Day two of testing on the full scene. Today’s plan was to create a few different styles which I could potentially use as presets, then get some feedback on them so they can be improved and development. Testing was done using one minute loops played back at around half speed to reduce the framerate for both performance and aesthetic reasons (so the glitches aren’t as fast).

Style 1 - Glitching all three together.

  • On my first day of testing I applied glitch effects to the individual models. This time I wanted to try glitching all three together. The benefit of this is that fragments of one model can appear on the others. However in reality it doesn’t look very good when they all glitch similarly at the same time. It looks choreographed and mundane, not as glitchy as i’d like.

Style 2 - Colourful compression.

  • This test goes back to glitching the models individually but adds in some compression with the saturated colours. Each one has a different level of compression and a slightly different frequency of glitch (20%, 25% and 30%). The compression makes the edges look pixellated, however after watching it for a while it seems a bit pointless to do this with the projection. The models themselves already look compressed and pixelated, doing the same to the projection just makes it look like its been mapped poorly. I think compression might be something to exclude from projections as the form of the models already does enough.

Style 3 - Just displacement.

  • This style of glitch doesn’t introduce any colour to the sculptures. Each one had displacement glitches only with slight different settings to add some more variety. I quite like this effect. Its subtle but still really interesting. With each sculpture occasionally glitching and at different times, it sort of makes you think ‘did that really just happen?’ then you try and catch the next glitch as sometimes the one you’re looking at doesn’t do anything. It would be interesting to add some colour into this occasionally too, as as variant to the style with constant colour.

Style 3.5 - Just displacement + Glitch FBO.

  • Using the same source video as style 3, this one has random use of the Glitch FBO effect too to add further distortion. I like how it adds in occasional displacement which is huge so that a whole model is unlit, and how it adds in some crazy distorted colours and patterns. It makes it actually look far glitchier and more interesting, though it might be worth using the FBO effect more sparingly.

Style 4 - Displacement with a splash of colour.

  • The video for this started similarly to style 3 however throughout the minute long loop, it is occasionally set to use colour in the glitches alongside the displacement. This adds some variety as sometimes it’ll do nothing, sometimes it’ll glitch and just displace and others it’ll displace and change colour. I think this works well as a style as there is more variety.

Style 4.5 - Displacement with a splash of colour + Glitch FBO.

  • Similar to how style 3.5 works, this one uses style 4 with occasional Glitch FBO. I slowed down the randomiser for this one so there are sometimes longer periods without the excessive glitching which I think is an improvement. This is probably my favourite style so far as it’s the most visually interesting.

Style 5 - Displacement + Convergence effect.

  • I decided to have a look through at some of the other effects in VDMX and found one called Convergence which seems to chaotically move the colour channels and blend them. I used the style 4 video as a base and had the effect randomly turn on and off. This style is surprisingly captivating. I’m not sure if its becuase its something new and I’m so used to the Glitch FBO effect, but I’m a big fan of this one and it definitely has a potential use in the final project.

Style 5.5 - Displacement + Convergence effect + Glitch FBO.

  • Again, similar to style 5 with the addition of the Glitch FBO effect for further distorted, glitchy wonderfulness. The two effects run on different randomisers so there’s such a variety to look at ranging from no glitch at all to full displacement and both effects at once. I think this could cut it as my new favourite.

Style 5.5.5 - Displacement + Convergence effect + Glitch FBO + Face capture.

  • This is an iteration of style 5.5 which adds face tracking into the mix. Well for the video its not live face tracking as its only me in the room and I can’t film and be in front of the camera. BUT it would work live, as long as the lighting in the exhibition room allowed for it. The faces are behind the glitched imaged in the mapping which use the add blend mode. So when the glitched video is white, the face cannot be seen, but when it is anything but white, the face can come through. As it is the glitches which make the video not white, the face appears when its glitched which is a very interesting and entertaining effect. The biggest issue with this is that its very resource intensive and makes my laptop run about 15 degrees hotter than it does without it. However in the show I think I might be able to work around this when I have an iMac to use. The plan for the show is to run VDMX on my laptop, then use syphon to get it to the iMac which is running madmapper and the 2 projectors. Depending on how powerful the iMac is, I could also run a face tracking Processing sketch and use syphon again to get that into madmapper. If I did it in Processing I could even have it display multiple faces so if there’s more than one person they all get involved.

