1.1 Problem Definitio
Through the century, film technology has grown up and makes miracle in our life, from the first projected black and white movie in 1895 to the 3D movie today. In the film industry, 3D movie has been very popular in recent years and it is almost becomes a need in the market. As 3D film technology is more and more matured, 2D to 3D conversion can be applied to live-action or VFX productions to add 3D depth to standard 2D material. Hence, filmmakers have different options of 3D film making to suite to their own styles and of course it will rely on the market needs as well. This project will look at which production method is better for filmmaker in term of quality, time consumption and cost.
1.2 Scope
The extent of this project is to present past and current knowledge of film technology and the development of 3D stereoscopic production. Any business and marketing strategies in film industry will fall outside the scope of the project. Where the current and new 3D film technology is discussed, this report will also outline a critical analysis of each method of live-action shooting in 3D and the process of 2D to 3D conversion.
1.3 Rationale
Stereoscopic filmmaking and stereo 3D conversion is an area of film technology that is constantly changing and expanding. Appearing in many forms and guides, many possibilities of production can be found, from the early development stages right through to post production. As more films are produced in 3D, filmmakers will experience a clearer perspective on the impact in time and cost of shooting in 3D relative to 2D. Hardly say that, 3D movie will take longer time to shoot and post produce than a comparable 2D films nevertheless it will cost more in overall. Hence, stereo 3D conversion will be considered due to being cost effective. Ultimately, the decision to shoot in 3D instead of 2D will depend on the investment return from the marketplace, and creative is one of the considerations too. However, affordably priced 3D camera rigs and other equipments are also being developed that will make stereoscopic 3D production accessible to independent producers on limited budgets. This project provides important information in 3D film making and post produced. Multimedia technology students with experience in 3D modeling, video production and post production may also benefit from this report.
1.4 Aim
The overall aim of the project is to compare and contrast in post production on stereo 3D conversion and live-action stereoscopic images and to define the reason for selecting each method.
1.5 Objectives
To develop an in depth understanding of 3D display, stereoscopic images and depth perception. Critically analyse research and explore the process of 2D to 3D conversion and live-action shooting. To evaluate the advantages and disadvantages between stereo 3D conversion and live-action shooting. To define which technologies on the way to best achieve a stereo conversion.1.6 Brief History of 3D
Here is the highlight of the history of 3D.
In 1838, Sir Charles Wheatstone first explained “stereopsis” – the perception of depth. In 1851, Queen Victoria saw a “stereoscope” still image, 3D became all the rage. Around 1890, first stereo film camera introduced. Circa 1915, first red & blue “anaglyph” movie was shown to broad audiences. In 1953, 45 films were released in 3D includes Kiss Me Kate, House of Wax, Creature from the Black Lagoon and more. In 1970’s, new capture and projection technologies were born but unlikely success due to eyestrain for the audience. Jaws 3D is one of the movie which projected in the cinema. Late 1990’s, IMAX begins projecting in 3D In 2005, Hollywood studios’ Digital Cinema Initiative, aided by Entertainment Technology Centre’s Digital Cinema Lab, a specification for a standard digital cinema package was created. In 2009, more than two out of ten movie screens are equipped for digital 3D In 2011, 3D home experiences no longer a myth. 2.0REVIEW OF EXISTING KNOWLEDGE2.1 Introduction
The review on this report provides a rich source of knowledge about depth perception, stereoscopic images, and 3D display that can be applied to the stereoscopic image production which will be discussed on chapter four. To understand how better to present information on 3D stereoscopic, a comprehensive understanding of depth perception is necessary. In addition, knowledge of the human visual system uses as a guide on stereoscopic viewing will be covered in this section.
2.2 Stereo Vision in the Human Visual System
The visual system consist retina, vitreous, iris, lens, pupil, cornea, optic disk and optic nerve. Each component of the human eye has a role that requires to work together with the brain, and the perceptive needs of the human mind. The human visual system has a physical configuration that able to build a 3D model of the world from two separate flat images taken in by the eyes. In the real world each eye sees a slightly different view of the world due to human beings have horizontally separated eyes.
