3 simple exhibits with VL - any resources?

[remony]

Hi beloved vvvv crowd,
There will be a show with 3 exhibits that I want to set up in VL. And therefore I am happy to scrape all links together where I could learn something about how it is done. They are all pretty simple, maybe some of you did that already?

• Photo Station: A Screen, a Camera and a text telling the visitor to press the button (3D printed, next to the screen) Press in order to take a picture. Countdown from 3 to take a picture. Might use face recognition, or not. No “Redo” option needed.

• Voice-in interface: Screen and Microphone. Text on screen asks the visitor to repeat a sentence. It reacts to a threshold of the Voice-In that triggers the text to jump to “Thank you” or “Please repeat”. (Quick concept explanation: The voice is not processed in any way, it is a performative piece where I want the visitor to repeat a sentence aloud)

• Score-board: Screen where the photo from the beginning is shown. Question on screen: Is that you? 3D-printed buttons for yes/no. The question is repeated until the right picture shows up (that only happens if many ppl visit the exhibition at once). As soon as the right picture is shown, there will be some text about the person´s performance throughout the exhibition.
  This one is super easy as far as I can see, the only question here: How would you save the pics from the beginning so they are accessible from another computer. I just thought about a LAN cable and a shared folder.

Thank you so much for your help!
And for the dev´s (@joreg): I could make tutorial videos of those exhibits for later generations. ;) Is that of interest?

Have a sunny day,
Rosi

[velcrome]

Please make tutorial videos, or just document your progress. Rad idea!

I can share some knowledge about a photo station, hope it helps.

Many people use a SLR like Canon, mainly because it’s easy to find a fitting lense, and it’s got a nice way to attach (or remotely connect) lighting, and release the shutter off-camera. So this is the solution if you cannot get too deep into code and solve stuff with hardware. In your case, a good webcam is also fine.

I found that people are much more likely to use a photo station, if it continuously acts as a mirror, so prior to the actual picture hair can be combed back, smiles can be tested, face-to-frame-composition can be optimized, and general vanity can take its tax. This is something a webcam should accomodate well too.

So VL.OpenCV should be on top of your list to check out. I assume it will carry most of the necessary code, like streaming a preview, taking a frame once the button is pressed, saving it to disk and displaying it for review.
A stopwatch node might be useful for the countdown, but logic-wise this entire thing should be super minimal.

The button next to the screen sounds like a suboptimal way to do it, because photos will always have an outstretched arm in the frame.
I think a longer cable with the buttons on the end is a better solution, or it can be released by foot. It depends on how sublimal you want it to be really, from a big-ass emergency button that doubles as a prop in the picture to a slick and almost invisible dance-mat… everything is possible. But ergonomics should come first.
Custom buttons can be done with an Arduino-type board. Dance mats can be done with DirectInput (an API from Microsoft). What do you think is best in your case?

What’s the deal with the face recognition? Just some more techy bells&whistles, or is it essential for the station somehow?

What happens to the photos? If you intend to keep them or even repurpose them afterwards, make sure to inform people about it and get their consent.

[joreg]

@remony

And for the dev´s (@joreg): I could make tutorial videos of those exhibits for later generations. ;) Is that of interest?

absolutely yes please!

[remony]

Hi @velcrome, Thanks for taking the time!

Yes a webcam will do and also I intend to make it a digital mirror before the pic is taken!
And I want the 3-2-1 countdown visible on the screen (I guess I need to convert the OpenCV to Skia in order to do that?). And that should resolve the issue of the outstreched arm?

For the buttens I want to work with a basic Arduino.
Question Nr. 1 right now: Is Arduino already possible and convenient to use in vl?

Face recognition is not needed for all costs, no.

And yes, the ppl will be informed about their pic being taken and used as part of the exhibition. The whole work is about data scraping, so I make sure they know what´s going on ;)

[ravazquez]

Hi @remony,

I think you could use Face Detection (not recognition) to prompt information/audio when someone goes in front of the camera, for example if a face is detected after X seconds of noone being detected you can show a welcome message with instructions or something on those lines. It would add some complexity to the project so take that into account, but could be a good excuse to use face detection if you want to explore that.

To add GUI elements to your images you will need to take your CvImage and convert it to SKImage (Skia) using a combination of ToImage (OpenCV.Conversion) and FromImage (Graphics.Skia.Imaging).

Cheers!

[sebl]

i made a photobooth with vl some time ago. to be more precise… too long ago and the patches aren’t working anymore in todays gamma.

but i can recap some things:

• i used an industrial camera from ueye (ueye package is out of date and doesn’t work with current gamma. i’ll fix that when i have time)
• such a camera combines both worlds: high resolution and continous stream of images, so you can have a preview
• back then i used VL.WorkflowFoundation as a statemachine to control the logical flow
• i attached a printer to the comp, so after the photo was taken, the user could decide to print it. for that i used system.printing and made a quick implementation of the things i needed. the patch looked like this:
  
  
  
  (a nicely wrapped VL.Printing would be nice at some point)

that’s how my rootpatch looked like, to give you some basic idea:

[joreg]

@remony

Question Nr. 1 right now: Is Arduino already possible and convenient to use in vl?

more convenient than in vvvv beta.

[remony]

Getting carried away here :D
(only bcs some parts of the video are outdated ;)
@ravazquez @joreg

[velcrome]

Bold! But it’s not like vvvvTV is a trademark anyway.
I presume @joreg agrees, so seize the means of production!

Did some procrastination in my now half-empty-half-devastated flat feels ok,especially if for the sake of testing pure conceptual interactive design.

I made a simple test device from an Italian magic arm and a Chinese Android phone.
GUI for the countdown was clumsy enough, but flexible - I could set the timer myself, and it was displayed on screen. True bedroom prototype, no vvvv involved.

My findings: 3 sec is not enough, nor is 5sec (unless you like motionblur and want to push that sporty look 🍷
).

I suspect, the sweet spot is somewhere in between, but only with 10sec I actually had the time to frame and focus something meaningful:

Test sample group was very limited while experimenting (because limited to 1 and somewhat constrained by time). I expect that true multiplayer might need more testing.
When a Mirror was used to simulate cognitive load for multiplayer, finding a good setup took minutes at a time, but enforced the idea that someone has to push that button for any progress at all…

Godspeed to your surveillance installation, Rosi. And really, ask for consent and shill free will. There is nothing better