Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Design and implementation of a voice-driven animation system Wang, Zhijin 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2006-0128.pdf [ 9.7MB ]
JSON: 831-1.0051586.json
JSON-LD: 831-1.0051586-ld.json
RDF/XML (Pretty): 831-1.0051586-rdf.xml
RDF/JSON: 831-1.0051586-rdf.json
Turtle: 831-1.0051586-turtle.txt
N-Triples: 831-1.0051586-rdf-ntriples.txt
Original Record: 831-1.0051586-source.json
Full Text

Full Text

Design and Implementation of a Voice-Driven Animation System by Zhijin Wang B . S c , P e k i n g U n i v e r s i t y , 2003  A THESIS SUBMITTED I N PARTIAL F U L F I L L M E N T O F THE REQUIREMENTS  F O RT H E D E G R E E OF  MASTER OF SCIENCE  in  THE F A C U L T Y OF G R A D U A T E STUDIES  ( C o m p u t e r Science)  THE UNIVERSITY OFBRITISH C O L U M B I A February 2006 ©Zhijin Wang, 2006  Abstract This thesis presents a novel multimodal interface for directing the actions of computer animated characters and camera movements. Our system can recognize human voice input combined with mouse pointing to generate desired character animation based on motion capture data. We compare our voice-driven system with a button-driven animation interface that has equivalent capabilities. An informal user study has indicated that our voice-user interface (VUI) is faster and simpler for the users than traditional graphical user interface (GUI). The advantage of V U I is more obvious when creating complex and high-detailed character animation. Applications for our system include creating computer character animation in short movies, directing characters in video games, storyboarding for film or theatre, and scene reconstruction.  ii  Table of Contents Abstract  11  Table of Contents  i n  List of Figures  V 1  Acknowledgements  1.  2.  v  Introduction  u  1  1.1.  Animation Processes  1  1.2.  Motivation  2  1.3.  System Features  3  1.4.  A n Example  4  1.5.  Contributions  6  1.6.  Thesis Organization  6  Previous Work  7  2.1.  History of Voice Recognition  7  2.2.  Directing for Video and Film  •• 9  2.2.1  Types of Shots  9  2.2.2.  Camera Angle  10  2.2.3.  Moving Camera Shots  12  2.3.  Voice-driven Animation  13  2.4.  Summary  15  iii  3.  System Overview  16  3.1.  System Organization  16  3.2.  Microsoft Speech API  18  3.2.1.  API Overview  19  3.2.2.  Context-Free Grammar  20  3.2.3.  Grammar Rule Example  21  3.3.  Motion Capture Process  22  3.4.  The Character Model  25  4. Implementation  5.  28  4.1.  User Interface  28  4.2.  Complete Voice Commands  29  4.3.  Motions and Motion Transition  31  4.4.  Path Planning Algorithm  32  4.4.1  Obstacle Avoidance  32  4.4.2  Following Strategy  34  4.5.  Camera Control  35  4.6.  Action Record and Replay  4.7.  Sound Controller  36  4.8.  Summary  37  :  36  Results 5.1.  5.2.  38  Voice-Driven Animation Examples  38  5.1.1  Individual Motion Examples  5.1.2  Online Animation Example  5.1.3  Off-line Animation Example  42  5.1.4  Storyboarding Example  45  Comparison between GUI and V U I  iv  38 •  40  •  47  6.  Conclusions  50  6.1.  Limitations  50  6.2.  Future Work  51  Bibliography  53  V  List of Figures Figure 1.1  Example of animation created using our system  Figure 3.1  The architecture of the Voice-Driven Animation System  Figure 3.2  SAPI Engines Layout  19  Figure 3.3  Grammar rule example  21  Figure 3.4  Capture Space Configuration  23  Figure 3.5  Placement of cameras  24  Figure 3.6  The character hierarchy and joint symbol table  26  Figure 3.7  The calibrated subject  27  Figure 4.1  Interface of our system  29  Figure 4.2  Voice commands list  30  Figure 4.3  Motion Transition Graph  31  Figure 4.4  Collision detection examples  33  Figure 4.5  Path planning examples  34  Figure 4.6  Following example  34  Figure 5.1  Sitting down on the floor  39  Figure 5.2  Pushing the other character  39  Figure 5.3  Online animation example  41.  Figure 5.4  Off-line animation example  44  Figure 5.5  Storyboarding example  47  Figure 5.6  Average time of creating online animation using both interfaces  48  Figure 5.7  Average time of creating off-line animation using both interfaces  49  vi  5  :  17  Acknowledgements I would like to express my gratitude to all those who gave me the possibility to complete this thesis. The first person I would like to thank is my supervisor, Michiel van de Panne, for providing this excellent research opportunity. His inspiration, his encouragement, and his expertise in computer character animation were very helpful to me during the whole project and the thesis-writing. I could not have completed all the work without his invaluable guidance and support. I would also like to thank Robert Bridson, who gave me very good suggestions and useful comments on how to improve my thesis. I am very grateful to Kangkang Yin, who spent a lot of time teaching me and working with me on the Vicon Motion Capture System. I also wish to thank Dana Sharon and Kevin