Monday, July 26, 2010

meowmix


Hoping to develop some of these guys in the coming days.

I've been reading a lot of books lately. I just finished up one on Aubrey Beardsley and Kim Elam's Grid Systems and am about to start on her book on typography. I'm also working through a book on pre-revolutionary Russian graphic design and just picked up Gruppe on Color and The Art of Robots today from the library. Gruppe's book is really amazing. I've only just started it but right away started taking notes - there's so many good things he says in it! I'll likely be posting some of my favorite quotes when I'm done with it too, seeing as how it's out of print and pretty expensive to acquire.

as of late

Relatively recent miscellanea!



True story. Good luck trying to copy/paste pieces of money in photoshop too. Had to scan in with acrobat reader and the shift the hues before copying and then shift them back again after pasting.


3 unrelated images layered up in photoshop. Fun stuff.


Saw this dude in the student union on my way to the computer lab. Instead of doing whatever I was supposed to I had a new important mission. You just can't pass up purple polo w/ sweater vest.

Sunday, July 25, 2010

if - double



For Illustration Friday, topic: double.
Kind of cool, kind of creepy. Happened upon some neat stuff in making this.

still life


This is the last post from the sophomore projects, I swear. The still life project was our last one, and one that I thoroughly enjoyed. It was nice to get such a project to finish the year out with. Not to say this project was a walk in the park, but texturing is very different from animation, and shares so much with drawing and painting, areas I'm far more experienced and comfortable in.

This project was a short and simple one. We were given files of assets, plain models and some simple props/sets that we were free to slightly modify/fix and arrange into a composition to light and texture. I ended up using a total of 8 lights (6 spot, 2 directional), and it was my first time using raytracing. We had to create shaders networks for all our materials, and it was my first real time getting into that too and building them from the ground up. I tried to keep things simple and render times low (I think at the end it took about :48 seconds to render the frame with depth of field, but then again the computers we use are wicked fast). I worked to keep all my maps low. Everything is procedural except for the fruit and paper towel bump, which is a small tiling raster pattern that I painted. The outdoors, bananas, apple, and pears are 1024 maps, the orange, cherries, and grapes are all 512. My knowledge of photoshop was very handy on this project, though sometimes there's just no substitute for enhancing and painting noise patterns yourself.


Not many renders/paint overs ended up getting saved on this project, but here's a few I found.

Here is a video showing, quite literally, a day's work. A lot of frames were dropped with the screen capture while scrubbing through the day's saved renders, and the quality's not great enough to show the subtle changes in things, but it gives a nice overview of the progression regardless.

Reference was of particular use of this project. Most of it was high res fruit photos pulled from flickr, but a fair amount too was just reference for color, mood, and design aesthetic. These three were some of the ones I kept open a lot to keep my head straight and stay focused.



Tuesday, July 13, 2010

etchings


Yet another post/image dump from last semester, this time from printmaking II, aka etching. I really enjoy printmaking classes. It's nice to get out of the computer labs and work with some real materials. It's also a great class to simply try stuff out in and play with ideas.


My first project was my most ambitious. During the winter I looked at a lot of beautiful work from the amazing Charley Harper (including The Giant Golden Book of Biology). When I got back from winter break I started drawing these simple birds that became more and more stylized and streamlined as I drew them. One day at work I drew some clustered together on a post-it, which ultimately became the basic prototype for this print. You can see the drawing colored over in photoshop at the bottom of this post. This project print used a three plate color separation and aquatints (more on the process of this piece at the bottom of this post). It's an edition of 5 prints which are signed on different sides - as there is no distinct "up" or "down" to these prints as they can be viewed from any side. Some of the birds are based on real life ones including pelican, seagull, flamingo, killdeer, and kingfisher.



Next I did a series of portraits of my CA friends. This was induced by my work at portrait night during the first half of the semester. I wanted to bring what I was doing there over into prints and at the same time explore the concept of portraiture through technique. The prints are on a full sheet of paper (22x30") and the print itself is made up of 9 separate coper plates that are 2x3". After several experiments with ink I decided on a mixture of ink base with a touch of black which gave nice gray prints that looked almost like graphite. The extreme amount of base had the unexpected effect of yellowing into a gold color within a few days, which I came to like.

