Shot 1 Background/Scene
The last shot is the same as first one, the same background and scene file was used.
Just another myblog.arts site
Category: Advanced and Experimental 3D computer Animation Techniques 1
The final rendered film, DPI was set to 70 and resolution to 1080p due to lack of time to render and render farm being to populated. I will render this again in higher resolution and camera settings but due to how close it was too deadline this could not be done before the submission. O
Overall I am happy with the final outcome, if I had more time I would like to add fog and particles to the scene it would add to the tone of the project.
All the backgrounds with lighting and post effect.
The extreme lighting effects such as the glow in shot 6 and the light in the image on the left were produced by manipulating intensity and exposure using a point light.
The floating body parts were animated in two different ways, they were grouped together in different groups based on their proximity to each other and the group was animated either using key frame animation and moving them manually into position or placing pivot in the center and animating setting key on different axis translations. The body parts on top of the pile were animating using method two and the floating head and the smaller body parts in the distance were animated using method one (key frame animation).
A total of 52 individual groups were used in the scene, about a third of them were animated.
The image on the left is the first shot, the background and all assets were made from the rig that was taken a part. The mountain some of the rocks and structures were modelled in Maya.
The textures for the fossilized body parts and the mountains were made in substance painter by me. The green moss texture for the surface is a free moss texture asset.
Mountain without texture
Mountain with texture
To visually symbolize repetitive failure I had this idea of implementing the rig into the scenery. The rig skin was unbinded and taken apart multiple copies were made of each body part in different sizes.
The main scene will be a desolate mountain area with piles of body parts scattered around, fossilized and part of the natural scenary implying that they have been there a long time.
The mountain was made using the sculp tool on a surface plane using a stamp. #
Duplicating the body parts hundreds of times made the scene very heavy to load and work on, to work around this the body parts were put into various groups and exported and then imported back to maya. This drastically reduced file size from 500-600 mb to 15-30 mb making each shot much easier to work with.
I was unable to secure funds for the assets, due to lack of time the decision was made to change the project to something more realistic given the time frame. Instead of idea 2 the artefact project will revolve around the body mechanic shot. I will contextualize the body mechanic shot thematically and expand on it. The body mechanic shot as it is now is a parkour wall climb that is executed successfully, the character then proceeds to celebrate which leads to him falling down the edge of the building he is standing on. The theme of the story is success and failure, this will be expanded on. I want to explore abstract imagery and aim for aesthetically pleasing visuals.
These are all the assets needed for the project, they are available on cgtrader and collectively cost $44.
The last two images are from the same asset collection which includes all the assets seen in the images, I will also model some assets that are not available cheap or for free. The buildings will be re textured.
The video above shows how to fake the tilt shift effect in Adove Premiere, the process is adding gaussian blur with high blur settings, use a feathered mask to select the area of focus. Although this method is simple and easy to use it will not distort the perspective and it is simply a gaussian blur effect with a wide area of focus, I will not be using this method.
Pixar has published a tutorial on how to produce the tilt shift effect in Maya using Mayas camera settings. I will rely on this method as it is more accurate and produces better results.
The process is as follows, create a polygon triangle and create a cluster handle on each vertex, select the vertex then deform the cluster using anim deform menu. Create a locator and snap it to the vertex and make the cluster handle a child of the locator so the the locator can move the cluster and the vertex.
Create and select a pxrcamera, load the camera and a transform locator in the connection editor and connect the locator to the pxrcamera focus.
If done right this should be the result. Using this method there is no need to model everything in miniature models instead normal size assets can be used with this effect to produce miniature look.
Eye brow and eyes were adjusted, random movement of the hands were fixed on ones and twos and the left hand has now been parented to the curve controller of the right wrist to keep the hand steady. Mouth shapes for O and the word damage has been adjusted so that they can be understood more clearly. The gun also now slams into the open palm of the left hand causing follow through action, more follow through is added to head and upper body movement, upper body neck and no longer move at the same time slight overlap has been added. When the character is pointing the gun at the end she is now squinting with one eye closed.
Compared to the body mechanic shot this turned out better due to better planning and better reference, however I still feel that I could have done better if I had planned my shot better and recorded separate reference footage for the face and mouth shapes. The slightly better preparation made a significant improvement on this assignment compared to the body mechanic assignment.