Loken for making the motion capture sessions a lot of fun to me. Finally, I am forever indebted to my parents and my wife Tong L i for their understanding, encouragement, and endless love.  vii  Chapter 1  Introduction By definition, animation is the simulation of movement created by displaying a series of pictures, or frames [1]. To create the illusion of movement, a picture is displayed on the screen then quickly replaced by a new picture similar to the previous one, but shifted or changed slightly. In the past when animation was created by hand, this process required a great deal of time, manpower, and recording equipment. However, today the animation process has become dramatically simpler. What was once done with pencils, sheets, and paints by a group of animators can now be accomplished by a single person with a personal computer. Not only has it become easier for people to create animation using computers, the application of computer animation has also spread into many aspects of our lives including advertising, architecture, education, industrial training, motion pictures, video games, and scientific visualization.  1.1.  Animation Processes  Broadly speaking, there are three types of processes to create computer animation: keyframing, physics-based simulation and motion capture. Keyframing is a popular method for animating articulated figures because it allows for artistic expressiveness by providing unlimited control to the animator. The animator defines a set of keyframes in which the properties of the character such as its position, orientation, and joint angles, are described at specific instants in time. The motion is then obtained by simply interpolating the property values between these keyframes, typically with cubic splines. The drawback of this process is that it requires significant time and effort for the animator. Defining a movement often  1  requires a high level of detail to ensure that the interpolation curves induce the desired collision-free motion. This is time consuming and generating new motions must happen offline. Physics-based simulation involves using mathematical equations to model natural phenomena and then using numerical methods to compute an efficient and accurate solution to the equations in order to create realistic passive motions. One class of physics-based animation uses forward-dynamic simulation and it has been used to analyze and reconstruct active human motions such as running, falling, diving and so on. A n alternative technique is that  of  "spacetime  automatically  constraints",  generate the  which  movements  optimizes  a complete  motion  trajectory  to  of physically simulated characters. However,  specifying the right control to achieve a desired result remains a difficult problem. Motion capture is the process of capturing the movement of a real object or person and mapping it onto a computer generated object or person to create natural looking animation. For character animation, motion capture systems are used to capture the motions of human actors by placing markers at various key points on the actors. Multiple cameras are used to observe and reconstruct the position of those markers in real time, and their motions are then used to reconstruct the motion of the underlying skeleton. A n advantage of motion capture over traditional animation techniques such as keyframing and simulation is the capability of real-time visualization and the high quality of the generated animation. However, there is no easy way to modify the captured data to interactively generate new motions for different characters or for different environments.  1.2. Motivation  No matter which one of the above methods is used to produce the animation, user interaction is always an important part of an interactive computer animation system. Without a good interface, even an otherwise powerful system cannot enable the creation of good animation.  2  Unfortunately, many current computer animation systems have complex interfaces and as a result, require extensive training or reading of manuals in order to learn how to use the system. There are usually a large number of buttons on various panels, and many keyboard shortcuts for an animator to select desired motions or to create diverse animation effects. The animator must use a mouse and keyboard to specify the path along which the character is moving as well as the specific types of motions to be carried at. This is time-consuming even when creating fairly generic motions. For example, if a character is to perform a slow walk, followed by running at normal speed, and then followed by fast walking while waving a hand, it is necessary to look for specific buttons among multiple panels in order to select the correct options. This kind of visual search is inefficient and distracting for an animator trying to focus on the animation. To improve user interaction and simplify the animation process, we present a novel voice-driven interface for computer animation systems. This has strong parallels with filmmaking, where there is a director in charge of filming, directing the actions of actors and the placement of cameras. If the director is not satisfied with a particular scene, he can either re-shoot the scene again or edit the footage in post-processing. The direction is often conducted through oral communication, which is the most natural way for both director and actors. In our system, the user becomes the director of the animation. He can instruct the movement of the characters and cameras using his voice, much like a movie director, albeit with a more limited vocabulary. Thus, we attempt to directly implement the director-actor metaphor for an animation system.  