For these plates I used a process called deep etch where you block out the plate and etch it for a couple hours, leaving large areas exposed to the acid which flat bite. The areas that flat bite consequently don't hold onto the ink and mostly wipe clean, leaving some ink clinging to the inside edges of the recess which then prints. Highlights were painted out while my friends sat for me and then after the asphaltum had dried I went back in with a scribe drawing in more detail from a photo.

Special thanks to my friends who took the time out to sit for me!
Tim Palmer
Adam Campbell
Josh Garlick
Matt Sullivan
Gordon Pinkerton
Betsy Bauer
Stevie Lewis
Alberto Beguerie
Uri Lotan
You guys are awesome!



My final project came from looking at glitch art for the first real time. It's everywhere and it's awesome. Most of my time on this project was spend in research and idea development on the computer because I was so unacquainted with the genre and hex editing (another post to come on that later as well).

I find the idea of translating something that's not am image into an image file to display graphically. Originally I wanted to get a RAM dump off of a computer, but my googling around told me pretty quickly that something like that would be near impossible with my level of computer knowledge. So, I turned to other files like logs, but couldn't find any that met my expectation, so I ultimately went with notepad.exe and translated it into a bitmap image by writing he image header and then dumping the file into the body of the file and filling the end with the necessary zeros.

The print itself is made in the following way: I printed out an enlarged and photoshoped version of notepad.bmp on the laser printers and then lacquer transfered the image to japan paper by hand so I could get a more imperfect transfer. Then I roughed the paper a bit, taking off the top layer of the transfer in a few places, then I stained the paper with acrylic gouache. Then using a template I cut out the rounded square from the japan paper which would ultimately be the the screen. The actual etching is a simple line etching printed with black ink. The substrate, as you can see, is a cardboard box. I saved all my cereal boxes through the year to recycle, so I conveniently had quite a few hand by the spring. On each box that I printed on I pulled up some of the top layer of the paperboard in the lower left corner (the signature and numbering of the edition is done in a similar way in the right corners). To print I dampened the cardboard with a spray bottle of water and ran it through the press with the inked up plate with the japan paper square with some glue on it's back over the computer screen. Printing it presses the japan paper to the cardboard, gluing it, and then the image is printed on top at the same time. The process is known as chine-colle, and is really cool. I'd like do do more work exploring it's possibilities further.


This is what it looks like to make an etching. That jar has asphaltum in it. It's very much like asphalt and full of tar and all sorts of other wonderful things. Whatever you block out on a plate wont react when placed in acid. The acid eats away at the uncovered area and creates a recess. During printing the ink gets in these recesses and stays once the plate is wiped. Lines and areas that aren't real thin will flat bite if they are not given a texture through a process like aquatinting (we spray the plate with a coat of spray paint =~ 50% dot pattern which blocks out those dots from etching and gives the plate some "tooth" to hold onto the ink in those areas when wiping. Because my lines weren't thin I had to use the aquatint process to create then and so had to paint around all my lines and other dark blue areas. Doing that took a very very long time, much longer than I ever expected (about 10 hours per plate if I remember correctly, probably more for the first one). To draw out the birds (because I didn't have a draft made up separately) used a waterbased marker and drew directly on the plate. Once I was done blocking out my plate I cleaned off the marker and tarnish that had developed in a water/vingar rinse and a light polish.


At this point my main plate looked like this. I then applied the spray paint for the aquatint, and once that was dry I put it in the acid to etch for several minutes.


The etched plate looks super cool when cleaned :). Next I ran a test print of this plate, then scanned the print, and then created the mock up for the colors/tinting/separation in photoshop. I then separated each layer of color and printed off a page with all of them so I could use it as a diagram to follow while working, knowing what to block out, what to not block out, and to make sure I didn't forget anything. The red plate was extra tricky because I wanted two levels of red, a light, and a more intense one. To do that I had to etch the plate twice, blocking out all the white areas first, then the light areas for the second etch.


