more break-downs on my work-in-progress

Hi Guys, my thesis is mostly done at this point. Sry for noting posting for a long time, cuz I get really busy recently in finishing up the work and talk with my thesis committee. The final result can be found in my website or the vimeo link here: https://vimeo.com/112202490
Here are some more break downs from my render passes for my foreground houses which I did earlier. I will post more of my work-in-progress and talk about how I approve the problem later on : )
Thanks!

final comp result from Nuke

beauty pass one with a high freq noise

beauty pass two with a low freq noise

a bottom mask pass create with a procedural shader 

a edges mask pass create with a procedural shader 

normal pass used for edge detection 

reflection pass with high freq bump

relfection pass with a smooth water surface

Boat Shading & Rendering Test

did some shading and render test last week, using the custom procedural ink shader I developed together with different layer for compositing.

NPR test on the boat. (top image: NPT render, bottom image: 3d model)

group of boats with reflection

building test and final layout

So in order to make the scene more interesting, I decided to go beyond just the Lotus flower, and include more in my subject of matters. I'm using Yang Mingyi's painting as my main references because I love the way he render light and shadow in his painting:

Yang's Paintings

I did some test shading and lighting on a simply building model before. There are still a lot to add right now but it seems to be a pretty good start : )

And here is the layout thumbnail I painted together with a rough 3D layout scene.

painted thumbnail

3D layout

reflection test

reflection test:

my references for reflection in Chinese ink-and-brush painting:

my render test for reflection:

different reflection passes:

WIP shader NPR development on Petal and Stem

WIP NPR shader development on Petal and Stem

I have been using RSL(renderman shading language) to develop my NPR shader and render it from Renderman for Maya.
I used one of Chang Dai-chien's Lotus painting as my main reference for composition and surface properties. And I started with developing the shader for the Petals.

1. First Draft Petal Shader:

Here is a screenshot of my first draft Petal shader:

In my first draft version, I tried to get a basic contour rendered and a basic interior shading for the surface to match the ink color. I wrote a function to detect the contour based on the surface normal N, the eye ray I, and a user input threshold th (for controlling the width of the contour).

For the interior shading, the color of the surface is determined by the diffuse factor of particular surface point (the angle between the surface normal and the incoming light ray hitting that surface point). And according to the diffuse factor (an float number from 0.0 to 1.0), an 1D color ramp is applied to the surface to determined the color. These colors are picked based on the different tones of the ink and also the changes of color when ink spreading on paper.

The render result with this draft petal shader is shown below:

2. Adding Painterly Details and Irregularity:

Apparently, this looks way too perfectly smooth and really CG. 

In next step, I'm trying to adding more 'painterly details' to mimic the effect of ink spreading on paper.

Here is a screenshot of the revised shader:

Here I  added a simply noise function to the diffuse factor to break the transition between different ink tones. The noise (from 0 to 1), is applied with a scale factor: (noise-0.5)/scaleFactor for controlling how much the ink spread. And I assemble different shading components back to a MaterialEnsemble node builded in Renderman Silm. The benefit of doing this is to take advantage of the build-in function renderman already has and get a better implement with maya (for instance, using light link with maya lights) 

Also, I used the similar method to create the stem shader (with irregular edges and same interior shading)

 

A render test in maya with the revised shaders:

A close-up compare of the improvements:

before

after

3. Thoughts on the Artistic Controls of Shading:

When using physical brush and ink, traditional painters control where and in what direction to have the brush interact with paper. Similar to this, CG artists also need to have artistic control on where to put the color and even how it spread over the surface.

Since the interior shading is control by diffuse factor, artist can control the look of the surface by using lights with different intensities in different directions.

Here are some screenshots with maya file and the render image side by side.

As shown above, the color of the petal is changed by manipulating the directional lights in maya. In this way, artist can control the shading by moving lights.

Master Thesis Introduction

Siran's Master Thesis Topic:

Bringing Cinematic Lighting Aesthetic into Traditional Chinese Ink-Brush Painting ‘Lotus Flowers’ in 3D Computer Graphics

In my research, I intend to merge traditional Chinese ‘Lotus Flowers’ Painting with cinematic lighting aesthetic in 3D computer graphics. The goal is to achieve a unique artistic style, which both preserves the tradition and essence of Chinese Painting, and brings more changes of light and color for mood enhancing and visual story telling. A workflow of shader-based Non-Photorealistic Rendering method combined with compositing technique will be developed to achieve the final look.

