Blog
Project 1
Creating Custom HDRs, and walking through Basic CG Integration with AOVs. And Shadow projecting.
Tools: Sony A6500
Programs: Photoshop, Maya, Arnold Renderer, Nuke
Taking Ref Photos
​
Check: White balance, ISO, Shutter speed are all at appropriate settings
​
Grey Ball: Shows the intensity of the light and light direction
Cube: Used as a point of reference for the persepective of the environment
These are essential for matching the environment in CG.
Shadow plate: Shows the area what parts will look like when in shadow, can use masks to control this.
​
Clean Plate: What will be used as the background.
Required: A matte colored ball, cube
This initial scene set up is to give us valuable information on the environment so that we can match it to CG.
​
This process was a bit difficult for me during my first attempts, since I took it on a cloudy day, so my exposure, colors, and shadows changed all the time. So All base photo credit goes to Bridget Gaynor's Library.
Making Custom HDR
Center your chrome ball for your photo. Set up continuous bracketing steps for exposure on your camera.
​
I like to take at least 7 photos with 3 exposure steps apart to get a good range.
------
Combine the images
-Photoshop: Automate>MeregHDR Pro> turn off remove ghosts. Save combined image as .hdr
-Crop in the image so that it is as close as possible to the edges of the sphere.
​
Unsheperize
-Nuke: Read the created .hdr
-Add in node unspherize
-write
Needed: Chrome Ball or Fish Eye lens, Photoshop, Nuke
Capture a full range of lighting from the environment to use in Maya.
​
Using Non-Commercial Nuke limited my output resolution for my flattened HDR on my first attempts. But since nothing in my scene was too reflective it wasn't a current issue.
CG Camera Prep
Find Camera information
-Locate reference images, right click> properties)
​
-Mark down the focal length, Crop sensor size, and type of camera
​
-Use Digified.net/focal length to convert your numbers to 35 mm lens if not already converted.
​
-Create new camera in Maya, input created focal length.
​
-Go into render settings, and input the same resolution for image size, and check 'maintain width/height ratio" so you can adjust to lower sizes for test renders.
Required: CG Software (Using Maya in this example)
Matching the camera to the one used to take your background plate. ​
CG Perspective Matching
First copy ref photos into source images folder
-Set up image plane. Select the ref with a cube.
-Match resolution gate
​
-Create a cube, move the pivot to one bottom corner and snap to origin.
​
- Move rendercam until it lines up flush to the cube in the image plane ** you can use "\" to zoom into the viewport for more accurate moving**
Required: CG Software (Using Maya in this example)
Utilizing the photos we gathered earlier, we will find the perspective.
​
If you originally took your background plate/ref photos with an off kilted angle.... This can make this process much longer...
​
AOVs and Render Layers
​
Setting up AOVs allows you access specific qualities seperately when you render your image, also it helps in the process of troubleshooting your look dev process.
The ones I like to use are
Z, Difuse, spec direct, spec indirect, motion vector, and if applicable SSS and coat.
Merge AOVs (right click and select driver) Check box for Merge AOVs to keep your file neatly intergrated into one
If not merged, the files rendered will be seperate.
​
​
-Using the new render layers you can easily create layers, organize visibility of objects, and assign attributes and shaders from here.
Required: CG Software (Using Maya in this example)
Making AOVs​
​
For the love of file organization, MERGE your AOVs,
If they're not merged you've get seperate folders for ever every AOV, for every renderlayer.
Project 2
Creating Custom HDRs, and walking through Basic CG Integration with AOVs
Tools: Sony A6500
Programs: Photoshop, Maya, Arnold Renderer, Nuke
Photo Preparation
Take normal ref photos like the project before. But this time make sure to catch your object's qualities in different angles of light, in direct light, and in shadow to have a better understanding of it.
Perspective Matching
Keep your clients up to date with what’s happening. To make this content your own, just add your images, text and links, or connect to data from your collection.
Research
For our refrenced mineral, we are given a random stone with no information whatsoever on it. Some students went straight away and began to try shading it by eye. But I spent a day to research my rock, it put me a little behind at first but it gave me a better understanding on the material I was making.
Though in pictures the rock looks dull it actually had a type of rolling aniostrophic shine. After a little research Discovered essential aspects of my rock.
Tiger's Eye,
This rock is a Chatoyant rock with a SILKY luster.
That means it actually has quartz running on the inside of it, like fine hairs, partially oxidized by other minerals. Chatoyant means that it has lusterous bands running perpendicular to the directions of the silk luster.
So it wasn't too hard after that to make a game plan on how to make a shader based on that information.
​
Shader Matching
Because my material depended mostly on rows of fine hairs basically on the inside of the rock. I used a ramp shader to made alternating black and white, and repeated it until it made a fine striping pattern.
I used this as a base to feed data to my bump, anisotrophy, and spec. Changing and customizing the inputs with layers and remaps.
I made several versions and attempts to not only replicate the rock in my reference, but how it feels like in real life- since many pictures can't capture the strange qualities of a Tiger's Eye.
AOV and Render Layers
I still can never merge my aovs correctly. And I made the same mistake as last time during the first projects- not merging my AOVs. So While doing my final comp I was overwhelmed with files, fumbling with file organization because I had to pull my AOVs from many ingrown folders from the renderfarm...instead of one neat, clean, file.
It seemed to go smoothly up until this point.
Project 3
Using provious practices and intergrating using tracking.
Tools: Sony A6500, Mixamo
Programs: Maya, Arnold Renderer, Nuke
Footage
Make sure to capture footage that has many points like grids, edges of signs, to track.
-Least motion blur the better.
-If using 3d tracking, parallaxing points makes for a better track later.
Tracking
I ended up doing 2d tracking, because all my other footage had an ungodly amount of motion blur, not enough points to track, or unstable points that couldn't be used.
Nuke: Add a 2d tracking node
-Mark at least 4 reliable points to track, and run the sim.
-within the 2d Tracking node, create stabilize, and 2d transform. These will help negate any extra shakes in your image, and the 2d Transform will move your rendered assets according to the movements of the footage.
Lighting
I used an LDR for here since it was a very overcast day, and there were no strong highlights.
Compositing
Merged AOVs made this much easier. With a cloudy environment I didn't need to edit the AOVs too much.
-Zdefocus
-Edge Blur
-Lightwrap
Adjusting those helped integrate my shot much better.
Though I greatly regret not organizing
Color Corrections
After you after you rendered your CG assets, and tracked it to the background plate, color correct both and adjust AOVs accordingly to try to match it to the shot.
Next Time I'd definitely do a sky replacement