Put together a layering tutorial for Edrian. I tried to make a photoshop action to speed things up. However, many of the steps need human input. The problem is photoshop doesn't let you do human input for these steps so I couldn't make a photoshop action that works for what we're doing. So I made this video tutorial instead that shows him how to do it.
The intersection problem in the middle has been fixed. It's now done in a way so that there's an equal amount of cyan, magenta, and yellow.
Let's explain the urgh. I'm actually annoyed that things aren't being done at the speed I want it to. And a huge part of it is because the original set up of the project made the entire process much, much slower than it should be. Hard and fast decisions weren't being made. What happens when decisions aren't made? Things don't get done. And what happens when things don't get done? Nothing happens. No progress, no nothing. We're better now, but I wish we got here sooner so that we ran into problems sooner. Then we could refine things further and improve everything. An example is how our project is about being physical as opposed to relying on something digital for interaction. Does the fact that a lot of the material we created use digital software contradict this? I think this is something that needs to be addressed in my exegesis.
Monday, June 17, 2013
Saturday, June 15, 2013
Second Layering Test
The great news is this test used Cyan-Magenta-Yellow triple and this triple works wonderfully. The bad news is that the layering doesn't work as effectively as I hoped it would.
Below is one of the tests where the spaces between the lines are 3pixels wide. In the intersection between the three shapes there are 3px by 3px squares. The other intersections are fine, in that you can still make out the form of the shapes fine. This is because there's an equal amount of the two colours in those intersections. In the intersection between the three shapes however, there's actually more yellow than there are of cyan and magenta.
Below is one of the tests where the spaces between the lines are 3pixels wide. In the intersection between the three shapes there are 3px by 3px squares. The other intersections are fine, in that you can still make out the form of the shapes fine. This is because there's an equal amount of the two colours in those intersections. In the intersection between the three shapes however, there's actually more yellow than there are of cyan and magenta.
Details of eggbeater gears
Realised some finer details we missed on the previous day about the eggbeater gears. There are lots of scribbles because later on I found that some of the things I wrote were wrong, so they've been corrected here.
Above is the gear that changes direction. We'll call that the change gear. This one needs teeth all 360degrees around. The sectors on the right show that the direction you rotate matters. Because you'll only ever rotate the change gear 300degrees. So the point at which a left/right gear needs to catch on to the teeth to make it change direction depends on what direction you're turning the change gear. And once you go in one direction you have to go in that direction the whole time. Turn it back and it ruins the set up, then it won't work properly.
As explained above, the left+right gears need to be on very specific points on the change gear so that the teeth catch at the right time. The diagrams on the right show this. I also explain why exactly half the gear has to have teeth on it. The answer is that only one set of teeth should be interacting with the change gear at any one time. If it's more than half then two sets of teeth would touch the change gear. The left and right gears would then be in conflict, stopping the whole thing from working. If it's less than half then you get an awkward pause where nothing happens.
This explains how to get the size of the left and right gear. I'm not sure if this is completely accurate, but that's up to Jenny to test out.
Friday, June 14, 2013
Potentiometer test
The potentiometers work with the RGB led. It should be fine with a 9v battery, but I'll test it just in case anyway.
Below are pictures of the circuit and a drawn diagram of it.
Basically the turning gear is attached to the "poll" of the potentiometer.
You can get a full range of colours just like when the colour cycle was controlled by the Arduino. True, controlling the colours with Arduino is more sophisticated while the potentiometer is more basic and crude. But that's the point of it. Reducing the interaction into something that is dominantly physical and less controlled through digital means.
Below are pictures of the circuit and a drawn diagram of it.
Basically the turning gear is attached to the "poll" of the potentiometer.
You can get a full range of colours just like when the colour cycle was controlled by the Arduino. True, controlling the colours with Arduino is more sophisticated while the potentiometer is more basic and crude. But that's the point of it. Reducing the interaction into something that is dominantly physical and less controlled through digital means.
