Category Archives: Term3

Showreel_term 3


In this term, whether it is a cooperative project or a personal project, I have a deeper understanding of this major. It made me more familiar with the film and television production process.

For personal projects, my main purpose is to consolidate and strengthen the content of the last semester. Make me to be more familiar with the post-production process. From modeling to texturing to compositing. This pratices allowed me to solve many technical problems. Examples include how to create hard surface models in Zbrush, layered rendering in Maya, and texturing. I think these things can make me more confident in my future creations.

For me, the main purpose of the collaborative project is the research on special effects such as smoke particles, as well as the knowledge of what I have learned before. For example, the clean-up, camera tracking and shooting. This is the first time most of us have the shooting experience, so many problems have arisen. For example, we did not carefully consider the final effect of special effects when shooting. Therefore, it is difficult for us to match character actions with the animations of the portal effects. In addition, the lack of preparation of props made it difficult for us to Roto out later. If we use a green screen when shooting, we may be able to create better effects and save a lot of time. But I am very happy to have those kind of experiences. It can help me avoid many problems in my future creations. I am also glad to be able to complete this project with my team members. Diana is a very good leader. In this exercise, she helped me solve many problems.

In general, I have gained a lot this semester. Regardless of the results, I am very happy to have these experiences.

Rendering and test for SHOT_5

This shot is quite tricky for me. This shot mainly shows that Diana draws a stroke on the wall, and then the portal slowly spreads out to form a circle. like the image shows.

I tried several different effects in After Effects. But the result is not very good. The portal looks like a flat surface. It is not very suitable for the scene.

Here is the first version for test.

Obviously, this result is not ideal. And the style of this portal is different from other portals. After a group discussion, we decided to use the effect of smoke instead of light. So I created the portal in the same way as in the previous shot. However, adjustments have been made to the animation. Make the portal appear as a line at the beginning, and then slowly expand into a circular portal.

Here is the final version,

Rendering and combination

I rendering the portal separately into two parts, the main object and shadow.


Shadow layer

I encountered several problems in rendering. The first is the voxel size of the portal. I found that when I adjust the voxel size to a smaller value, it will bring better and clearer results, but at the same time, the simulation cache and rendering time will be longer.

The second problem is that when I perform multi-pass rendering, there will be a problem as shown in the figure. In the rendering process, there is no white edge in this circle, but it will suddenly appear when the rendering is completed. It still exists when I composite it in Nuke. In the end, I found that the problem is the alpha channel.

This kind of problem will not occur after canceling the option of straight alpha in the rendering settings.


Here is the process of combination.

I clean up the material first, remove the signs and pipes on the wall. Then extract the projection through the shuffle node and merge it with the portal and the original material.

Then add another scene video to the screen. This expresses the meaning of Gherardo being transferred from another scene.

Then add the portal and shadow.

The last step is to Roto out Gherardo and use another roto node as a mask to make Gherardo through out of the portal. In fact, this is more difficult. Because this camera is in the front. Many problems cannot be concealed by covering up. And due to our early considerations, Gherardo’s actions are not very match with the portal animation. Through discussions with adlo, his suggestion to me is to adjust Roto to make the shot looks real.

Portal creation and test

I started the first portal test after we finished shooting.

According to our concept, I did a test of two versions of the portal. After group discussion. Everyone will prefer the second version. So next I will develop the production plan of the portal in the second version.

For the second version, I decided to use the Turbulence plug-in in C4D to help me complete the production of the portal. In the transmitter, a divergent smoke ring can be obtained by adjusting the turbulence, temperature and density. In order to make it closer to the effect of the portal, rather than just a single-layer ring, I added another ring behind the first ring to make it attractive to the smoke of the first ring.

According to the tested version, we found that the portal did not interact with the surrounding environment and the arch. The portal looks like it is pasted on the screen. It is not very harmonious with the scene. Therefore, ambient light and projection are very important to us. So I imported our footage in the taste. And an arch was created based on the shape. This allows a good interaction between the portal and the arch.

Fix the problem

This week I started the composition of the second shot and fixed the problem that appeared in the previous shot. Besides the model material has been improved.

last week Nick suggested a few thins that need to be changed in the shot. When the robot rotates its head, there are several parts that stretch and deform. So this week I adjusted it.

I used the Blend Shape and sculpture tools to fix these problem.

here is the result after I changed.

Here is the final outcome. I also added some scratches to the texture.

Another problem is that due to the insufficient number of faces of the model, the surface of the model is not smooth enough. So I increased the subdivisions of the model to make it better.