When making the color plates I would have to print the main plate on paper, then run it back through the press with the color plate exactly where the main one was to transfer the image to the plate. Once it was transfered I took a scribe to delineate the edges of that shapes that I intended to block out (the light lines scratched into the plate above). Once all three plates for the the blue, red, and green were made I could print them all together. To do this all plates had to be inked up and wiped and then printed one after the other in the exact same place so they would register. Because of the inks I was using wiping the plates was taking way too long, but paper wiping helped to pick off the plate tone much much quicker. Registration was such a pain. The plates, while really quite close in size, didn't match up completely, despite being cut and filed carefully, transference of images, along with drawing over them, only served to compound any possible error, plates and paper could shift slightly in printing and just to make matters worse, the plates actually started to warp during the printing of the edition. All in all it wasn't too bad though, and I was able to touch up a lot of spaces with watercolor afterwards.

Monday, July 12, 2010

concept2


Story's not my strong suit, but here's some shots from animatics from concept design class!






Sunday, July 11, 2010

sitStand MEGAPOST



As promised here is the giant sit stand post. Of course this still isn't everything, but it surveys the process pretty well. Not pictured are many many other screenshots, renders, tests, diagrams, and sketches. Oh yeah, and there's 556 maya save files, including model, rig, props, lights, scene, and a great many incremental saves. These are mostly divided up by subject, but of course a lot of this stuff overlaps, especially toward the end.

First comes story, that's where you always start. Along with that comes character design, figuring out how everything's going to look, how everything's going to play out in terms of posing and acting, and how it's going to be presented and staged. At this stage too it is important to look ahead to the technical requirements of your model and set. What is everything going to need to do and are you going to able to make it do that? An example of this is the expression sheet which is great for getting a feel for what the character's face needs to do so you can know what you need to achieve with your blend shapes.



Starting out in maya we modeled our characters. This was my first time modeling a fullbodied/human character and my third time modeling with polygons. Modeling isn't exactly my cup of tea and I'm rather slow at it. It's kind of cool though to build your guy from the ground up though even if 3d is rather dead and unappealing until until things start to get rigged and textured.



If I remember correctly I modeled Carson from a cube using a lot of extrudes. He's modeled using a proxy that would take the low res poly mesh half and create a subdivided smoothed copy and mirror it across the axis. At this point too you quickly run into situations where you have to make decisions about how to translate 2d into 3d. I had to look at my flat illustration many a time and ask myself "what does this shape mean in 3d?" Pictured above is one of those instances where I had to interpret the nose shape. I went with the one on the right.



A lot of the time I had the pleasure of getting the nitty gritty of the mesh worked out. Working to get everything into quads and create a good edge loop flow isn't exactly a walk in the park. Pages and pages from my sketchbook have scribbled mesh diagrams in them from this period. Sometimes even that wasn't enough and I'd have to print off screens like this one I compiled to think through things with paper draw overs. Click the picture to view a .gif of the face mesh at this stage.



This is a screen of the "final" model textured and without a smooth node. (I say final because I later realized that in the excitement of being done with my model I forgot to go back on top of the symmetrical mesh and make the pant cuffs uneven, which I later fixed.) The hands were a pain to do but things became much better when Beavan showed me this awesome hand tutorial. It's not in english but the pictures spell out everything. The feet were a lot of fun and I'm rather happy with how they turned out. The foot is made from his hands. I copied the hand and mostly just pushed and pulled and scaled things to make it into a foot before I attached it to his leg. His geometry is all one piece, except for his hair, and of course, his eyes.



After modeling comes rigging where you set up all the joints, deformers, and controls for your puppet to move. After first semester we really don't do any real rigging. We used the setup machine for our characters which is a paid for third party plug in ($40 I think) which automatically rigs characters after you fit widgets over you mesh, but even that has it's own tricks and issues to learn to work with. And of course after that you still get all the joys of painting weights and tracking down all those little influences that find their way into vertices they don't belong in. Here are some screenshots.

1. It seems like this only really happens to me... but when I'm modeling in polygons I get graphics card issues when using a proxied mesh where whenever I extrude or cut faces parts disappear. Things look normal again but it's still kind of annoying to restart maya every 15 minutes when modeling.

