Reflection, Refraction, and Translucency

Welcome to the blog about my project focusing on reflection, refraction, and translucency. Overall, this project was a lot of fun and very productive for me in learning more about shading and compositing within Nuke. Take a look at my daily process and see the final breakdown of the project below.

Purpose: - To experiment and understand the uses of reflection, refraction, and transmission in shading objects in a CG environment.  Day 1: Thursday, April 16th, 2020 - Today I started the project. In class we learned a lot about all the different properties that are involved in aiStandardSurface shaders. I wanted to learn more about them, myself, so i dd some testing shown below. I took one of the Stanford models provided and went through 5 different versions of each parameter, showing how each parameter affects the shader. The collage of all the tests turned out pretty interesting and I think it was well worth the time spent putting it together. Showing, from top to bottom, different levels of: Diffuse, Specularity, Transmission, Subsurface Scattering, and Coat from an aiStandardSurface. Tasks for tomorrow: - choose plate for project - camera match - gray ball match - start on rock shader

Day 2: Friday, April 17th, 2020 - The plate I chose has some pretty dynamic lighting. Paired with the yellow shader of the rock, i think the harsh sunlight could look really cool shining through my model with some subsurface scattering the light rays throughout it. I think the hardest part about this scene will be to replicate any sort of yellow bounce light that would be on the ground plane in real life. Below are the clean plate as well as the matched camera with the cube. Dimensions: 6000 x 4000 Focal Length: 55mm F-stop: f/9 Camera: Canon EOS 80D Tasks for tomorrow: - match shadow and keylight - gray ball match - animate gray ball​​​​​​​

Left: clean plate, Middle: cube plate, Right: camera and ground plane match

Day 3: Saturday, April 18th, 2020 - My progress for the day has come to a halt. For a reason that I can't explain, this specific scene file or group of objects and shaders, will not render at all in Maya. I tried Maya and Arnold renderers, but they both leave the error:

// Warning: file: C:/Program Files/Autodesk/Maya2018/scripts/others/renderOptions.mel line 19: Renderer "arnold" does not provide render options. // 

// Warning: file: C:/Program Files/Autodesk/Maya2018/scripts/others/renderOptions.mel line 19: Renderer "mayaSoftware" does not provide render options. //.

Even when I try to open Arnold render view it leaves me with a fatal error and crashes Maya. The solution I have had to use was to create a whole new project file with everything replaced, so my time spent today will be to recreate the scene with all the shaders and render tabs unfortunately.  Tasks for tomorrow:  - match shadow and keylight - gray ball match - animate stanford model

Day 4: Sunday, April 19th, 2020 - Yesterday was spent getting my scene back to normal, and having everything set up the way it should be. Today I want to at least get the keylight and shadow matched to have ready for class tomorrow. the sun through the window is the main source of light for the scene. However, the light is so bright on the table that it causes a lot of bounce light on the backside of the cube. The majority of that bounce light can be resolved with the projected plane under the object later on. I matched the position of the shadow and set up the shadowmatte plane under the ball. I also managed to organize the render layers to make it easier to render in the end: Shadow, Beauty, and Occlusion.  Tasks for tomorrow: - gray ball match - animate stanford model - start on stone shader

Left: my shadow match, Right: the reference image's shadow

Day 5: Monday, April 20th, 2020 - For today's work I managed to get the gray ball match pretty close to the reference image. I brought the sphere in and projected the clean plate onto the ground plane to give the bounce color back onto the sphere. I also gave the ball a bit of subsurface scattering on the shader for the ball to match the Styrofoam material shown in the original reference image. Shown below are the three images; the reference image, the ball's beauty layer, and the ball's shadow layer. - Also, today in class, we learned about the Fresnel mask. This mask can be used to blend two different shaders for a glassy shader. It can be done in the compositing with a subsurface scattering layer and a transmission layer. By taking the red channel from the Fresnel mask, the two layers are merged onto each other seamlessly giving a kind-of marble look to final composition. Tasks for tomorrow:  - start rock shader

Left: sphere reference image, Middle: sphere beauty layer, Right: sphere shadow layer

