Transdiciplinary Collaborative Project

At one point, i turned to the VVVV forums to get some advice about patching whilst i was trying to work it out myself. This was great since VVVV has so many nodes, it can be difficult to know which one to use for the job at hand.

Myself and Toni went off to find something to project onto in the meat time to have something for the presenation. For this, we went to the modelling course in another art building. We spoke to some tutors and were introduced to a student that had a car suitable for us to borrow. It may not have been a porsche, but the colour was perfect to project onto to show a working concept.

The evening before the presentation, i had spent some time working on setting up the projection. I measured the distance from the projector to the car and the height at which the projector was at. It was from this angle that i retook the image of the car and redid the image to be projected. This was to try and get the best result possible. On the day, i found out how to warp and adjust an image using the homography node. This meant i was able to adjust the perspective and position the image on the car more accurately. There was still some slight overhang but i think i could fix it with more time. This was the best angle to view it at and using the GUI Toni had refined and built in flash, were able to show a good concept that got people interested.

The deadline was fast approaching and i was still having no luck working out how to create a clickable icon in VVVV. We decided to just leave it and work on an animated GUI whilst operating the projection from the VVVV patch. A few days later, i began to think about the VVVV patch again. I thought about just having the icons placed there and creating a quad that was a button. It was then i realised that the quads were the buttons and the file texture would just be attached to the quad. I began to work on the patch again utilising a Button node which was an exsiting patch. I got this to work and when the desktop mouse was over the quad in the renderer, it would change colour. All i had to do then was work on a mouse click that would trigger something, creating a button.

During my attempt to keep the project going, i looked into different ways to program a GUI to work in conjunction with the projector. This is where i found Shoes, Python, Glade, GTK+, MSYS2, all of which are programs to code something. I do not know the first thing about any of these programs so i decided to go back and try VVVV again since i was familiar enough with how it worked.

Myself and Toni decided to try and project the car ready for the main presenation. We had some issues with VVVV working and getting the projector to work. Once we switched computers in the base room, we were able to get the projector to work but not VVVV. From here, we just decided to try and line up the images to the car.   We moved the angle of the projector and the car to try and get the best results but really struggled without the use of projection mapping. This was only because we were unsure of how to do it in VVVV. This shows that part of the car was not being projected on despite lining the front of the image up. I decided to take all the equiptment home and redo the images, measuring my distance and angle from the car. Toni began to redo the animate GUI i did and make a cleaner, polished fully functional version.

I considered our options and whether we would need to use VVVV or not. If we werent able to complete the main patch, it may be worth just using some simpler software to map out the projection. I suggested using a combination of software i had previous experience with. I was happy to refamiliarize myself with it and projection map that way. Upon some discussion in another session, we were happy to get the VVVV projection patch working and to carry on using it.  I then began to try a trial version of HeavyM as buying it was far too expensive. Unfortunately, i was unable to project a file texture but it may be worth looking into more to see if i could map the car and just use a colour selector to project the desired colour. https://www.kickstarter.com/projects/theheavymteam/heavym-bring-projection-to-another-level   As stated before, i have used resolume professionally as a VJ at live events. It may take me a little while to readjust myself to it but i feel confident...

During easter break, i spent my time trying to find  a solution to the button problem. once i met back up with the group, we all decided that maybe need to look into alternative options in order to at least show concept. It was disheartening to realise that the project we wanted to complete may have been too much for us. But we at least wanted to have a good working concept. In order to show the GUI, i suggested doing it in animate. it would be a fully working interface but it wouldnt control any projection. To project, we could still use VVVV but i recommened maybe looking into something simplar if we didnt have to use VVVV. I had used some software before but said i would look into it and see how i got on.

The rest of the group designed a GUI for the project. The idea is that you would select what you want to change and then adjust it. I did have a few adjustments to make with this. Firstly, i suggested getting rid of the car from it as it may be too distracting to the main idea of the project. Secondly, a few options would have to be changed in order to fit in with the brand and what they offer. The colours would have to match what Porsche offer with the car but suggested there could be a free design feature where this is used to freely create the car. Another thing is that they do not offer window tint with this vehicle and it would also be difficult to project so this would have to be replaced. Afterwards, i moved a few things around and came up with a new one. Although much more simple, it would allow space for options. The icons would be animated and would be triggered on a click whilst making it a bit larger to show the option selected. The options would be displayed u...

The issues began to arise when i started to try and create a button out of the icons we had premade. I hadnt done anything like this before so i turned to forums and google for any advice. I couldnt find any help with the specific ideas we had. At this point, i just began to play around with VVVV to see what i could manage.

In the beginning, we did not have any issues with VVVV. What we have in mind seemed fairly straight forward. It wasnt until i began to work on the main patch that i came across issues. The first thing i did was work on an intro screen that would be displayed before the user began to change the options. This took a bit of work to get right, but i eventually got it to work. The idea is that a user would walk up to the tablet and see this and be prompted to touch the screen to begin. The images would then transition out of the frame and the main menu would be displayed. How i am going to get the rest of this to work is yet unknown.

Since the last post, i have been working with VVVV trying to figure out how to make buttons using icons we have premade. I have tried just using a button but am unable to figure it out despite looking online. The group and i have met since and we are currently working on a concept test with a small scale vehicle, just to get the projection aspect working. We did think about usinga program called ReconstructMe, to generate a 3D model of the vehicle, but since we are only working with a portion of the car, it would be much easier with the time we had, to work with a 2D image. Myself and Toni have began to work on an Animate project that would simulate what our GUI would be like in VVVV. This would also be useful during a demo, just so the user could get a good idea of what the final project would look like.

We have met with the Porsche dealership and they are happy for us to project onto a vehicle. They have also expressed an interest in showing the project in the showroom which means we have to pull it off and do it well. We had to examine Porsche as a brand and realised the kinect idea may not work best for the brand. We have decided to work with it as a tablet that controls the projection, still showing what the chosen car will look like. I have already began to patch it in  VVVV.

To projection map onto the car, my thought is that we will initally have to model it or find one already modeled for us. This is to have the right measurements when designing what will be projected onto the map. Im not totally sure what software we are using at this point as we wish to have interactivity but i will look up some CAD designs for now. We are also unsure on the vehicle model.

In my search to find the right software for the task, i came across some software by Disguise (formally known as D3). This would be great for what we are after but how could it work with interactivity?

Whilst visiting Blechley Park, i came across an interesting part of it that featured a room of interactive games. I decided to play around with it to see how it worked.  

These were a few examples i found of projection mapping and vehicles. http://eventprojection.co.uk/case-study/ford-smax-case/ http://eventprojection.co.uk/work/dirt-four-promo-video/

I thought i would email a few visual companies to see if they had any technical specifications from old projection mapping events for us to look at. This could give us a base to work with and also show us a few extra things we could throw in. As of yet, i am waiting to hear back from one place in regards to my questions.

The idea of the project, is to create something that would be in car dealerships. Its to change the user experience of buying a new car. Rather than going into a dealership and buying a car from a catalogue or picking what you want from a website, the user will be able to have a physical representation of the car they will be getting. As well as this, we also thought about ways to make the experience a bit more personal. This would involve the ability to design your own graphics to go on part of a car. I mocked up a quick concept to represent what we would like to create.