Gears
Sorted out gears with Jenny. At first there were problems with timing and it didn't look like anything was happening. At least, not fast enough to meet the deadline. I began to wonder if the gears were actually that crucial. Does it add more to what we want to say about the project? The fact that it's a gear isn't important. It's the physical element in a gear that makes it relevant to what we're doing. So I'd say it doesn't necessarily add more, but it reinforces what we're aiming for in terms of a very physical interaction as opposed to a digital one.
We went through a long list of different gears that we thought might work. But the thing is they don't work. Either because the teeth don't make sense or the gear doesn't change direction. Below are drawings of all the failed ones. It's rather messy, but you get the idea.
In the end the set of gears that do work for what we want is the same as those found in eggbeaters.
Now that that's sorted we'll leave it up to Jenny to get that bit done and add any more developments.
We went through a long list of different gears that we thought might work. But the thing is they don't work. Either because the teeth don't make sense or the gear doesn't change direction. Below are drawings of all the failed ones. It's rather messy, but you get the idea.
In the end the set of gears that do work for what we want is the same as those found in eggbeaters.
Now that that's sorted we'll leave it up to Jenny to get that bit done and add any more developments.
Another RGB test
Different printers give different results. Got two sets of one test printed. One set on the home printer and another from the printing service at Warehouse Stationery.
There's a huge difference in how the colours came out between the two sets. I tested both to see how well coloured light works to make colours appear and disappear.
The home set didn't work as well as the WS set. This may be because the WS set's colour values are much closer to the colours I set on the computer compared to the home set. The colours from WS set are also far more vivid.
Part of this latest set was layering the colours on top of each other as well as testing triples.
Above is an example of the layering tests I did with the triples along with the video of that very same test. The layering I did works in revealing the forms of the shapes when colours appear and disappear. But I think there is a better way to do it. Here I've cut horizontal lines to each colour. The only problem here is that a section of the circle contains no yellow. A better way to do it may be with dots or a different way of setting up the horizontal lines.
The current triples shown here, in this chart are based on the previous test prints, which were done on the home printer.
Since the WS set is shown to be different I printed the full list of seven original colours in addition to cyan, magenta and yellow. Same digital values as the previous one, but printed by WS printer. This is to look for differences and to make the results more up to date. Below is the latest results. The top line refers to colours of light while the rest refer to printed colour. Basically, what happens is some colour turn black and some colours disappear depending on what colour the light is.
Triples are found when they meet these conditions:
-At any one time two of the three colours would disappear on a colour of light.
-All three colours turn black on a different colour of light.
Various triples have been tested but the best is cyan-magenta-yellow. You can see this in the video below.
The first row of colours are yellow, cyan, and magenta from left to right. Note that they all turn black on a different colour. Two of the colours also disappear on certain colours when one is turned black. Although yellow and magenta do darken on blue, the difference in darkness between the two is enough to set it apart from the other triples. With the other triples the difference in darkness is too small.
There were some key differences. For instance, as stated before the cyan-magenta-yellow triple worked most effectively compared to other triples in the WS set. When in the home set it didn't work at all when blue light was shone on it. Therefore, the set to use would be WS as the colours are closer to the values I set. And the triple to use would be cyan-magenta-yellow. Some other triples do work, but not as effectively. For example, some colours would darken at the same time as another. That you don't want.
An interesting thing about the WS set is that purple light turns certain colours (red, orange, yellows) into very bright neon colours (can also see this in the video above). This could be something that we can use if we need it.
There's a huge difference in how the colours came out between the two sets. I tested both to see how well coloured light works to make colours appear and disappear.
The home set didn't work as well as the WS set. This may be because the WS set's colour values are much closer to the colours I set on the computer compared to the home set. The colours from WS set are also far more vivid.
Part of this latest set was layering the colours on top of each other as well as testing triples.
Above is an example of the layering tests I did with the triples along with the video of that very same test. The layering I did works in revealing the forms of the shapes when colours appear and disappear. But I think there is a better way to do it. Here I've cut horizontal lines to each colour. The only problem here is that a section of the circle contains no yellow. A better way to do it may be with dots or a different way of setting up the horizontal lines.