In the second shot, I rendered different channels and adjusted them in Maya.

This week I made adjustments to the shadow and skin problems that Nick proposed. For my previous version. Nick thinks the character’s skin looks too dark. Compared with the surrounding sofa, the dark part is a little deeper. There is also the wrong direction of the shadow.

Here is what the shadow looks like before the change, which is completely wrong compared to the shadow of the sofa on the right. The correct shadow should be straight down.


So I re-adjusted the light source in Maya. Take the light in the right corner as the main light source and add another light under his feet to enhance the shadow effect. Afterwards, the color and blur of the shadows were adjusted in Nuke.

This is the effect of the shadow after the change.


About the modification of skin color. I chose multi-pass rendering when exporting from Maya. Afterwards, there are more possibilities for model adjustments in Nuke. For diffuse and specular layers, I added a Grade node to adjust the contrast and color saturation of the model.

Here is the result after adjustment.

Texturing and Creation of HDRI

Last week, I just finished the model of the robot. It is actually not perfect. But I may take time to continue to improve this model in the future. The main task of this week is texturing the model, shooting footage, and images making HDRI images by myself.


For the texturing. I was using Substance painter to finish the work. Firstly is because I have seen some tutorials about Substance painter. Very cool. So I want to learn and understand this software a little bit. The second is because, in addition to the material, I also need some damage effects in the field. Like the image shows. Substance painter can help me accomplish this.

The workflow for the texturing. In the first step is to export the model from Zbrush to Maya as two level versions. This step can help me reduce the burden on my computer. Maintain the effect of high mode in the case of low mode. After we import the model in Substance painter, we need to bake it first. Then I can get a low-level model with a high-level model surface.

Then i can start to paint on it.


For the HDRI. I borrowed a 360 camera from the equipment department to help me finish it. This is a very simple method, compared to SLR cameras and metal balls. There is no need to stitch and cut out pictures. Only a camera and app can help us complete the production.

Here is the image results and the rendering results in Maya.

First version for test

Shooting and Division of work

After last week’s field survey. We voted to select several filming locations. On-site shooting started this week.

Here are the photos Diana and Gherardo brought back last week.

After that, we quickly decided that the shooting time was Wednesday and Friday.

Here is a work photo

Diana and Gherardo reorganized the footage and upload it. In the meeting on Friday, everyone received the shots they wanted to make. Here is the time planning table.

Here is the shot information

Detalis of each location.

Modeling and test

For personal project. I decided to focus on 3D Composition. The topic is related to the epidemic.


There was a melodious piano sound in a deserted apartment building. The camera shifted to the inside of the apartment. There is a robot playing the piano.

Work Process

modeling, texture and render – Maya, Zbrush

composition, tracking -Nuke

Shooting – Black magic or Canon 5D 3


In the first week I started using Zbrush to create my initial model.This is just an initial model, and it is not perfect in many places.

In the next few weeks I have been trying to create hard surface models in Zbrush. Nick told me that it might be easier to create a hard surface model in Maya. I tried to combine the two software to help me complete the model. But I found that it is quite difficult to use these two software together. So I started learning to create hard surfaces in Zbrush.

The first method is to use cutting and tilting brushes to help us create a smooth and regular surface. These two are my more commonly used cutting tools. Trim brushes are often used to crop the outer edges of objects. The Clip brush is used to crop the inside of an object or a flatter surface.

These two are the brushes that help me trim the surface. Make the surface of the object more regular.


When encountering this kind of surface, I will use the Zsphere tool to help me with topology.

The last tool for creating hard surfaces is to use Zmodeler. This tool is powerful. It helps me create objects that are as regular as the polygons in Maya.

Here is a video introducing ZModeler. It is very useful for me.

Another is the ZRemersher tool to help me rebuild the surface of the object. In order to paste the material and color later. The Subtool tool helps me to import each part into Maya hierarchically for animation and rendering.

Here are some reference pictures I looked for when modeling.

I also looked for some books related to machinery or robots. The first one is “How to draw”. Another is “Industrial Divinities”.

Collaborative project Introduction and concept

In the first week, Diana collected everyone‘s ideas in our group. In the end, two ideas were adopted. The first one comes from Diana. The main theme is that several characters use portals to travel to various places in London. We want to use 3d characters such as mice and bees to guide the protagonist in different directions.

This is the inspiration.

The second idea comes from Gherardo. It is mainly the multi-layer relationship of an object in the picture.

This video may be able to express the meaning better.