12 AM, Alexandros Drymonitis adr...@gmail.comwrote: Say I have a full sine and a full cosine cycle stored in two tables. I'm trying the following to get a sinusoid from [rifft~] but it doesn't work. [phasor~] | [*~ sizeOfTable - 3] | [+~ 1] |\ | \ | [tabread4~ sine] | \ [tabread4~ cosine] | \ [rifft~] | |[tabreceive~ hann] || [*~ ] | [/~ 1536] I've set the block size to 1024 in this subpatch, and there's a hann window in the parent patch as well. The tables have three guard points, that's why I'm multiplying [phasor~] by the size of the table minus three and then add one. The output of this is a waveform with very low amplitude that kind of bounces up and down within a sine like mask. Don't know if I'm making my self clear. My main question is, how do you get a sinusoid out of a sine and a cosine? Also, what's wrong in my approach? ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] cosinesum vs sinesum in arrays
* Alexandros Drymonitis adr...@gmail.com [2014-01-29 21:02]: When sending the cosinesum message to an array, don't you have to specify the number of points too? And specifically a power of two? Plus, for a cosine you have to specify the first partial to have zero amplitude and the second an amplitude of 1 (or less). So the message should be something like [; array999 cosinesum 1024 0 1( First the name of the array, then the cosinesum function, table length and partial strengths. Hope I got your question right.. Thanks, you are right about the message format. My questions is with regard to a different behavior of sinesum vs cosinesum in Pd(0.45.0) when specifying the partials though. On Wed, Jan 29, 2014 at 6:01 PM, Peter P. p8...@aol.com wrote: [...] not sure if I am getting something wrong here, but it seems that sending the message ; cosinesum array999 1 should of course read: ; cosinesum array999 512 1 does something different (besides an obvious phase shift between sine and cosine functions) to the resulting contents of that table: the first partial on cosinesum seems to be the DC component, while the first partial in sinesum is the, well, first partial. Accordingly, the second partials have double frequency in sinesum as in cosinesum. Perhaps this is intentional, but was just wondering... best, Peter ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Data structures and their clickable area
On Mon, 2014-01-27 at 21:34 -0500, Jonathan Wilkes wrote: On 01/27/2014 05:35 PM, Roman Haefeli wrote: Hi I'm using a template consisting of a rectangle done with [filledpolygon] and a number [drawnumber] in it. While mouse clicks anywhere in the area of the rectangle are detected, it's only possible to change the number with the keyboard when I exactly click on the number. Is there a way to make the number catch the keyboard no matter where I click in the rectangle? One possibility is to make the hotspot bbox settable. Actually, something like this would be on my wishlist. Knowing it does not exist yet, I hoped for some kludge solution. Or maybe have a method to forward widgetbehaviors to another drawing command. Would certainly be interesting too, though having the hotspot area be configurable would make this less important. Anyway, thanks for your thoughts. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Data structures and their clickable area
On Mon, 2014-01-27 at 21:34 -0500, Jonathan Wilkes wrote: On 01/27/2014 05:35 PM, Roman Haefeli wrote: Similarly, I'd like to be able to mouse-drag anywhere in the rectangle in order to change the value of the number. You could probably do it if you use a field variable to define hotspots on every 6x6 tile of the rectangle. But you'd also have to constrain movement of the rectangle by abusing the quanta syntax, something like (-whatever:whatever)(0:0). Interesting idea. That would presumably constrain the field variable's screen coordinates so that it doesn't move when you click-drag it. I'm not sure, if I fully understand the quanta syntax. I'd assumed that something like x(-30:30)(0:0) would not allow any movement, as you suggest. But it is still movable as if I'd use plain x (without quanta). When I use something like x(-30:30)(-1:1), it jumps between -30 and 30. Then use the same field variable for your [drawnumber]. Unfortunately, when using quanta, the variable doesn't return the input (my mouse movement), but the result. So the number jumps between -30 and 30 as well. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Data structures and their clickable area
