#cg-toolkit

[sindy-cgvirtual]

Final Submission

(KNB221 CG-Toolkit)

If you were to go back to the start of this assessment to do it again, what would you do differently to complete the task?" 

Due to lack of consideration, I did not consider the sizes of the assets initially, so it has taken me much times to resize the assets and assembling. The sizes are very important for modular assets to make each other assembled seamlessly for re-use.

So if re-do this modular assets again, I will start with the sizes at the very beginning. Plan before modelling, and calculate the sizes of the main assets to make sure they are proportional to each other, which are good for scaling and flexible dimensions. 

Also, as will also make a beveled version especially for baking the texture, as in real-life everything will have bevel edges, so using high-poly version for baking texture and then use for low-poly models can improve on the appearances as well as the real-time rendering efficiency. 

Further more, I think it will be efficient if assembling from bottom to top, which enables initial snapping to the ground without moving the whole assets afterwards, which the mouse-clicking is just too messy. In my case I just assembled randomly which has caused much extra work such as keeping snapping downwards to the ground after many assets have been assembled.

Additionally, it is very important to ensure the models in 3D work-space is set to zero in position and rotation, as well as the rotation of the XYZ-axis. As for this time, one of the window is set to 0.067 degree of rotation, even though the amount is very minor, but this can still cause trouble with snapping. It also caused me extra hours to figure out the ‘minor’ problem which is not minor actually.  

[sindy-cgvirtual]

Assessment 1-Milestone 1C

( KNB221 CG Toolkit)

The first video is rendered directly from Unreal Engine 4 without compositing in After Effects.

In Unreal Engine 4, I added the level sequence for the lighting environment and imported the skeleton and animation of the female to the environment. Similar to the previous videos done for milestone 1a and 1b, I added keyframes to the camera with appropriate focal length and focal distance, and then render out the final result.

And this is the composited video done in After Effects.

Multi passes rendered from UE4 and compositing in After Effects with different blending modes to achieve the final result.

Using TrkMat for the text

To achieve the result of “putting text behind the character”, I rendered the character with custom pass of base color, and added effect of tint and level to make it alpha channel, so that I can trkmat this to the text by using Luma inverted matte.

Using TrkMat for the background

By using the same method applied to the text, I have inverted the color of alpha channel to adjust the color of the background while excluding the character.

For the background, I just added black and white effect to it and also apply the curve to adjust the brightness and contrast to achieve the final result. 

Inverted alpha channel to adjust the background while excluding the character

Through this assignment, I have become more familiar with the level sequencer function as well as the import of the skeleton and animation.

Furthermore, I have learnt how to do the lighting in UE4 by using different kind of lights and different setting of sky sphere to create different kind of atmosphere. I have also learnt how to create the god light to simulate light casting through the building by using bloom effect to the light. In order to create the fog, I have created the emitter for the first time, which is a new knowledge for me.

Additionally, unlike milestone 1a, of which the multi passes I have rendered for compositing in After Effects, this time, I have rendered different passes from the first time where I found it very useful with better effect in terms of color and shadow detail to achieve the final output. 

Lastly, the alpha channel is definitely a powerful function for mixing different layers by using different TrkMat mode. So it helps me to get better understanding of this function. As initially I would like to use mask to block the text from the character, however the character is animated while the mask stays still, which means the mask should be animated every frame according to the movement of the character, which is rather time-consuming, so I began to think the most efficient way to doing this, which is using the alpha channel. 

Time spent:

6 hours on compositing in UE4,including

add different lighting to the environment

add color to the greybox

import character into the environment

composite character and environment with level sequence and camera movement

render out multiple custom passes

render character animation without the environment

2 hours in After Effects, including:

import different passes for composition

multiple tries with different layer blending mode

add text to the video and animate the text

adjust the character base color channel into alpha channel

using alpha channel to achieve the final result in terms of text and background color adjustment

render out as video and upload to YouTube

1 hour on reflection, including:

Upload video and relevant images to Tumblr

Post reflection, progresses and documentation.