The current triples shown here, in this chart are based on the previous test prints, which were done on the home printer.
Since the WS set is shown to be different I printed the full list of seven original colours in addition to cyan, magenta and yellow. Same digital values as the previous one, but printed by WS printer. This is to look for differences and to make the results more up to date. Below is the latest results. The top line refers to colours of light while the rest refer to printed colour. Basically, what happens is some colour turn black and some colours disappear depending on what colour the light is.
Triples are found when they meet these conditions:
-At any one time two of the three colours would disappear on a colour of light.
-All three colours turn black on a different colour of light.
Various triples have been tested but the best is cyan-magenta-yellow. You can see this in the video below.
The first row of colours are yellow, cyan, and magenta from left to right. Note that they all turn black on a different colour. Two of the colours also disappear on certain colours when one is turned black. Although yellow and magenta do darken on blue, the difference in darkness between the two is enough to set it apart from the other triples. With the other triples the difference in darkness is too small.
There were some key differences. For instance, as stated before the cyan-magenta-yellow triple worked most effectively compared to other triples in the WS set. When in the home set it didn't work at all when blue light was shone on it. Therefore, the set to use would be WS as the colours are closer to the values I set. And the triple to use would be cyan-magenta-yellow. Some other triples do work, but not as effectively. For example, some colours would darken at the same time as another. That you don't want.
An interesting thing about the WS set is that purple light turns certain colours (red, orange, yellows) into very bright neon colours (can also see this in the video above). This could be something that we can use if we need it.
Thursday, June 13, 2013
Key Theme and Animation Content
The story of the project is how, as a group, we struggled to find unity. Or at least a thread that connects each part of the project together. As actual people, it was trouble in finding something that connect our ideas, ways of thinking, and visions into one.
This aspect has dominated the project throughout this semester. Manifesting itself as conflicting opinions, difficulty in translating ideas into someone else's way of thinking, and sluggish, ineffective decision making. So naturally, in the content for animation this is what the story is about.
Now that we have a theme, we had to create symbols and imagery to represent how we feel about this semester. This is for the animation. First thing we did was summarising how the semester was for us in one word. Jenny's word was lost. This described her experience of the semester where she has an idea of what she wanted to do, but didn't know where to begin. Edrian's was confusion. His one is similar to Jenny's in that he could see and had a vision of what he wanted to achieve. However, he didn't know how to make it happen. My word was trapped. I knew how to do what I wanted to do, but I didn't have a vision of how it fits into everything else.
We also pointed out key traits that make us different to each other along with some similarities that made us become a group at the very beginning. This included our ways of thinking, personality, and work habits. Then they were all summarised into one package for easier understanding: animals.
Jenny's personality is personified as a bear. She has a dominant personality and is very opinionated. Also very assertive and sometimes stubborn. Has big dreams and enjoys eating.
Edrian identified most as a dog. He is very friendly and loyal to friends. Questions a lot; is curious. Also good at talking to people and making them feel comfortable. Style of thinking of strongly visual.
My animal is a cat. I'm independent and introverted, preferring to go off and do things on my own. More inclined to be experimental; is curious as well. I switch hobbies often.
With the theme sorted out and a starting point for inspiration in imagery we have enough to brainstorm on things going in the animation. At this stage we left it to Edrian to work on this aspect. I think it will go great because this is where he will be in his element.
Tuesday, June 4, 2013
Presentation and Final Plan
Presentation feedback was helpful and fair in my opinion. It's true that we don't have enough time to do a full blown complex animation that's layered. But there IS a better way which I'll cover later further down, I talked about it briefly in the previous blog post.