My Motivation and Introduction to My Topic:

Chinese painters have been using the ink-and-brush medium to create beautiful paintings on paper or silk scrolls for thousands of years. Traditional Chinese painting has distinctive characteristics. It does not emphasize one-point perspective nor does it aim to express the shades of color of a subject in relation to a fixed source of light as traditional western painting does. Thus, it usually does not have any physically accurate lighting, shading and shadow.

Fig 1. Lotus Painting by Chang Dai-chien(May 10, 1899 – April 2, 1983)

However, I have always wondered about and wanted to explore how traditional Chinese Ink-and-brush painting would look with the introduction of more change in lighting, and also how it would appear in digital media forms and displays. The inspiration behind this idea is closely related with my interests in cinematic lighting aesthetic. In the world of filmmaking and photography, cinematographers and lighting artists use light as a powerful tool for conveying mood and visual story telling. This is achieved by applying combinations of different lighting aspects such as quality, direction, color. Although they are very different considering the methods, tools and media being used, traditional Chinese ink-and-brush painting essence and cinematic lighting aesthetic share an important similarity – they both emphasize expressing the characteristics of the subject and the artist’s vision and emotion rather than just depicting a scene in a physically accurate way. This makes me interested in attempting to merge these two types of art together by introducing cinematic lighting aesthetic into Chinese ink-brush painting.

(a)

(b)

Figure 2: Examples of cinematic lighting principles applied in movies for enhancing visual story telling: (a)screenshot from film ‘Blade Runner’(1982), in which, the director of photography (DP) used strong rim light and an extreme high lighting ratio to enhance mood. (b)screenshot from the film ‘House of Flying Daggers’(2004) where the DP used atmosphere effect to create drama and to separate out the main characters.

Instead of working with traditional media – ink, brush and paper, my research will be using tools and methods in 3D computer graphics. In recent years, lots of works have been done in computer graphics to recreate painterly rendering styles (including traditional Chinese ink-brush painting). And different Non-Photorealistic Rendering (NPR) workflows have been applied in 3D computer graphic animation productions to create beautiful works like the short-films Paperman (2012) and Twinings ‘Sea’ (2011). Both works have a graphic quality of a 2D hand-painting style and interesting changes of light and shadow for enhancing visual story telling. This motivated me to try bringing the Chinese ink-and-brush painting style into the 3D computer graphic world， and utilize the techniques in computer graphics for applying cinematic lighting principles into ink-brush painting

(a)

(b)

Figure 3: Examples of 3D animation with painterly rendering style. (a) screenshot from short film ‘Paperman’(2012), in which NPR technique was developed to create a traditional 2D hand-drawn rendering style, while lighting and shading qualities of 3D animation are still maintained. (b) screenshot from the commercial ‘Twinings Sea’ (2011), where special shading and compositing workflow were used to create the oil painting look.

The goal of my research is to create a new type of Chinese ink-and-brush painting in CG, which preserves the tradition and essence of the original art form, while introducing changes of light for visual story telling. I chose Lotus Flower – a very popular subject in traditional Chinese ink-and-brush painting, as my subject matter in the work. And I will be mainly using Chang Daichien’s Lotus painting as my references (Fig. 1). The research will focus on 1) How to recreate an ink-and-brush painting style for different parts of Lotus (leaves, stems, pods, petals) utilizing NPR techniques. 2) How to bring in and handle the changes of lights and different render effects in Chinese ink-and-brush painting, which is commonly seen and comes free in physical cinematic lighting. This includes: the subsurface scattering effects which happen on organic surfaces, the reflection and highlights on water and plants, refraction, atmospheric effects, etc. 3) How to use cinematic lighting aesthetic in Chinese ink-and-brush painting to enhance mood and story-telling. The research will be focusing on the rendering aspect of 3D animation. Such aspects as animation and camera movement will not be explored in this project as they exceed the scope of the study. Also, the goal of this research will be recreating Chinese ink-and-brush painting through 3D computer graphics, so the tools and workflow developed in this research are not intended to be applied to any physical painting process.