Final Video

My final output, finally!

Tweak twig

Low res test render of final composite with camera shake and motion blur.

Spent the whole of today tweaking my comp to give the scene a greater sense of depth via color correction and saturation, using the depth pass as a mask. The render of the kelp and jellyfish were also duplicated, transformed and color corrected to create more layers in the scene. This way, together with adjusting the values in their respective depth passes, I could have the kelp fade into the distance. A color lookup was also used on the merged depth passes to create a sort of exponential falloff that gives the feeling of a medium like water instead of air.

Also, lots of tweaking to get the comp to look right.

Rerendering again..

And so I'm re-rendering all my elements, again. The reason for this was that I had to change my animation and simulation attributes so the tentacles would not look really stiff near the head. That was caused when I unknowingly edited the gluetoanimation attribute while trying to solve another problem.
And since I had to re-render anyway, I took the opportunity to tweak my animation for the hero jellyfish.

The background jellyfish were also re-simmed, however, only 5 were simulated this time, instead of 20 in order to save ROPing time. The 5 were then duplicated, rotated and translated to populate the scene. The kelp were also scaled down as I found that they filled up too much of the screen and were taking the focus away from the jellyfish. Another problem I found was with generating a zdepth pass from semi-opaque objects where the information was inaccurate. To solve this, another mantra node was created to render the objects with matte shading and the micropolygon engine to speed things up. The zdepth passes were rendered out separately for each element to be merged in Nuke. This was done so that I could duplicate the renders of the kelp and background jellyfish to create the illusion of even more jellyfish further away without the extra render time. By keeping the zdepth passes of each element separate, I can then duplicate the required one and increase the value so the ZBlur would work on the duplicated element as well.

I also ended up using the light emission pass as the main pass for one of the hero jellyfish as there was some nice transparency going on there which was lost in the beauty pass. The other pass were then graded and added on top of it. Render times from Nuke also seem to be going up, at about 4 hours for 400 frames now.

Nuke Network

Comp Render Test

Here's a quick render. It's really compressed, so ignore the sparklies, they won't be in the full res render!

Almost there..

Test composite, render going solid for a few days now

Also, tested out the motion blur 3d node in nuke to get some camera motion blur going in comp.. and camera shake in Nuke!

Some methods I’ve tried and determined to be useful were: 1) F_MotionBlur, which analyses the image sequence and generates motion vectors which are used to blur the image sequence accordingly, 2) MotionBlur 3D, which accepts an input camera from Houdini and uses the information to generate motion vectors, which are then used by the VectorBlur node to blur the image sequence during a camera move. Lastly, there is the CameraShake node which is insanely useful for, you guessed it, camera shake. The frequency jitters the image left, right, up and down to simulate camera shake. As a bonus feature, motion blur is also included!

Everyday I'm Rendering...

Did render optimisation for the whole of today, adjusting the pixel and ray samples to have a decent yet fast render.

Also tweaked the width attribute used in the wire sim so the oral arms would not fly off to the side. An interpenetration problem with the tentacles and head geometry was also solved by including more points in the gluetoanimation attribute. This way, I could get away with not simming the head geometry as an RBD collision object.

Comp Update

Here's a video rendered overnight from yesterday. The swimming pattern comes from the refraction, which, while looking really good in a still frame, looks plain weird when moving. Tried smoothing the refraction out so it wouldn't be so noticeable, yet retain the complex patterns caused by the refraction. Hopefully I'll have a pleasant surprise tomorrow!

Updates

Updated render with some refraction going. I actually plugged the output of the pattern into the attenuation colour  of the refraction, giving the head a nice folded look. Hopefully it'll hold up when it's moving.

Also working on a particle sim for floating debris now!

Yet another slapcomp!

Here's the flipbook of the background jellyfish ropped out yesterday night

And another slapcomp! Experimented a little with the Godrays node (no I didn't make that up) and the iDistort node with noise to lend an underwater feel to the shot. In order to get the iDistort to work, the data must be shuffled into the uv channels of the main image. I also used a constant, a ramp, a radial and some noise nodes to create a background for the shot.

Render update

Another lighting/render test!