Some suggestions were offered to us both before and during the presentation. Prior to the presentation an option was to have still images where the animation would be revealed through changing colours. During the presentation there were a variety based off the still image alternative. For example, changing patterns on the wall or something similar to comic book style changes in image. So it's a little like an animation. Another was to simply have a series of still images revealed by different colours and to spin around or run around to animate them. These are "nice" but inelegant solutions on what we can do with hiding and revealing certain colours. These dumb down the idea and it isn't what we do. So we've clearly ignored these and come up with a better solution that fits what we want.
At presentation I showed the latest RGB light test. Some colours work, some don't. In the final it won't need the Arduino at all and to clarify, none of the colour orders would be predetermined. That depends on where the user turns the gears.
How do the gears connect to the lights? With potentiometers. Jason says you can get potentiometers that turn right around continuously. Before we knew that though I found an example of a particular gear type that can change directions (Chou, 2007).
We were planning on using that to be able to turn the potentiometers smoothly so that there's not a big change from full brightness to no brightness.
Andy suggested that the gears push buttons at specific points to turn the LEDs on and off. I think a saw-tooth shaped gear like those in escapements would work great for that.
Figure 1. An example of an escapement gear with saw-tooth shape. Retrieved from http://en.wikipedia.org/wiki/File:Scappamento.gif
However, a smooth transition is preferred rather than flashes so we're going to stick with the original gear in mind. This is also because we couldn't actually find any potentiometers that could spin around forever. As for taking apart the potentiometer, that's another option. However, it allows more chances for human error and wastage. We don't have time for that.
So what's the final plan? The final plan is to have gears controlling the turning of the potentiometers --> the potentiometers change the colour of light --> the colour of light hides and reveals certain printed colours --> moving printed bars on transparencies over them reveal an animation.
Content of the animation and the theme that pulls all elements together conceptually: struggling to find unity as a group when it comes to this project. So basically it's about unity and putting pieces together into one whole. It's the story of how this project is the way it is.
Reference List
Chou, R. [Roger Chou]. (2007, August 2). Gear System that Changes Direction [Video file].
Retrieved from http://www.youtube.com/watch?v=3aobPgGzB-U
Some suggestions were offered to us both before and during the presentation. Prior to the presentation an option was to have still images where the animation would be revealed through changing colours. During the presentation there were a variety based off the still image alternative. For example, changing patterns on the wall or something similar to comic book style changes in image. So it's a little like an animation. Another was to simply have a series of still images revealed by different colours and to spin around or run around to animate them. These are "nice" but inelegant solutions on what we can do with hiding and revealing certain colours. These dumb down the idea and it isn't what we do. So we've clearly ignored these and come up with a better solution that fits what we want.
At presentation I showed the latest RGB light test. Some colours work, some don't. In the final it won't need the Arduino at all and to clarify, none of the colour orders would be predetermined. That depends on where the user turns the gears.
How do the gears connect to the lights? With potentiometers. Jason says you can get potentiometers that turn right around continuously. Before we knew that though I found an example of a particular gear type that can change directions (Chou, 2007).
We were planning on using that to be able to turn the potentiometers smoothly so that there's not a big change from full brightness to no brightness.
Andy suggested that the gears push buttons at specific points to turn the LEDs on and off. I think a saw-tooth shaped gear like those in escapements would work great for that.
However, a smooth transition is preferred rather than flashes so we're going to stick with the original gear in mind. This is also because we couldn't actually find any potentiometers that could spin around forever. As for taking apart the potentiometer, that's another option. However, it allows more chances for human error and wastage. We don't have time for that.
So what's the final plan? The final plan is to have gears controlling the turning of the potentiometers --> the potentiometers change the colour of light --> the colour of light hides and reveals certain printed colours --> moving printed bars on transparencies over them reveal an animation.
Content of the animation and the theme that pulls all elements together conceptually: struggling to find unity as a group when it comes to this project. So basically it's about unity and putting pieces together into one whole. It's the story of how this project is the way it is.
Reference List
Chou, R. [Roger Chou]. (2007, August 2). Gear System that Changes Direction [Video file].
Retrieved from http://www.youtube.com/watch?v=3aobPgGzB-U
Subscribe to:
Posts (Atom)