Top Left: subsurface example, Top Right: transmission example, Bottom Left: Fresnel mask, Bottom Right: composite

Day 6: Tuesday, April 21st, 2020 - Today I started the shader for the rock. Below is a zoomed-in view of the rock whose shader I am attempting to copy. Alongside it is my first attempt at copying that shader (created earlier in this project, thus why the shadow isn't matched). Tomorrow I want to have a better version that is more demonstrative for the project blog. Ideally this would be a compositing job using the Fresnel mask like yesterday with two different layers combined. I also think the shader will show up a lot better on an object with thinner parts like one of the Stanford models or a sphere.  Tasks for tomorrow: - further work on the shader - choose Stanford model

First attempt at the shader, alongside the rock as reference.

Day 7: Wednesday, April 22nd, 2020 - During class today it was spent going through everyone's project and troubleshooting problems. I didn't have too much to show but when showing the image above, they all seemed to like it. I got advice for the occlusion shadow coming from the original rock. Notice the strong yellow glow bouncing onto the shadow of the rock. The advice came from my classmate Nick Kubik, he said that to capture that occlusion I may need to render two separate passes. One with a high depth value on the subsurface, and one with a low depth value. This will effectively serve the same purpose that the occlusion shadow does for the main cast shadow. One of the passes will act as a richer version of the bounce color and the other will act like a feather for the color.  Tasks for tomorrow: - attempt occlusion shadow - more shader progress - shadow - composite

Day 8: Thursday, April 23rd, 2020 - Today I took a stab at making the occlusion shadow mentioned yesterday. I figured out a solution that may be a little simpler than rendering out two different versions of the occlusion glow. In Nuke I can simply change the brightness and add two of the same layer which will have one bright part and one less bright, more feathered part. Shown below is the test I did for this idea. I think it came out pretty well. It should serve the purpose intended for the composite. Tasks for tomorrow; - more shader progress - shadow - composite

Occlusion glow for the dragon.

Day 9: Friday, April 24th, 2020 - I started testing the different settings within the shader in order to make the dragon fit the scene better. I totaled 6 different renders/ tests for the two different layers of the shader. The inner beauty part of the shader is meant to capture the less transmissive part of the glassy rock. On the other hand, the outer layer is meant to be the more see-through part of the shader, with more subsurface scattering. Below is a collage of all of the tests. Notice the slight changes in all of them and the vast difference between the last two outer-beauty layers. The change comes from a higher value in subsurface. These tests were helpful for me to really visualize what specific parts of the shader needed to be changed to better fit the rock's surface. 

Left: Inner beauty tests, Right: Outer beauty tests

Day 10: Saturday, April 25th, 2020 - Since I have finalized the look of the shader, I can start finalizing the look of the shadow. I had to make the shadow a little artificially to make it look the way I wanted. The shadow that was cast naturally from the key-light through the dragon was not as bright as it should look for a transmissive object. I mainly needed to color the shadow to make sure it fit in the scene with the intensity of bounce light onto the table that it is being cast on. Another issue I had with the animation of the shadow is that it floats off the edge of the table. I can easily fix this with a roto mask in Nuke as seen below.

Left: Final shadow, Middle: Shadow without mask, Right: Shadow with mask

Day 11: Sunday, April 26th, 2020 - Today the project is due and I have the whole animation rendered out. The last thing I need to do is to composite the entire thing together. Shown below is the complete node tree in Nuke. You can see that I added a few extra nodes under some layers to adjust the color and brightness to get them to fit a little nicer into the scene. Also below is the breakdown of the final composite for this project. Of course, I will most likely add or alter this composite for the future after critique to make sure it is the best it can be. 

Day 12: Monday, April 27th, 2020 - Today was critique. The main issue that I got out of the project was that the dragon seemed to be too blown out, and I agree. Below is a side-by-side of the original dragon and what I came up with after fixing it. I simply used a color correct node under the outer-beauty layer and toned down the gain and gamma a little bit. This gave the dragon a little more contrast and helped it to feel less blown out by the key-light. Ideally, this would be fixed within the shader and the key-light intensity in Maya, but the color correct served the short-term purpose for fixing the issue.