Ravine: Animation Sequence

Final Animation: Playblast Preview

   A preview animation is underway. For now here is a screenshot of what has been developed so far.

Reworking the environment

   With a few hours to go before the deadline there was a little discussion on the development of at least some kind of animation. This was set back however with a disagreement over the scene that had been produced. The discussion was resolved with a look into adding to an earlier environemnt model

  With two and a half hours before the deadline, creating new textures was a tall order. Instead focus was made on adding the constructed assets into the version 2 environment.

  When complete two render sets were made: The first set above are ambient occlusion passes while the set below are the objects with a default shader but illuminated with an ambient light.

 There was some discussion on how upward travel was being conveyed. In the scene presented in preview render passes with this matter was resolved with a dead end. Exploration was made using geography to imply the only way is up in an alternative direction to the use of a dead-end style barrier of rock to imply the fish could not go further.

Coral and Scale Development

   Time was spent correcting the UV map for the coral. The reasoning for this was an attempt to have the shader's progression up the coral correlate better with the contours of the geometry.

  The methodology seemed straightforward: Cut apart the UV map at each split zone, modify the UVs to give a flat-on shape, straighten everything out. As can be seen in the picture below, the new strategy allows the gradient map to better follow the flow of the model.

  Using a shader gives me some freedoms, as it means we do not have to completely unfold the model in order to get the desired result and anywhere that appears non-existent (such as side-on faces) on the UV plane is not stretched to moment-ruining distortion.

  A after a tweak to the shader the quality is solid enough that it can be consideredpresentable. As well as well as a sequence showing the coral growing from a bud, a turnaround was also created to show the model as it would look in the final render.

  As well as these tasks, The strange growths to the body of our fish were added to the body. The initial plan of using a lattice as discarded. This fell flat when the eye wasn't considered in the original plan, but this was compensated for, as the manual application turned out to be rather therapeutic and progressed relatively smoothly and promptly.

  Above is a breakdown of the animation, with each line being a cluster of blendshapes.

Coral Reacher Nearing Completion

  As on schedule, the coral has been assembled, staged and animated. I might look into rigging and perhaps a controls for animation purposes but at present this is the construction so far. There was talk of including a gradient on concept art and there may need to be a little of colour modification for the final model, the core animation however does appear solid.

  In terms of geometry this is not too complex a model when a single instance is around 1.8MB of data. That may multiply considerably when there are lots of these but as mentioned in an earlier post but work was put into minimising the geometry so that lots of these could be on screen at once.

The Primary Hazard

  The development of the scales generated enthusiasm. The task today was making the coral/organism that would chase after our fish. Originally the plan was to use coral, but there were a few scenes that involved the coral reaching out. Doing some research there was some similarity to the tendrils of sea anemones. Which are a lot more flexible.

  First tests however resulted in a few errors from the way the joints were placed down. These took some time to rectify and were a little annoying.

  Initial attempts at a cleaner model involved joints at every lateral edge, but I discarded this as a control group would have 30 different joints to manage.

  To cut down the number of joints, instead of one joint per lateral edge I looked into one joint for every two lateral edges, which halved the original number.

  Each joint has three controls: X, Y and Twist. One bends the joint forwards, another sideways and the twist helps with curling. It is very flexible with little deformation. Hopefully it is simple enough to be used en masse. And it should also be useful for wrapping around things.

  The final parameter I added was a length control. Unlike the others, this one control affects the entire skeleton, which could really help with any scenes of the construct launching itself at our hapless fish. 

  The coral wasn't entirely forgotten about. I have developed a fairly sound method of reproducing the branches of a tree, which are very similar to branched coral. The plan is to work in a blendshape that will have the coral branches growing from their individual points of origin. 

Testing the Alien Scales

   My primary task this afternoon was to get a working system for having a set of alien scales grow from the body of Ravine's main fish characters. To do this it was suggested to use blend-shapes, which I spent the afternoon working on in order to get the transformation right.

  The core idea was to create a layer that would go within the skin. This layer would fit to the body by being wrapped around using a lattice deformer. A rough shape for the fish's body was used in palce of them odel itself for the time being. The final setup would need more precise positioning as the scales would grow from around the eyes and sweep towards the tail as they spread.

   Because blendshapes were used, the 'duplicate special' tool was used. particularly useful when duplicating large grousp of objects.

  To give the progression a natural feel the scales should ideally look like they're growing independantly. This was achieved by giving variance as to when they would start and how quickly they would stop growing.

 This is the UV map of the scale. Structurally it appears simple but detailed enough. Woudl lower-polygon versions be useful? Perhaps on more distant shots

Ravine: Further Refined Soundscape

   Some further work has been made on the soundscape for the animation, with the principal elements modified being the sound of water as it interacts with the character. As well as some refining of a few other sounds and the way the music is used within the sequence.

Ravine: Environment Modelling and Mapping

   I looked into tightening up the model of the ravine. Some final tweaks merely to make things a little easier come the later stages. This included adding more crags (it was suggested that given the harshness of rock and the concept art we might be able to use an unsmoothed model), UV maps and some modification of the seabed.

  The ravine floor could be kept smoothed. The material could have been sand that was blown over the ledges by the ocean current and settled on the floor, depositing itself unevenly down below. This could also act as  rooting material for some of the plants, as sand is easier to push roots into than rock.

   The lattice deformer tool was used to give the ravine floor a rise on the ends, while an accompanying lattice deformer near the surface was used to give the sense that the split was more extreme in terms of elevation nearer the centre. This will also allow the connecting part of the floor to better sync with any ground plane extension we use

 Some of the geometry along the ledges was removed as it was giving some slightly odd shapes in the shader engine that did not feel natural. While also tweaking some of the vertex positions to smooth transitions between faces.

  Finally some workable UV maps were created. Because of the sheer scale of the model the UV maps had ot be cut up not just to fit them on the UV square but also to ease the rendering times as it will allow us to use smaller texture maps.

  Perhaps an alternative when it comes to rendering is to create two models. The seabed will be out of focus or distant so we may be able to get away with a small-ish texture map. But the ravine walls there could be two maps - high-resolution and low-resolution, each on their own separate model that we key to alternate depending on the camera distance. But this might involve more maps and more cutting up.

Ravine bed

 

left-hand wall

right-hand wall

Ledt-hand ravine floor

right-hand ravine floor