1.3.  System Features  To enable the voice-driven feature, our animation system uses Microsoft Speech API as the voice recognition engine. We define a set of grammar rules for recognizable voice commands. All speech from the microphone is processed by the system. If the spoken words match one of  the grammar rules, it can be recognized as a voice command that then triggers a motion for the character, a camera motion, or a system function. Our pre-defined voice commands are divided into three classes. The first class of voice commands allows users to control the actions of the characters. Our characters can perform simple actions such as walking, running, hand waving, as well as more abstract actions such as following another character or walking to a specified location while avoiding obstacles. The second class of voice commands allows for the control of camera movement. The user, who is the director of the animation, can make use of different classical film camera angles such as an "over the shoulder shot", "panning shot", "zoom in" and "zoom out". The third class of voice commands relates to global functions such as opening and saving animation files, resetting the scene and so forth. Our system can be used both online and off-line. During online use, the characters and cameras react in real-time according to the instructions they receive from the user. To create off-line animation, the user gives a series of instructions followed by an "action" command to indicate the start of the animation, analogous to what a film director might do. Our system is also able to record and playback spoken dialogue. This is a useful feature for directly adding scripted dialogue to a storyboard or animation. This is treated as a special mode where the voice recognition process is temporarily disabled in order to avoid any potential false recognition by the speech engine.  1.4.  An Example  Figure 1.1 shows an example of creating animation using our voice-driven animation system. The user directs the movement of the character by speaking into the microphone connected to the computer. The speech engine of our system will then analyze the speech input. If any of the voice commands is recognized, the animation engine will find the corresponding motion clip in our motion capture database. Spoken dialogue can be added and additional mouse  4  input may be provided by the user to specify certain locations. The desired animation will then be generated and displayed on the screen as a visual feedback to the user. More specific and detailed examples are given in Chapter 5.  1.5. Contributions The primary contribution of this thesis is to provide a method for the multimodal interactive control of 3D computer animation using spoken commands in order to approximate the director-actor metaphor. Specifically, we present a voice-driven interface for creating character animation based on motion capture data. This interface is able to accept voice commands and pointing input, which allows a user with little knowledge of computer animation techniques to quickly create character animations. We develop a system that implements this voice-driven interface as well as a functionally equivalent GUI-based interface. We qualitatively and quantitatively compare the performance of creating computer animation using this voice-driven interface with that of the GUI-based interface. Our system demonstrates the feasibility of directing computer animation using only spoken directives, as well as the potential advantages of voice-driven interaction over the traditional mouse and keyboard interaction.  1.6.  Thesis Organization  The remainder of this thesis is structured as follows. After reviewing previous related work in Chapter 2, we provide an overview of our system in Chapter 3. Chapter 4 describes the components of our interface and explains the details of our implementation. Chapter 5 demonstrates a variety of results obtained using our interface. Finally, conclusions and future work are presented in Chapter 6.  