I`m new vvvv user. After playing with hierro skeleton plugin I have one question
Is kinect plugin capable of showing rgb and depth video streams simultaneusly?
Or is it possible to modify script?
I’m trying to overlay depth video over rgb one, now using separate webcam, which is stupid…
the RGB texture is missing in the current hierro’s version, you can modify the plugin yourself or you can wait a bit untill hierro will release a new version with RGB… i think it will come shortly :)
i had the luck to playwith a alpha version and i can say it runs quite well ;)
Thanks Sapo, good news. I have no programming skills. Is it possible to take a look at alpha version as well?
yeah RGB is cool
i tried Brekel stuff, try it it’s free
Could be interesting to be able to set min/max distance treshold in Hierro’s plug eheh
Yes, I tried Brekel too, nice stuff, but I need vvvv plugin to prepare my multimedia piece. On OSX Max/Msp there is J.M.Pelletier kinect plugin, unfortunately I’m working on PC…
i’d share it if it’s ok for hierro, but i think that early version is still incomplete, and you must put files in a strange location… so you might need to wait anyway for a fully working one. Btw you can use vux’s kinect plugin if you need just rgb and depth images!
and…cant you use OSX to prepare your multimedia piece ? you could try IPelletier plugin ? kiddin :D
Anyway, dont have time to put hands on plugin, but i will try to share soon the alpha :)
I will appreciate that , if it is OK. BTW, where can I find Vux kinect plugin?
I’ve downloaded kinect_pong, but kinect part is red- no kinetc and kinect exture. I’m also not sure is kinect driver included, or I have to start niserver.exe or nitracker.exe?
OK, found it, it’s for CL-nui driver, isn’t it?
yep… no text …
Hi there, i´m actually looking for depth info in vvvv, too. I tried the vux- plugin, but didn´t get it to work on single time. The OpenNI- Skeleton did work kinda well after some licensing mess was detected and cleaned up :-)
Would it be complicated to break down Skeleton to just the Depth- and RGB- Signal? I want to get rid of the necessarity to calibrate, as i only want to blobtrack hands and bodies.
I have some skliss in modifating other coders work to my necessities, so i´d try my best to get this to work, but I´m afraid that this would be a little too complicated.
If anyone else would be interested in doing this, I´d be the first to alpha/beta/andsoon- test :-)
Hi cznickesz, try using vux with CL nui driver, which works on win7.I tried on winXP, cl nui installs OK, but do not transmit camera RGB and depth streams There are some vvv patches- heightmap, pong. Good start for playing with rgb and depth data.
I did get it to work on my NB (which is the relevant one for my plans) and i guess it all was a problem of wrong drivers (did use the original primesense- drivers instead of the hacked ones). CLNUI didn´t ever work. As i use hierros stuff now, this is no problem, though.
Im tracking multiple persons via colortracking now. What i was wondering is if this wouldn´t be possible via plugin- code and be way faster than the freeframetracker as it is obvious, that the kinect already does the tracking work. i do not mean skeleton tracking which already is done, but simple midpoint- based positiontracking without the necessity to calibrate.
About multiperson detection,if you right click kinect node, you can read the code, search for GetCoM, you should find something referring to detectedUsers Center Of Mass, in that line you are already handling a main coordinate, the user id, and related info , for user detected.
Hi, I’ve found GetCoM, and two variables com y and com y. To use them, I need to transfer data to output pin. Now I have to learn scripting language to do this. Is there any good tutorial how to write cs?
Thanks Hierro, I must be blind…, btw any progress in adding RGB camera to your kinect plugin?
This sounds very fine for me… i guess i´m going to check that tutorial and start coding plugins very soon. I messed around with shader- code and had loads of not very useful fun those days (i got a shaderbased kind of colourtracking up and running then, but never finalized the project) and i bet i will have even more now, hehe.
hierro, where would i be able to find some docs about what else i could read out. I was wondering if there´s something like width and height of the detected user so get a scale/ z- position. And there would possibly be some more data that could be interesting.
Well, using some tricks I have both rgb and depth simultaneusly in vvv!
Well, what kind of trick?
Would you share the code?
PS: Nice installation