This week I have been researching Composition and Color Theory. I have also been sketching to develop my idea, these concepts will be posted next week. Following below are some of what I have learned this week.
Composition
A painting or photograph needs good composition to lead the viewer’s eye and make the viewer feel like they are in the scene. Leading the viewer to a Focal Point is essential, composition is a tool for this purpose. Placing objects in a conscious way, creating shapes and pathways for the eye and create balance is important, this leads the viewer to certain elements of the image. These paths can be blocked off by objects to avoid leading the viewer out of the image frame, but instead make the viewer focus on the Focal Point. Giving the Focal Point area the highest contrast, by using the lightest and darkest values in that area of the image is another effective technique. The details in this area is also more important than in the rest of the image, because it is where the viewer looks.
Following are a few of my notes from the book d’artiste: Matte Painting(Ballistic Publishing: Altiner, Cole & Stoski), regarding Composition:
-Add recognizable objects to the scene to make it more believable for the viewer.
-Making the foreground darker and the background/middleground lighter leads the viewer to the lighter area. If you make the background dark and foreground light, you can make the background more unknown and scary looking.
-Its is good to use “tricks” to separate the foreground with the middle and background. Maybe adding a slight almost invisible blur on the edges between them.
In a CGSociety tutorial on Composition in Art, Philip Straub explains the following:
http://features.cgsociety.org/story_custom.php?story_id=3275
«Without a place for the viewer’s eye to rest, the painting will lack impact and won’t hold the attention of your audience.»
«Since most Environment Concept Artists work in the entertainment industries, its expected you will be asked to create cinematic moments or “memorable moments” utilizing the environment as a stage.»
Rules of Third
The Rules of Third is a technique adapted from the original Golden Rule, but is slightly altered to fit photographic proportions. It involves adding two horizontal and two vertical lines over the image, splitting it into 9 pieces with equal proportions. This creates 4 »power points» where the lines are converging. These can be used to lead the viewer toward the Focal Point in the image. Important objects in the scene are placed on or near these points, guiding the eyes of the viewer to important elements of the image, creating a balance. Normally the area with highest interest and weight in the scene is put directly at a power point, while smaller objects like trees, birds or houses are placed close to, but not directly at some of the other Points.
«The two best “power points” are the Upper Right and the Lower Right because the eye enters the picture frame at the lower left hand corner of the picture frame, travels to the center of the picture area and then reaches the right hand ‘Golden Mean’ position where it stops to look at the ‘Center Of Interest».
«The reason the eye enters a picture at the lower left side is because we are taught to read from Left to Right. This is a psychological fact that has been proven over the years. Next time you’re in an art gallery or art museum that shows the Old Masters paintings, notice how many have the Center Of Interest in the “Golden Rule” positions.»
The Circle
Circular composition can also be used to lead the eye toward the focal point in the scene. It can be used more obviously by adding objects with circular shapes that gradually become smaller and aims toward the focal point. It can also be vaguely suggested. Circular composition is a good way to create motion in the piece.
Dynamic Composition
This type of composition involves tilting the camera to either side. This creates a more dramatic scene. This approach is especially useful when you are trying to depict action in your environment.
Color Theory
In an environment image the sky gets warmer as it gets nearer to the horizon line, while the land gets cooler as it gets closer to the horizon. The area in the foreground of the image is high in contrast and saturation, this decreases gradually when you get farther into the image.
Shadow areas in an environment often has a color ranging toward the cool spectrum. A complimentary color can be used.
Color contrast
Most of the ground elements of my image will be in a green color, but some of the dirt and rocks will be toward the red. This creates a a complimentary contrast of Green/Red color.
Color/Value perspective
Color perspective gives the image depth. The farther objects are placed into the picture, for instance mountains, the more sky color they get. Mountains visible in the far distance almost to vague to see, basically has the same color as the sky, and get less saturated.
Lighting
I want my image to be set early in the morning, when the sun isn’t very high in the sky, barely visible over the horizon line. Depending on the geometry in my scene, I might add Volumetric Sunlight Pass in Maya to create fog and haze. If this doesn’t work I will simply paint it.