Also restructured my network and guess what? It now runs at least 10 times faster! Previously, what my network was doing was building geometry every frame based on the simmed wires. I restructured the network to construct the tentacles and oral arms and using wire capture and deformers to transfer the wire sim data to the geometry. A problem I faced here was capturing all the tentacle geometry as they might cross over into another tentacle's capture region and mess up. to work around this, I gave each tentacle a group using the connectivity and partition combo and did a foreach based on that.

Another problem I faced was having the wire sim in my asset and running in a foreach. This caused problems with my wire sim because it was iterating through my asset at each frame and overwriting the values of the jellyfish in the list before it. So the sim data from jellyfish 2 is overwriting the data from jellyfish 1, and it messed up the simulation, big time. So I had to do a workaround where I ran only the head in the foreach, and used the output to drive the tentacle and oral arm part of the network. And it seems to be working pretty well, since I'm nearing the end of my project and would spend the time more wisely by working on lighting than reworking the way the tentacles are simmed.

Progress update

Progress is really slow at this stage, and I can't really get a nice rim on my jellyfish. And I really don't want to have to light each of my background jellyfish individually. That would be just crazy.

Also fixed my wire sim today after having it hang Houdini. I knew something was wrong when it took ages to rop, but I didn't expect the wires to just explode. Turns out I turned cache simulation off while trying to fix a problem some time ago and the wire solver doesn't like that.

Latest render test

Latest render! 

Rendering.. Argh

Encountered a whole ton of rendering problems, first it was the noise caused by the stochastic transparency setting that affected my SSS (which took me a whole afternoon of troubleshooting to figure out). And now it's the look development that's giving me a headache

Tried out a whole bunch of colour combinations, which obviously, look weird.

PBR!

Created a repeating pattern around the head of the jellyfish the same way I created the frills, only in SHOPs this time. Rendered with PBR.

DOPs challenge in Houdini!

Jellyfishhhh test comp

Here's a test composite with a gradient background

Spent the most of today tweaking the shaders... And finally copied my kelp onto the proxy curves.. And a test composite. 

Jellyfish Shader!

Further refined my jellyfish shader and add extra functionality to my Mantra Surface by creating an opacity ramp going from the top of the head to the bottom in the mantra surface so the top would be more opaque.

Also adjusted the noise for the oral arms so it'll be more frilly.

The problem with the detached arms were also fixed by soft transforming the inner surface of the head down to intersect with the arms. Hopefully, it'll be hidden with the motion blur on.

Mr T.K. also dropped by and gave some tips on making my jellyfish look Pixar-ish. So I copied and peaked the head of the jellyfish and wrote a shader with an opacity ramp the goes from a colour to black, giving an extra layer of transluceny.

Jellyfish update!

Reference

Render test

!

Test for the random motion paths

Did the jellyfish's shading for the most of today and finally managed to get closer to the translucent look I wanted. Charles also came by today for the weekly mentor meeting gave me some tips on optimising the background jellyfish.

Kelp References, Modeling and Shading

Render test of kelp created by copystamping scale and rotation values on to a leaf model

Model Update

Created the oral arms of the jellyfish using noise and twist after hours of trial and error.

Sine to create the wavy patterns, aanoise to make the curves non-uniform. Another layer of aanoise is used to displace the points three dimensionally. Without this, the arms look pretty flat. Also, since this VOP SOP is after the wire sim, the point position changes every frame and the noise would swim. To solve this, I did a time shift, locking the oral arm on the first frame and referencing it into the VOP SOP. This way, the noise would "stick" to the geometry as it moves.

Mr Daniel and Mr Albert dropped by the office to see us today!

Modeling Update for the Jellyfish

Managed to model the oral arms of the jellyfish, not that great, but progress nonetheless. Also passed them into my wire sim, apparently if a surface is passed in, the wire solver will treat each edge as a wire. Due to this, my simulation is slowed down by a LOT. I'll have to find a way to use the back bone to run through the simulation before using a wire deform to transfer the data to the high resolution geometry.

Also found a way to have the jellyfish move along a curve (using the carve) at a preset speed, no matter the length of the curve. I needed this because I have an expression which animates the jellyfish forward  to the animation of the head. Hence, since the carve node works on a 0-1 range, I wouldn't want to go in and adjust it for the many background jellyfish I have. Hence, I used an arclen() expression to translate my translation value into a 0-1 range that can be used by the carve node. fit01() does not work.