@zimt

Hi,

I don't have a specific solution for your issue, unfortunately. I have projects using mediapipe that I have come back to and found that there have been a number of changes to both Mediapipe and Pythoner compatibility that required rewinding and using an older, supported Mediapipe and Python version (such as Python 3.10) to avoid dependency conflicts.

My Mediapipe projects have been usable again after doing that. Linked here is a Mediapipe face-tracking patch in Isadora 4.5.1 but it was made before a json/pyhthoner skeleton decoder actor was available, so the face marker decoding is scripted manually using a Text Parser actor. 

Mediapipe Face by Bonemap

Best Wishes

Russell

@michel

Wow, that's pretty impressive!!

I wonder how much code if shares with the new TouchOSC... most of it I would think.
If you were able to share the TouchOSC file so I could poke around and see how the programming works, that'd be great, especially the dynamic filling aspect, which is what I'm trying to do too.

Thanks!

Mark (not Mark)

@michel

Hey Michel how are you? Thought you weren't working on the forum anymore.

Hope all is well and thanks for the answer.

Raay

@sommerdi Hi! Thank you for your feedback - really useful! I have the IEM plugins, I will experiment. 

With my Focusrite Soundcard set to 96KHz and Isadora set to 96KHz with a 128-sample audio buffer I get the least delay - it's very slight, but annoying.

However, when I try this in Reaper with the same audio settings and audio interface the sound is crystal:

Hopefully I'll report back with a solution - possibly the IEM plugin suite.

P.S. The Sennheiser AMBEO plugin seen above is discontinued (Isadora doesn't recognise it) and has been superseded I believe with the dearVR plugin that I'm using in Isadora.

Best Simon

@bonemap

I have discovered that the issue is with Mac OS and that this is a common problem with touchscreen overlay devices using a Mac.

I have found this utility on Github that identifies and attempts to fix the issue:

https://github.com/shueber/Tou...

Best Wishes

Russell

@mark_m said:

Can I achieve this by hacking the preferences somehow?

 Unfortunately not. I believe Isadora binds to the first available device, which on PCs can change depending on config/build.

Do you want to start a support ticket, and I can dig into see if I can workout a workaround?

I think this add-on might be what you need: TROIKATRONIX : ISADORA - DX – Rotate Canvas

It uses Matte to create a new canvas at the required 'rotated' dimensions, and then drops the image into this new canvas. 

@woland

Sure, here we go:

Liveaudio 2 OSC:
https://alexxaudio.gumroad.com...

Audio-Automation 2 OSC:
https://alexxaudio.gumroad.com...

@jaio

If you can find a suitable cube LUT, for converting rec 2020 to rec 709, you could use my 'color grading' shader (TROIKATRONIX : ISADORA - DX – PY – Cube2LUT – ColorGrading).

This should allow you to convert the color, but will not be as true to source as making the conversion in something like Davinci Resolve (my recommended option).

a workaround is to use a font with a capital version like Bodoni 72 and Bodoni 72 small caps and to switch between both.

Best regards

Jean-François

Honestly this confused me too when I was buying my first projector. Lumens measure total light output from the projector itself, while lux measures how bright that light appears on the screen - lux depends on both lumens and screen size. So same projector will have higher lux on smaller screen and lower lux on bigger screen. For comparing projectors, always use lumens since that's the consistent measurement regardless of screen size.

I use an Optoma projector and when I was evaluating, I focused on ANSI lumens specifically - that's the standardized measurement. Some cheap brands advertise "LED lumens" or "peak lumens" which are inflated numbers, not real brightness. For dark rooms 2000 ANSI lumens is fine, ambient light rooms need 3000+. Don't get fooled by high lux numbers on small demo screens in stores - always ask for ANSI lumens rating before deciding.

@woland Many thanks L! Hopefully it’s possible.

Best wishes Simon

@emullis

Actually, if you humor me one more time, there's one thing we didn't try. Based on the fact that you stated this same setup used to work for you with an older version of Isadora, we should try that.

Can you:

In Isadora, go to Help > Open Plugins Folder > TroikaTronix Actors PluginsSelect everything in that folder, then right-click one of them and select "Compress"You'll get a .zip file that you can later unzip and use to restore the contents of this folder later if need beDelete everything except the zip from that folderQuit IsadoraInstall an older version of Isadora using our Legacy Downloads Page (I'd suggest starting with 4.1.1 and working backwards)If you want to keep the recent version of Isadora on your computer too, see this article on how to install multiple versions of Isadora on your computerTry streaming data from Rokoko again with each older version of Isadora and see if any of them are stable with your v2 suit.

One other idea, that unfortunately requires different hardware, is seeing if a Windows computer running Rokoko Studio and Isadora is stable for you. I know this worked for an artist I collaborate with in the UK who uses a v2 suit.

@dusx

Thank you. Managed to do this with the help of Claude Sonnet 4.6  which showed me how to retrain the data using something which has the same landmarks as in your Mediapipe pose tracking examples.
This seems to work very well, thank you. 

Not sure how to share the process? 

I took maybe 25 photos of each of 14 poses (yes, different handstand shapes, of course!)

Then organised them into this folder structure

training_data/
    arms_crossed/
        photo1.jpg
        photo2.jpg
        ...
    arms_raised/
        photo1.jpg
        ...
    lunge_left/
        ...

And then followed Claude's instructions. 

Here's my Colab notebook:
https://colab.research.google....

After I'd successfully trained that and got the files out the other end...

It then gave me the python mediapipe code. 
https://www.dropbox.com/scl/fi...

And that outputs the OSC into Isadora. Two streams, one telling me which pose it is, and the other giving me the confidence about it.

Hope this is useful to someone!!!

Cheers

Mark (not that Mark, who I'm sure doesn't use Claude to do his coding!)

@woland Thanks a lot. That makes sense... ;o)

@jjhp3 said:

Any ideas about getting the VSTs?

Depending on the names, sometimes you can look them up to find them (if they're from a specific company).

@dbini

That helps. Thanks for the quick response.

many greetings

Matthias

@piotrtwon

Hi!

I am going to direct you to our Support Ticket system, so that one of our Support staff members can help you further with this request.
You can open an Support ticket here: https://support.troikatronix.com/support/tickets/new

Short answer; we have options available within the product to do object-tracking in Isadora.

With kind regards,

Juriaan

i think i remember - a long time ago - there was a feature request to get the Eyes monitor feed as a GPU output, but that has not manifested.
it is possible to create a visual representation of the blob tracking system using Eyes++ and a lot of Limit Scale Value actors to process the outputs from Blob Decoders and control Shapes and Lines actors, maybe with a Threshold on the video feed if you want to simulate the blobs. The useful thing is that Shapes, Lines and Text Draw all have GPU inputs so you can daisychain them together to add graphics onto your video feed.

but maybe Isadora isn't the ideal solution to achieve this effect? object tracking in browsers is getting really good these days, i'm sure that a quick search around github will find an app that can do what you need instantly, then you can pull the results into Isadora with Screen Capture