In the book d’artiste: Matte Painting(Ballistic Publishing: Altiner, Cole & Stoski) Alp Altiner explains the importance of correct Light & Shadows in a Mattepainting, on page 90:
“The most important part of making your painting photo-realistic is creating believable light and shadow. Just the right amount of light, shadows, and highlights will bring a photo-realistic feel to your painting.”
-Øyvind-
I have now updated my research document with live footage and 3D implementation research.
Live greenscreen footage:
I will use live greenscreen footage as a part of the 2.5 mattepainting. Using the software Eyeon Fusion and the the Ultrakeyer function the footage will be keyed into the forground of the mattepainting. Color corrections, fog and various other adjustments will be added on the final composite.
Hopefully I can borrow a camera with high resolution. HD720 (1280×720) is an ideal quality. It will not be good for the realism in the image if I have to scale up the video.
Following are some excerpts from a Wikipedia article on High-definition video:
Source:http://en.wikipedia.org/wiki/HD_video#HD_content
“Older (pre-HDTV) recordings on video tape such as Betacam SP are often either in the form 480i60 or 576i50. These may be upconverted to a higher resolution format (720i), but removing the interlace to match the common 720 p format may distort the picture or require filtering which actually reduces the resolution of the final output.”
“Non-cinematic HDTV video recordings are recorded in either the 720p or the 1080i format. The format used is set by the broadcaster (if for television broadcast). In general, 720p is more accurate with fast action, because it progressively scans frames, instead of the 1080i, which uses interlaced fields and thus might degrade the resolution of fast images.”
“720p is used more for Internet distribution of high-definition video, because computer monitors progressively scan; 720p video has lower storage-decoding requirements than either the 1080i or the 1080p. This is also the medium for High Definition Broadcasts around the world and 1080p is used for Blue-Ray movies.”
This is something I should have in mind when filming and encoding my final video. Ending up with Interlaced or otherwise heavy compressed video would ruin the blending of live footage, 3D elements and painted areas, making it look unrealistic, pixel distorted and therefor separated.
Codecs
Source:http://www.larryjordan.biz/articles/lj_video_codec.html
“As a sidelight, codecs are often divided into “lossy” and “lossless.” A lossless codec preserves all the original image quality so that when an image is restored it is indistinguishable from the original. TIFF and PNG are examples of a lossless codec.”
“A lossy codec “throws out” visual information as part of the compression process, which means that the compressed image does not have the original quality of the source. All video codecs are lossy.”
Since I’m making a video clip I use a “lossy codec”. My final video will be watched on the computer and should be suitable for viewing on the web. For this I should keep the filesize within limits. It is important that the video isn’t laggy and slow. The H264 codec is suitable for this purpose, since it is designed to reduce file size and decompress quickly.
3D implementation:
When adding 3D rendered elements to the mattepainting, it is crucial to make sure the direction of the sunlight is correct. The light clusters in maya for skylight and bouncelight must also have the correct intensity and colors. For the bounce lights from the ground for instance I need to take colors from grass, rocks and the surroundings to make the 3D object blend in and look realistic in the scene. I have to make sure the surfaces absorbs and reflects light in a natural way, based on the objects material type, its surface color and the color of the sky/bouncelight. The same thing goes for the skylight, simulation the colors of the sky above.
Rendering
Renders in Mental Ray with the correct settings will give a much better result, but will take a lot more render time compared to Maya Software. Since my main 3D objects doesn’t have any animation I only need to render one frame. For this reason Mental Ray is a good choise. I will do test renders at a small resolution first, then increase to full 4096 pixel size when it looks correct.
My animated element will be quite small and have a simple topology and texture. Mental Ray should be good here also. This all depends on the size and amount of the animated objects. Maya Software will give the same result if I for instance animate a flock of birds in the distance.
I will add occlusion, specular and bump map to the shaders, then paint over the renders in Photoshop if needed, to better blend them in. The renders will be done with an alpha masked background so it is easy to bring them in as layers into the Photoshop file.
-Øyvind-
The Milestone 1 documentation is completed. To access it press the lightblue link to the left, or use the menu on the upper right. Ive got a good start with developing my project and have done some valuable experiences over the past few weeks. I now continue forward to Milestone 2 with an enthusiastic will to consume more learning.