2.2.1 Depth Perception
Depth can be divided into binocular depth cues that image received by both eyes, and monocular depth cues that image received by just single eye. Studies with random-dot stereogram have shown that the binocular and monocular depth cues are independently perceived (Julesz, 1971).
2.2.1.1 Monocular Depth CuesAt longer distances, pictorial cues such as textures, size, colours, shadows and perspective lines are more important than the binocular cues. Failure to present credible pictorial cues might cause undesirable effects and thus destroy the sensation of depth (Boev & Hollosi, 2008).
Texture in sufficiently symmetrical or constant patterns is a good cue to depth. Shadows work as a depth cue by indicating the size, shape, orientation and parallax of different objects. The functioning of scale as a cue to depth relies on the fact that the sizes of familiar objects are known. Focal depth has been proposed to be an extensive cue to depth that is effective on different distances (Horii, 1992).
A person with one blind eye has limited ability to judge distance and depth. However, they able to extract 3D information from a single 2D view through the basis depth assessment as below:
Scale – through the evaluation of the relative size of objects. Superposition – from observing one object partially obscuring another. Shadow – through examining the shape of an object’s shadow. Focus – through changing focus on the eye. Perspective – through observing that parallel features converge as the distance increases.The ability to judge distance can be developed with experiences; however, it can never reach the same degree that can be obtained with binocular vision (Westone Resource, 2002).
2.2.1.2 Binocular Depth CuesFigure 1 shows how geometry of binocular vision gives rise to slightly different images in two eyes. If the two eyes are fixating on a point P, then the images cast by P fall at the centre of the fovea in each eye. Now consider a second point Q. If the images of Q fell five degrees away from the fovea in both eyes we should say that Q stimulated corresponding points in the two eyes, and that Q had zero disparity. If instead the image was located 6 degrees away from the fovea in one eye but 5 degrees away in the other, we should say that Q stimulated disparate or no corresponding points and that Q produced a disparity of 1 degree. In general, if Q’s image falls x degrees from the fovea in the left eye and y degrees from the fovea in the right eye than the binocular disparity is (x – y), measured in degrees of visual angle. The amount of disparity depends on the physical depth (d) of Q relative to the fixation point P. In fact, disparity is approximately proportional to this depth difference divided by the square of the viewing distance (v). Thus disparity increases with the amount of depth, but decreases rapidly with increasing viewing distance (Chalmers & Lo, 2003)
Figure 1. Geometry of binocular vision
Schreer (2005) stated that human eyes are separated from each other on average by approximately 63mm.
Beyond a distance of 400 metres the change in parallactic angle becomes so small that depth perception cannot be discerned. The smallest difference in parallactic angle which can be interpreted as an impression of depth is an indication of our stereoscopic acuity (Westone Resource, 2002).
2.3 Stereoscopic Depth Cues
Stereoscopic depth cues are a kind of motion parallax cues. Human is using two eyes as two points of view, and then make comparisons between these two views. Retinal disparities can be defined as discrepancies between two images. Two pictures to be processed together with specialized neurons in the visual cortex looking for these disparities, hence more information, more accurately can be extracted from motion parallaxes.
Mostly horizontal parallaxes, occlusions revelations, some shape changes, and convergence cues. One could rightfully argue that there is no stereoscopic depth cue that is not a motion cue (Mendiburu, 2009).
Horizontal Parallax
Human brain will extracts and computes the size of the disparities to assess the distance of the objects when looking at a stereoscopic image.
Occlusion Revelations
Occlusion occurs when objects overlap each other. Occlusions are the most powerful depth cues. In monoscope, some parts of the background object are hidden. In stereoscopy, there is a thin stripe of the background object that is seen by only one eye. This additional texture is a major cue for the brain to reconstruct a scene, to the point that occlusion will supersede any other cue, and your brain will twist reality in every direction to make it work (Mendiburu, 2009).
Shape Change
In stereoscopy, there’s an additional factor. The distance between human eyes is fixed at an average of 2.5 inches; therefore, the amount of “side view” for per eye is the function of an object’s size and distance. A dice you hold in your hand will reveal more of its sides than one on the other end of a craps table. And a building may not let you see its sides because it’s more that 2.5 inches wide. If you see more or less of an object, and you can locate it in the distance or have a reference shape to compare it to, then you’ll have enough information to infer the missing information and assess its actual size. Most of the time, this will tell you the size and distance. When shooting in 3D, filmmaker will play with the camera’s interocular distance. This will create size effects on objects, landscapes, and actors, and make them feel giant or small (Mendiburu, 2009).
