Kinect + Isadora Tutorials Now Available

camilozk

Hi and thank you very very much for these tutorials. Well done, great job!

I have a little question about the last part

@ http://troikatronix.com/support/kb/kinect-tutorial-part3/

on "SETTING UP THE SYPHON FEED"

"Windows: Add a SpoutReceiver2 actor to the Scene Editor, and click on its ‘server’ input"

there is no server input in my SpoutReceiver2

thanks!

DusX

It's the 'Select' input. (trigger type)

It should open a little SpoutPanel dialog (if you put spoutpanel.exe beside isadora.exe in the Isadora folder). In that panel you can select the input source.

If you want it 'hardcoded' to the SpoutReceiver, you will need to type the name show in the SpoutPanel into the first input of the SpoutReceiver (Sender Name).

It is case sensitive. Also beware of whitespace (trailing spaces etc..)

Once sent and working, save your Isadora file, and the Sender Name will remain.

DusX

This tutorial may also be helpful:

http://troikatronix.com/support/kb/using-spout-ffgl-video-sharing/

camilozk

ok thanks dusx

I was actually expecting a "drop down dialog, displaying all the active Syphon servers on your computer", as the tutorial states, but I only see the "depth 640*480"

I realized however to modify the sketch in processing in order to send either rgb camera image // infra red camera image // depth without colored bodies of tracked bodies // depth image with colored bodies of tracked bodies

DusX

@camilozk

It seems you might be confusing Spout with Syphon.

If you are on a PC you will be dealing exclusively with Spout.

If on Mac it will be Syphon.

if you wish to share you code modifications regarding the video sent to Isadora, I am sure other users would appreciate the information.

(if you post code, wrapped in html PRE tags, the forum will auto format the code. It must be done in the html edit mode.)

Skulpture

Am I right in thinking this will only track one person at a time?

mark

Dear @Skulpture,

Well, yes and no. I had at least six people in the frame at one point and they all had Skeletons. But I didn't revise the code to _handle_ those multiple skeletons. We should make this change. I can't do it at the moment with my coding responsibilities for the main program, but maybe someone here has the desire to make it happen?

Best,

Mark

leben

Hello,

I am running Izzy version 2.1 on a mbp retina on yosemite. I have followed all the steps for this tutorial, I have run the processing sketch and have an image coming from the kinect in the viewing box, but there is no skeleton image..am I missing something?

thanks

Pascal

Hello,

It works great...thanks a lot for all.

Best,

Pascal

DusX

@ leben

It can take sometime. You need to stand in front of the kinect, allow it to detect your figure (will color your form a solid colour)

After a little while (if your form is clear) it should switch to adding the skeleton. (this process is very cpu intense, and on my PC almost freezes things for a couple seconds) but once in this mode returns to performing well.

gapworks

yes it works great! i was just wondering if there would be a possibility through processing to get kind of a ghost image and/ switch between different possibilities?

DusX

@gapworks

Like @mark mentioned in regards to tracking multiple skeletons, switching the video feed would have to be added to the Processing code.

I imagine that OSC control of the video type could be added.

Perhaps a Processing Pro in the forum might take this one and share the final script.

leben

@ DusX

All is working fine. Thanks for the help, and to those who wrote the tutorial. this all very exiting

mc_monte

Thanks all for the wonderful comments and helpful feedback. Get out there and make some cool art :D

Skulpture

I am guessing from the screenshots and from my own patch that all the red text (errors?) that show up in Processing patch are normal?

Marci

"I am guessing from the screenshots and from my own patch that all the red text (errors?) that show up in Processing patch are normal?"

Copy & paste a sample - all library initialisation output is generally red by default... so __shouldn’t__ be anything to worry about, but I can take a look.

"Like @mark mentioned in regards to tracking multiple skeletons, switching the video feed would have to be added to the Processing code.

_I imagine that OSC control of the video type could be added._

_Perhaps a Processing Pro [user] in the forum might take this one and share the final script."_

Can do... but probs not til the weekend.

DusX

@Marci

It would be a greatly appreciated contribution.

Pascal

Oh yessssssss

Marci

Have hoisted the code over to my GitHub repo, so just keep an eye on that... feel free to report issues via that too as we go along, or fork / contribute etc. Not done anything to it at the moment, just preparing the ground so to speak.

@Mark / @mc_monte / @dusx / @skulpture / @michel et alle - if you want some specific text putting in the README.md acknowledging orig sources etc, just fire it over to me via message or similar and I’ll swap it in. At the minute I’ve left license type as “none”... suggest ‘CC0' seeing as it’s all open source anyways. Thoughts?

See https://github.com/PatchworkBoy/isadora-kinect-tutorial-files

Skulpture

I've also done a video walkthrough (unofficialy)

https://www.youtube.com/watch?v=0HY5U6QSyhM