More concept
Yesterday I worked more on the concept:
I tried to make the canyon rougher looking compared to the previous concept. But I feel I lost the realism of the scenery, and the colors are somewhat off. There are no real(man made) objects for reference in the image. I want to focus less on the canyon aspect and do a simpler version with trees and more rocks on the sides, make it into a dried out river, maybe with roads and a city. Or add some tall mountains with snow in the far background. I will continue working on my concept.
-Øyvind-
This week I have continued looking for reference, sketching and reading though the project literature. Ive been researching the techniques of how to build the mattepainting in Photoshop. It involves using Adjustment Layers and Clipping Groups to have interactive control of the color and detail adjustments. And the use of Layer Sets and multiple versions of the document, to keep it organized and to save memory. I will be working in a 16:9 format, with a with of 4096 pixels. When I use allot of layers and adjustments it will be very memory demanding, therefore researching and finding solutions on how to minimize this will improve my workflow and saving time. Having a clear view of the worlflow for a mattepainting is crucial for my planning process.
In the book d’artiste: Matte Painting(Ballistic Publishing: Altiner, Cole & Stoski) Dylan Cole explains the following regarding his work on the piece “The Tower of London”, on page 20:
“The key is to layer the PSD to give a lot of flexibility. When darkening images
down this much, one must be wary of losing bit depth. That’s why I work with
a lot of adjustment layers and layer masks, so that I always have the freedom
of change.”
On page 26 in the same book Dylan Cole explains about the piece “Desert City” and writes:
“A painting of this size can be very cumbersome in Photoshop with all
of the layers, so I like to make use of layer sets so that I can navigate
through my layers easily. I also save out multiple versions of my
painting so that I can flatten certain areas, yet retain all of the separate
layers in another file.”
Concept phase:
I want to do a deep volcanic canyon of some sort, with something important at the base of the canyon, or maybe some sort of constructions in the background. I’m still working on the idea/setting. It will probably be for a fantasy or modern time movie.
This is a concept collage of my shot. I might move the perspective a little to the side to make a better composition:
Some reference, I haven’t decided on time of day and lightning/color conditions yet:
Camera mapping:
By looking at tutorials from the Internet Ive been researching the camera projection techniques, making my first test render. I wanted to get familiar with the technical side of camera mapping and projection as early as possible, since this is a new field to me, making it very hard to plan how long it will take me to do in the final camera mapping. By becoming familiar with this I will become more efficient later on. This is important for my future project planning in regards of time.
I took a photograph of my Wii with a DVD cover on top, this was imported to Photoshop. A cutout of the Wii was copied to a separate layer. I used the clone tool on the floor layer beneath to make it all floor. No I have two layers, one with just the Wii and one with the floor.
I set up a scene in Maya with the original photograph as an image plane. I created a polygon box roughly the size of the Wii. Then I set up a camera with an aim point, and matched the camera angel and Focal Length with the image plane. I called this Camera 1. I created another Camera named Camera 2 with the exact same settings and snapped it to the exact same position. In Hypershade I made two new Surface shaders. I used Surface shaders because it isn’t affected by the maya default lights the same way a lambert is. With the lambert the sides became almost completely dark. Surface shaders makes a better render for texture projection. One shader has the Wii cutout image, the other just the floor. I added these with an image node with filter set to none, for a smoother render result. A place3Dtexture node was connected to the projection node, (output:world.inverse.matrix and input: placement.matrix), then the projection node was connected to the shader. In the “projection attributes”, Proj Type was set to Perspective. I linked it to camerashape2, meaning the texture will be projected from the camera named: Camera 2 . Fit Type was set to “Match Camera Film Gate” and Fit Fil set to Horizontal.
Now the two textures are projected from camera 2. I animated camera 1 for 80 frames and did a render. First I had some problems with the texture on the sides, but I fixed most of it by changing my initial projection settings. I didn’t line up the box perfectly, so the edges are a little off. I should have done a bevel on the side also.
This is what a render looks like:
Here is the test clip
Reference:
When researching camera projection I used these online tutorials:
http://www.digitalartform.com/archives/2004/11/camera_projecti_1.html
http://www.g-3d.com/Newer_AAC_3d_499_Cam_Map.html
d’artiste: Matte Painting:
http://www.ballisticpublishing.com/books/dartiste/matte_painting/
-Øyvind-
This week I have made some changes in the project plan. Ive continued reading from the litterature list and started to look for reffarance images on the net and in movies.
-Øyvind-