Convergence, Parallax, and Depth
Parallax is the relative position of an object’s image in a set of pictures. When the parallax value of the pair of images of an object is negative, the left-eye image of the object is seen on the screen at a position that lies to the right of its right-eye image. When a viewer converge the image pair, the 3D object seems to be located in front of the screen plane. When the parallax value of the image pair is positive, the left-eye image of the object is seen on the screen at a position that lies to the left of its right-eye image. When a viewer converge the image pair, the 3D object seems to be located behind the screen plane. When the parallax value is zero, the two images of the object overlap at the screen plane. In this circumstance the object seems to the viewer to be located on the screen plane. When a viewer’s eyes are focused on the screen and converged on an object that appears to be in front of the screen (i.e. an object with negative parallax), the viewer’s left and right eyes cross. And when a viewer’s eyes are focused on the plane of the screen and converged on an object that appears to be on or beyond the screen (i.e. an object with zero or positive parallax), the viewer’s eyes remain uncrossed. Because crossing the eyes can cause physical discomfort, filmmakers are often cautious about using excessive negative parallax, choosing instead to limit the frequency of use of the effect and the degree to which objects are allowed to intrude into the viewer’s space (Clark, 2010).
See figure 2 for a clearer picture.
Figure 2. Parallax (Autodesk, 2008)2.4 3D Display
Stereoscopic display efficiently presents a left eye image to the left eye that is isolated from a right eye image that is presented to the right eye. This allows the visual system to merge the two images resulting in the perception of depth, or stereopsis.
In current market, there are several methods to create stereoscopic visualization and the most common approaches used today are summarized on next page. Each one has advantage and disadvantage that should be considered by the user.
StereoMirror
Features:the light path from two polarized LCD monitors is combined using a 50/50 beamsplitter that transmits light from one display and reflects light from the other; find additional information here.
Glasses: passive, linearly polarized
Pros: high brightness, full stereo resolution, excellent color, no flicker, best stereo image quality, very comfortable, low stereo crosstalk, continuous image
Cons:form factor, requires two input signals
3D Vision
Features:left and right images are displayed frame sequentially with black frame insertion between, an emitter synchronizes glasses so that the left eye is blocked when a right image is shown and the right eye is blocked when a left image is shown
Glasses: NVIDIA 3D Vision active shutter glasses
Pros: full stereo resolution, good color, low stereo crosstalk, low cost
Cons:low brightness, mild flicker
Other Active Glasses
Features:left and right images are displayed frame sequentially, an emitter synchronizes glasses so that the left eye is blocked when a right image is shown and the right eye is blocked when a left image is shown
Glasses: active shutter glasses
Pros: full stereo resolution, good color, low cost
Cons:low brightness, mild to significant flicker, mild stereo crosstalk
Anaglyph
Features:stereo paired images are colored with red/blue or red/green tint for the left eye/right eye image
Glasses: red/blue or red/green tint passive glasses
Pros: can be used on any color display (even paper), low cost
Cons:very high stereo crosstalk, loss of most color information
Patterned Polarizer
Features:odd display rows are circularly polarized in the opposite direction as all the even display rows, left/right images are split to either the even or odd rows
Glasses: passive circularly polarized
Pros: continuous image, comfortable
Cons:half resolution, limited vertical viewing angle, stereo crosstalk increases with viewing angle
Head Mounted Display (HMD)
Features:a display in front of the left eye shows the left image and a second display in front of the right eye shows the right image
Glasses: electronic glasses with two displays and appropriate optics
Pros: zero stereo crosstalk, private view
Cons:heavier glasses, eliminates view of surrounding, not multi-user, limited display technology
Dual Panel
Features:a pair of LCD panels are laminated and controlled with a custom algorithm that allows one panel to control the pixel intensity and the other to control polarization distribution
Glasses: passive polarized glasses
Pros: full stereo resolution, simultaneous left/right image displayed
Cons:low brightness through two panels, complex driving scheme, viewing angle restrictions
Autostereo
Features:Uses LCD with an added optical element (either lenticular lens or parallax barrier) to create a combined left eye/right eye presentation of the image on the screen
Glasses: no glasses required
Pros: glasses free, multiple viewing zones are possible
Cons:resolution reduced by half or more, high stereo crosstalk, complex content development
2.4.1 REAL D Technology
The REAL D 3D theatrical system delivers a high quality 3D movie experience to the viewer. The system is designed to operate on existing digital cinema systems, which include a 3D enabled server and a DLP Cinema™ projector (Cowan, 2007).