On Die, 2014-01-28 at 12:40 +0100, João Pais wrote: there might be a complicated and confuse way of doing it - by clicking the scalar, you activate a click message to the drawing window, that clicks in the drawed number. For that you would have to look around the click messages in the pd documents, which I didn't really understood so far. Can you intercept mouse events done in the canvas? Can you even do it without externals? Roman On 01/27/2014 05:35 PM, Roman Haefeli wrote: Hi I'm using a template consisting of a rectangle done with [filledpolygon] and a number [drawnumber] in it. While mouse clicks anywhere in the area of the rectangle are detected, it's only possible to change the number with the keyboard when I exactly click on the number. Is there a way to make the number catch the keyboard no matter where I click in the rectangle? That's not possible. Essentially what you want is to take a click from one draw command-- [filledpolygon]-- and map it or forward it to another-- [drawnumber]. Scalars don't give you any tools to hook in to a parent drawing command's widgetbehavior that way. Similarly, I'd like to be able to mouse-drag anywhere in the rectangle in order to change the value of the number. You could probably do it if you use a field variable to define hotspots on every 6x6 tile of the rectangle. But you'd also have to constrain movement of the rectangle by abusing the quanta syntax, something like (-whatever:whatever)(0:0). That would presumably constrain the field variable's screen coordinates so that it doesn't move when you click-drag it. Then use the same field variable for your [drawnumber]. I'm almost finished with some new drawing instructions for data structures in Pd-l2ork that implement a subset of the svg spec. I've got some mouseover/mouseout widgetbehaviors working, but still nothing particularly sophisticated in terms of mapping mouse/keyboard interaction to field variables. One possibility is to make the hotspot bbox settable. Or maybe have a method to forward widgetbehaviors to another drawing command. -Jonathan Any ideas? Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Data structures and their clickable area
On Mon, 2014-01-27 at 23:35 +0100, Roman Haefeli wrote: Hi I'm using a template consisting of a rectangle done with [filledpolygon] and a number [drawnumber] in it. While mouse clicks anywhere in the area of the rectangle are detected, it's only possible to change the number with the keyboard when I exactly click on the number. Is there a way to make the number catch the keyboard no matter where I click in the rectangle? To answer my own question: I had some thought about dealing with the 'capture the keyboard' part. When clicking the rectangle, I could use the pointer to route keyboards events from [keyname] to the clicked scalar. This way, I could even use 'Left' and 'Right' key events to move the scalar selection with the keyboard. This would allow to set a whole array of numbers by only using the keyboard. It's still not clear how to unselect the whole thing, when data entry is completed, though. I believe many interfaces allow de-select something by clicking anywhere nearby. Don't know if that is feasible in Pd. Using a key would be another option. The 'Escape' key, for instance. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] Data structures and their clickable area
On 01/29/2014 05:40 PM, Roman Haefeli wrote: On Mon, 2014-01-27 at 21:34 -0500, Jonathan Wilkes wrote: On 01/27/2014 05:35 PM, Roman Haefeli wrote: Hi I'm using a template consisting of a rectangle done with [filledpolygon] and a number [drawnumber] in it. While mouse clicks anywhere in the area of the rectangle are detected, it's only possible to change the number with the keyboard when I exactly click on the number. Is there a way to make the number catch the keyboard no matter where I click in the rectangle? One possibility is to make the hotspot bbox settable. Actually, something like this would be on my wishlist. Knowing it does not exist yet, I hoped for some kludge solution. It would literally be five minutes of dev time. But I'm not sure it's the ideal solution since often you want a hotspot to exceed the formal bounds of an object, and this wouldn't do that. Still, I'll code it up and see how it works. Or maybe have a method to forward widgetbehaviors to another drawing command. Would certainly be interesting too, though having the hotspot area be configurable would make this less important. Probably best to just fool around with Raphael.js or some such library to see what it does, and see what can be ported. -Jonathan Anyway, thanks for your thoughts. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list