Author: Vitus Lasrado

This week’s class was about CG. We saw different channels. For now, we get to use RGB which is red green and blue but when you render in 3d software we can add more passes. This help to get better control over the highlights, mid-tones and shadows. A rule of thumb is whenever you’re following to rebuild the CG asset always use “plus” in your merge for lights and always use “multiple” in your merge for shadows.

There are different types of AOVs (Arbitrary Output Variable)

Material AOVs these used to adjust the material attributes or shader of objects in the scene like using the passes so you have control over highlights, mid-tones and shadows. 

With LightGroups, we can adjust the Individual lights from a 3d scene.

Data Passes also known as Utilities, in combination with these tools can be used to achieve various effects like defocus, motion blur, re-lighting, and many more.

IDs help to select an area

The layer Contact Sheet node lets you view different passes on one screen. 

Different renders from different software will be having different passes so when changing any highlights, mid-tones and shadows you needed to back it up by using a shuffle node.

We can add UV texture in Nuke with the UV passes, by rendering the geometry. You start with your geometry then add whatever texture you want directly to the scanline render. The other way is to add UV geometry scanline, render it and then to STMap to stamp channel and add the render src for the texture 

The position pass represents the 3d scene in colour values. Red is the x coordinate, green is y, and blue is z. You can use a PositionToPoints node to visualize it in the 3d space. A depth channel stores the distance from a point in 3d to the position of the camera (and will be moving with the camera).

PP Mask hub is a gizmo in which you can select a part of the plate and with the help of a position pass and relight an area, it is a simple way to do it and this also can be done using a normal pass.

For our homework, we had to add a car to a background of our choice and match it to the background.

This week we saw an advance projection.

We saw simple way of using roto paint, first we needed to find a reference frame, then add a frame hold node, then do your roto paint and under the paint add one more frame holder. Under that, added a premult, after which the 3D project node for the projection’s camera, added a one frame hold on the same frame, then pass the camera from it. Added the Scanline render add a camera without the frame hold, at last merger it with the plate.

Gonzalo showed us the importance of taking the right reference frame. Without the right reference frame, the projection comes out blurry and pixelated. If you see the small door we can see the differences between taking the wrong and right reference frames.

The next thing we were taught was how to project twice in the same areas.

We also learnt about different types of Projection Artefacts

Smearing: The image looks fine from one angle but the image smearing or streaks across the object.

Doubling: This is when the projection multiplies in the geometry.

Resolution: Not enough resolution in the painting because the camera gets too close.

We learnt how to change a scene from day to night, but the important thing was how to zoom in or out with the parallax (the motion of things close to you moving faster than those far away). 

As our homework assignment, we had to roto the wall and remove the tracking points in the Garage video and add some stickers or drawings on the Greece street video.

This week’s class was about 3D tracking and Projection, the project3D node projects an input image through a camera onto the 3D object. Firstly, we saw how to project texts on a video by using a card node, scene node, camera 3D node and scanline render node.

For the 3d patch – project on mm geo, first we need to find a mind point of the video and add a frame hold node then use the roto paint to do the cleanup, add a roto and use it on the cleanup areas so there won’t be any problem in QC We then have to give it a bluer on the edge to blend it in and add a premult, following which we add one more frame hold on the same frame. So now to project the clean-up in the moving camera, we needed to add a project node, card noes where the projection is taking place in the scene then a camera and scanline render, so in the project node add a frame hold on the same frame as before then attach a camera to that then to the project where you needed to attach the camera attach the 3D project to the card, card to the scene then scene to scanline render also added the camera to the scene and scanline render.

3d patch for project roto, from the plate, plug it into project3D then add a card at the area where you want to do the roto, plug the project3D at the cared img input, under that scanline render then roto after that project3D to revise it then copy past the same card from the top, one more scanline render, for the first scanline render just add a frame hold then for the project3D later gives the same frame hold for that last scanline render just add the camera without frame hold.

Project UV was fairly easy to understand than others, from that plate you needed to need to add project 3D then a card for which area you want then added the scanline render for that reader on the background. Add a square for changing the layout added your roto pant to do your clean up then change the settings to lifetime then premult it and add a reformat then copy paste the card from the top then scanline render and merge it with the plate add a camera to the project 3D and the last scanline render

Point cloud helps to visualise the video in 3D space so that adding an object or a card will be easy and you will know exactly if the object is teaching the floor or the safaris.

Poisson Mesh helps to visualise the video in 3D space and converts the point cloud to geometry.

Baked Mesh from pint cloud, to create this you need to change the select node to vertex and then select all the points from the point cloud then go to groups in the point cloud then create a group after that select the group then bake select groups to mesh this helps in geometry also. The other one is Bake select group this helps to understand where is your card placed in the 3d wold.

To work with the Model Builder geometry node, we need to add the camera and sounds to the node following which we go in the node and then create a card to match the corner with the object you want to cover and treat it as it is a corner point then play the video if you feel that it is going off the place move it to match it again.

For homework, we had to prep the room and then place the trackers and the grind.