3.0 METHODOLOGYIn order to fully obtain the research required and to gather the accrued information to achieve the aim and objectives of the project, the research method is mainly based on secondary research. However, primary research could be considered as part of the method on this project.
3.1 Primary Research
Semi-structured interviewing is perhaps the most common type of interview used in qualitative social research. In this type of interview, some specific information which can be compared and contrasted with information gained in other interviews and also in other sources. (Dawson, 2006)
The using of questionnaires is important for a project. In this project, a combination of both open and closed questions can consider to be taken. That way, it is possible to find out how many people have good or bad experience with 3D movies and what public think about 3D movies compare with 2D movies on the same form.
Below are advantages and disadvantages of interviews.
Advantages:
Very useful in obtaining detailed information Obtaining a realistic picture of the way of people view Able to control the search design to fit the project needs Focus on specific subjectsDisadvantages:
Can be costly Time consuming, by the time the research is complete it may be out of date May not get response if emails or direct mailing are used3.2 Secondary Research
Sharp (2002) defined that secondary data is data collected by others and published in some from that is fairly readily accessible. In order to fully understand the development of live action and post produced stereoscopic, and its current role within the industry, this process requires gathering general information found through secondary data.
The secondary research was divided into three parts:
The principles of 3D stereoscopic, 3D display, and 3D glasses. The process of 3D stereo conversion. Stereoscopic filmmaking – Live-action shooting in 3D.In the first part, a subjective research was done by collecting the information of how human able to see 3D view both in the real world and on the screen through a viewing system. In the second part,
Advantages and Disadvantages of the secondary research method
Books and journals are likely to be the major source of the secondary data. Furthermore, they are extremely helpful for providing an accurate and in depth information on this project. However finding information on a specific subject is one of the challenges in this project unless there is a journal that focuses on it in its entirety. Although, some information that found from books may be out of date, yet, the history and the concept of 3D is never change. Hence, books consider one of the important sources for this project. Specialist publication such as 3D world and Post are very useful and can be found in the learning centre.
The data from the internet is a simple approach to rapidly find information relevant to the project. Especially for this project is considered very useful, considering the lack of any out of date information that’s books have provided. However it’s hard to determine what information is valid on the internet unless the information is clearly referenced such as entertainment technology center website and whitepaper.
3.3 Comparison of the Chosen Methodology
4.0 REFERENCESLo, C. and Chalmers, A., 2003. Stereo Vision for Computer Graphics, ACM Inc, pp. 110.
Julesz, B., 1971. Foundations of cyclopean perception. Chicago: University of Chicago Press.
Boev, A., Hollosi, D., and Gotchev, A., 2008. Classification of stereoscopic artefacts, tech. rep., Mobile3DTV Project.
WestOne, n.d. ‘Resources – Principles of 3D Stereoscopy’. [online] www.westone.wa.gov.au Date accessed: 17/01/2011
Schreer, O., Kauff, P., and Sikora, T., eds., 2005. 3D video communication. Wiley.
Mendiburu, B., 2009. 3D Movie Making. Oxford: Elsevier Inc.
B. Clark, 2010. ‘3D Production and Post’. [online] www.jhfestival.org Date assessed: 25/10/2010
Dr. Dawson, C., 2006. A Practical Guide to Research Methods. Oxford: How To Books Ltd
Sharp, John A., 2002. The Management of a Student Research Project. Burlington: Gower Publishing Ltd
5.0 BIBLIOGRAPHYWandell, B., 1995. Foundations of vision. Sunderland, MA, US: Sinauer Associates.
