@citizenjoe  Good idea, thanks! 

@armando

Hey there,

I am going into a deep dive over in TouchDesigner land to try and get good results from dotSimulate's implementation of realtime Image Diffusion with his "Stream DiffusionTD" project. 

So far I've been able to get it to work on my M3 MacBook Pro @ ~4-7fps and on a Windows computer ~18-20 fps with an Nvidia graphics card.

I'm still very new to all of this but I think that I can confidently answer at least your first question "Question Does this means that I might end up with multiple virtual environments that might contain mumpy for instance ?":

1) - Yes, you might end up with multiple VENV(ironments) for different instances of something you are working on. I have ended up with between over 5 different attempted installations of StreamDiffusion on my computer as I've tried to get it to work. I just added a new installation recently because there was an update from DotSimulate and I didn't want to accidentally break my previous installation so I just made a new one and now have both in a working state.

In terms of all the common libraries. . . . I know that NumPy was one of the ones I had to struggle with getting the correct version installed of to make this work . . . 

Here is a video describing the install process for StreamDiffusionTD: 

@skuvenI've used Data OSC. . . the facial tracking was causing it to crash a lot on my own and my students phones last semester - but just tried messing with it again before classes started last week and it seems it may be a bit more stable with whatever updates there have been in the past year.

. . . Very curious to think about how to use it's GPS data as I have never really had a wifi network large enough that I think I could get that much useful data from the resolution it offers. . . . but a neat little app!

@swda

The best way would be to publish it on the derivative.ca website, there is a job offer page…

@eva said:

Im trying to add the project

All I need is the .izz file. Maybe you're trying to upload a .zip file or something else? Failing that, you can always send in your file via a ticket using the link in my signature.

@eva said:

I used this and extended it

https://troikatronix.com/add-o...

Unfortunately what you started with doesn't help me in this situation, I need to see how you implemented it. I suspect that you may not be using a background Scene for lighting, but I can't confirm that without looking at what you've programmed.

@woland couple years back i tried using a microbit connected to the mimu glover software (trial) and it worked. Pretty cool. I didn't persue because wasn't a priority at the time. But if you have a microbit around....

@dillthekraut

Hiya,

I linked to this earlier in this thread, but it's now become a long thread so easy for things to get lost :-)

https://www.adobe.com/devnet/x...

and this in particular shows how it's embedded in files.

https://github.com/adobe/XMP-T...

It really does seem to be cross-app, not just Adobe specific.

I know, for example, that if I export a standard .mp4 (H264) file from Premiere, that has markers in the timeline, those markers appear when I re-import it in Premiere. So that's an .mp4 file with no XML sidecar file, meaning that all the info is in the header.

@rwillats said:

Now, when I try to open Syphon (to then re-install the OBS Virtual Cam), the icon appears on the dock, but the actual window doesn't appear. If I go to help>install OBS Virtual Camera Plugin it waits and then closes the Syphon app. I can't get any further. Any ideas? 

 FWIW I have the same problem and I'm on Ventura 13.6.7.

I installed Syphon Virtual Webcam after already having the latest OBS installed and when I launch it the icon shows up in the doc, but no main window. Interacting with the Virtual Webcam in any way (like the help -> install OBS Virtual Camera Plugin) causes the program to close..

Just a bit more info for debugging in case it's helpful

Just like the fill tool in Photoshop. I imagine an actor with these inputs:
Video
Horz (the horizontal position of the target pixel)
Vert (the vertical position of the target pixel)
Threshold (0-100)
Contiguous (off/on)
Colour
Trigger
So you can specify a particular pixel on the incoming video screen. Hit trigger to colour every adjacent pixel of that value with the specified colour. (if contiguous is set to off, it will colour every pixel on screen that has that value, but this becomes more like a Luma or Chroma Key)
The Threshold would enable you to select pixels of similar value.
Maybe it would need a selector for different types of value (Hue, Luminance)?
What if it could fill a moving image? Would it need constant triggering? or maybe it doesn't need a Trigger at all, just a Bypass?

Any ideas? Would this be useful? Would it even work?

@fred

Hello Fred,

As I mentioned earlier, I can’t provide detailed answers to your specific questions because of the long overdue staggered vacations that our staff are going on. However, I want to do what I can for you right now.

After reviewing your licenses, I realized you currently don’t have access to a license compatible with Isadora 4, which was an oversight on my part. I’ve now renewed your Beta Testing License and I am about to create a ticket in our system to send you the info via email. As a longtime power user of Isadora and a talented member of our community, it would be a real shame to see you go, so there will also be something in the ticket that will hopefully help make the financial decision of whether to upgrade easier for you.

Also, to briefly address one of the important points you brought up, as I said earlier, one of our primary goals right now is to ensure the future growth of the company. More specifically, we're aiming to bring more developers on board, as doing so will allow us to enhance feature delivery and assist with the ongoing development of Isadora.

Best wishes,

Woland

@dillthekraut said:

How is a JavaScript template meant to be uploaded to the plugin db? I guess as user actor? There is no specific category like the 'Open GL Shader', which has probably the same work path, or is there?

You could go with Category User Actor (or Patch if you want to make an example file) and select the tag "Javascript"

@dillthekraut said:

Maybe this is might be worth an extra category, as well for the new Pythoners?

The Javascript tag has it covered. There's a tag for Python as well, so that's also handled :)

@dillthekraut said:

I would also like to kindly remind on the feature request regarding the search function for the Plug-In DB.

 Oh, believe me, I'm waiting on that one too. I'm dying to have it.

@woland Thank you! With the proper changes, using a background scene for the ArtNet actors worked like a charm!

@jtsteph said:

Davinci Resolve'

 Maybe this old thread of mine could help to start on something. Back then I worked with Premiere EDL exports to interprete them as timecode triggers. I'm not familiar with Davinci, but maybe there is some way to transform the script to work with similar files from it (I found some youtube videos which claim to describe the export of EDLs from Davinci, but I didn't check them out).

I didn't work with this since the project I used it for (I needed to tranfer thousands of surtitle cues). So no guaranty that it still works. But maybe i'ts a starting point.

thanks for the shaders. Enjoy playing with these YouTube

Hi hi, 

This is Chun-Shing. I will perform my piece at Zuricher Theater Spektakle from Aug 22nd to 24th 2024. 

https://www.theaterspektakel.ch/en/program24/production/chun-shing-au-gpo-box-no211

As part of the show, I created 14 paper robots that, using Isadora 3, send an OSC signal to ESP32 to control the servo motor ( wireless).

Please come to see it. And, please share my performance with people who are in Zurich or friends who are planning to go to the festival ~ 🫡❤️❤️I would love to hear some feedback.

@woland The future can't be fast enough...

@dbini

Great, your steps caused it for me and I've logged it. Thanks for the assist :)

@woland

In the GUI I can change IPs ,  SSIDs + password, and port numbers, but not OSC Adresses. That's surely a unwanted lapsus on their site.

As they say, the system is meant to be used by assigning a different port to each sensor.

Cheers

@juriaan Thanks a lot. I know that new systems, especially when still in beta would not run on a computer with a mission critical izzy patch especially on a new Isadora version.... too many moving parts. But thanks again. Please keep us posted if you find some problems between sequoia and izzy 4.0

Thank you very much! I can work with that!