Style 6 - Complete Chaos

  • This final style is similar to 5.5.5 but uses a different source video. The origial glitched source video has quite large blocks of colour in it which are a lot bolder than previous iterations. While this style looks quite nice, it doesn’t live up to the glory of styles with extra effects. So I added the random Convergence effect and Glitch FBO effect as well as the face tracking which makes it a lot more lively and interesting to watch. I also added the FBO effect to the face itself so that it glitches independently too which works really well.

Feedback

Here are a few comments from friends and family after showing them the above video. The nested bullets note my commentary on the feedback:

  • The ones with the faces look the best
  • The face doesn’t have to be in the same place, it could move from block to block or something.
    • This is a great idea, though at the moment I’m not yet sure how to implement it.s
  • It would go great with music (but I understand that isn’t what the project is about)
    • Maybe not music, but I have done audio reactivity so I could possibly control the glitches with ambient noise levels.
  • There should be more blacked out pieces which look like they’re missing
    • This could be done by making the displacement values higher.
  • Style 6 works really well because of the colours and the faces
  • Style 3 is also interesting due to the simplicity, makes it seem more ‘real’ as pieces of the sculptures are removed
  • Perhaps try mixing style 3 with style 6 to get a good balance of both - Add long pauses of cut outs like in style 3
    • I will definitely try something like this, I can slow down the glitches even further to enhance the effect
  • The first few styles are better as when it gets too flashy it can be hard to focus on whats going on
    • This is an interesting comment as most other people prefer the chaotic ones. I think it would be good to include a variety of styles in my presets ranging from mildly glitchy up to crazy glitchy chaos.

Tutor’s comments

  • Glitch happens when the illusion of reality breaks down by an electronic disturbance that wakes the viewer up and reminds them that they are looking not at reality but a “construct”.
  • What you have at the moment is not really doing that. It is more decorative than disturbing and that’s a problem.. I think that you might need to shoot some video which is closer to the structures themselves and then create animations in which they unravel, split open or rupture or are disturbed in some way that momentarily undermines or opens up their structural integrity.
    • I agree with this point. I have gotten a bit carried away with creating ‘glitch spectacle’ focusing on finding something which looks cool rather than considering the deeper meaning. As for creating certain animations, I will definitely give it a go (its on my to-do list in fact!). Its worth noting however that I am somewhat limited in the complexity of animations I can create as when I relocate to the exhibition space next Monday and set everything up, I’ll have to recreate all of the content again which there may not be enough time for more complicated effects. This is why I’ve been exploring a few of the real-time effects I can use so they can be changed and played with on the fly.
  • The actual figures are configured in a rather flat way. They need to operate against each other somehow operating more in space.
    • My current placement for the sculptures and projectors is to take up as little space as possible so I can be considerate in the actual exhibition. I agree with you that there should be more depth and they could do with being spaced apart a little more. I’m currently using a layout which has a smallish footprint and I can project onto the most surfaces. Once we’re into the exhibition space I’ll be able to see how much I can spread and better determine where exactly each model goes for optimal effect.
  • Your background in the exhibition space will be important. It should obviously be black or white but maybe work with panels operating at different angles.
    • This is something that depends on my placement in the exhibition space, so i’ll cross that bridge when I can. Admittedly in my current testing space it’s not ideal having a sofa behind the models!
  • Currently when you do include a body part it is just a face you should maybe work with more holistically with the human form.
    • It is currently just a face as I had the idea of doing live face tracking for it so that members of the audience will become part of the work. I’ve written a Processing sketch that will detect up to three faces and can display them, one on each model (faces are used multiple times if there are less of them).
    • I have an idea which I will attempt today. I’m going to get someone to take some pictures of me which I can line up on each of the models so that they could ‘glitch’ into placement occasionally - hopefully I can do this.
  • Maybe you need to re-visit some of your earlier experiments from last semester with the classical bust
    • I think what worked really well with the classical bust is that firstly its far more realistic than my models, so the glitchy aesthetic seems more out of place. It might also be due to the classical bust being more recognisable, more iconic, something that you really wouldn’t expect to be glitching as you’re already familiar with it.