2. Always remember to hide your joints. Weird stuff happens when you put a key on one by mistake...

3/4. It's just fun to mess with rigs in ways you're not supposed to :).



View for a .gif on what happens when you forget to parent eyes in a rig.



Next is blendshapes which is used to rig the face and uses target geometries which have been sculpted in their own ways and deforms the character's face to match the deformations on the target geometries. I wasn't really expecting much out of blendshapes, I just knew going in that I wanted a good set because I found I enjoyed facial animation so much while working on the animation exercises at the beginning of the semester. Turns out I love making blendshapes. It is perhaps one of the more tedious things to do and particularly vulnerable to some technical and user screw-ups but for me it was the first time I saw life flow into my character and it very much revitalized my enthusiasm in the project. Used for this were some very helpful and time saving taper scripts as well as js facial scripts for setting up the controller sliders.



Posing is really just a lot of fun. Often while I was animating and navigating around in the perspective cam some possibility or another would spark my interest and I might just tweak a few things or just use that particular perp view and render it out. Since I used a painting for my background instead of modeling a set and used a 1024 texture map for Carson my render times were very low, I remember waiting at most 15 seconds for a render and usually they were much closer to 5.






Background time! Early on I decided that instead of modeling my background I wanted to paint it. I probably could have stood to model it but it did teach me some very good things about lighting and light quality and applying the concepts of color and value to 3d.



Click for background progression .gif!




This was the first time I had significant issues lighting my scene. I found it very difficult to light Carson since he's so tall and skinny. You've really got to fake a lot of the ambience so that he doesn't look so dark but at the same time you want to preserve a sense of strong afternoon daylight with solid shadows, while at the same time not blowing out highlights on the arms, shirt, and feet. My most annoying problem was dealing with the issue of the hat not casting a shadow at the top of his bangs. I tried messing with every setting on the key light that I could before throwing in the towel and emailing my teacher. Apparently the light was simply too far away so I supposed that just messed up the shadow map calculations. d'oh. The top render illustrates the shadow issue I was having while the lower render shows early lighting and test compositing. The dark stripe is due to trouble with transparency in after effects and I soon just used a surface shaded plane with the image mapped to it and rendered it together natively to avoid the issue.



Some lighting test renders for after effects taken into photoshop and had fun with instead :).



The overall scene with cameras and lights.



Prop modeling! All props had to be modeled in NURBS. This chair probably took 5-10 hours to model. Lots of little pieces.



The cinderblock obviously didn't take as long. I often made plans/thinking diagrams like the one above on paper. This one didn't even get followed really because I changed my mind on certain shapes, but it's always a good thing to figure out how you're going to do something before doing in to do it. Displacement maps are also super cool when rendered at their default settings.



Part way through and all the way to the end you work on animation. The screenshot above shows how I like to work a lot of the time. Having maya on that whole monitor is just too much maya for me. I like to keep it smaller and then put other stuff around it like reference, related inspirational art and photos, drawings from the project, and text files with notes. It helps keep me focused on what I want to get across with this instead of getting sucked in to working ahead in maya without the end goals in sight.



Above is one of the time when I got so stuck with the animation I had to just walk out and figure out just what the motion was. I ran through it a bunch of times and different ways in the lobby, sketched out the intermediate poses, made notes on the timing, even went back and drew over screenshots with notes too work out the silhouette and details further. I still got nowhere. I just couldn't get that arm reach to work and animated it from scratch a few times. At that point I just had to step away and move on because I wasn't getting anywhere with it that day. Later on when I went back to work on that section I realized that that action I had wanted wasn't even the right kind of action for this situation. It's yet another example of unwittingly getting caught up in the here and now of detailed work instead of keeping your eye on the overall piece at hand.

Here is a 3 min (=~16MB) movie illustrating the animation progression from some of the many playblasts that I made and what I remember from the time.
http://webspace.ringling.edu/~emcmahil/misc/animProg.avi

And of course, here's that director's cut, which you can thank Adam for:
http://webspace.ringling.edu/~emcmahil/misc/emcmahil_ssDirCut.mov