6  Chapter 2  Previous Work Voice recognition technology has evolved rapidly over the past decades. It is now being widely used in user interfaces to allow the user to provide input to an application with his or her own voice, typically as a substitute for typing on a keyboard or clicking a mouse. In this chapter, we first provide a brief historical overview of voice recognition technology. We then give a short introduction to directing for video and film. Lastly, we review work in animation that relates to the use of voice and natural language.  2.1.  History of Voice Recognition  Interestingly, a toy company logged the first success story in the field of voice recognition many decades before major research in the area was considered. The "Radio Rex" from 1920 was a celluloid dog that responded to its spoken name. Lacking the computation power that drives recognition devices today, Radio Rex was a simple electromechanical device. The dog was held within its house by an electromagnet. As current flowed through a circuit bridge, the magnet was energized. The acoustic energy of the vowel sound of the word "Rex" caused the bridge to vibrate, breaking the electrical circuit, and allowing a spring to push Rex out of his house. While Radio Rex was not a commercial success, the toy was an important first demonstration of voice recognition [2]. During the 1940s and 1950s the US Department of Defense (DOD) and MIT became the largest contributors to the speech recognition field. The US D O D was driven by an interest in automatic interception and translation of Russian radio transmissions. The actual output their devices generated from Russian statements was more amusing than accurate however, and the  7  project was deemed a dismal failure [3]. Speech research was hamstrung because systems were designed that acknowledged only complete words and sentences. They would attempt to statistically guess the words following another word. But they took no account of the sounds used in the formation of the words themselves. These word formation elements are called phonemes and are sounds such as T H , P H , etc. Work at IBM derived the first statistical correlations between phenome sounds and the words they ultimately formed [4]. This was to prove the cornerstone of future research. By 1975, Jim Baker from Carnegie Mellon had developed his revolutionary Dragon speech recognition model. I B M continued to pioneer their statistical approach to speech recognition, adapting many algorithms from information theory. By 1985 they had developed a system capable of discerning 5000 different words and phrases [5]. Although much speech research took place over the next ten years, little in the way of commercially-viable voice recognition-related products materialized. Then in 1995, I B M , who had conducted intensive research for 50 years but had released nothing commercially, finally entered into the consumer marketplace with VoiceType® and later ViaVoice®. In April of  1997, Dragon Systems countered, introducing Naturally Speaking®, the first large  vocabulary continuous speech software available. The  development of voice recognition technology  has facilitated the design of  multimodal interactions. One of the earliest multimodal interfaces was built in 1980 by Richard Bolt [6]. The "Put-That-There" system used basic speech recognition and equipment that could track where the user was pointing. This simple system allowed users to move colored shapes around on a screen. It was able to allow the free use of pronouns and figure out which object the user was talking about. The system also allowed users to give objects names, and then use those names to refer to the objects. The system was a first attempt at creating a more natural user interface.  8  2.2.  Directing for Video and F i l m  Video and film directors are artists who can take a completed script and imaginatively transform it into exciting sounds and images. Directors creatively organize many facets of production to produce works of art. They know how and when to use different types of camera shots and have mastered the use of composition, image qualities, transition devices, and relations of time and space. Directors know when and how to use different types of sound and how to control sound and image interaction. By using all of their creative powers, directors are able to produce films and video programs that have lasting value [7]. Directors prepare a shooting script by indicating specific types of images and sounds to be recorded within each scene. In order to record the different scenes and sequences described in the script, the director must organize the activities of many different people who are involved in production. The director must be able to communicate precisely and quickly with cast and crew. A specific communication system and language must be understood and followed by each person as a command is directed at them.  2.2.1  Types of Shots  A director's ability to select and control visual images begins with an understanding of specific types of shots. The camera can be close to or far away from the subject. It can remain stationary or move during a shot. The shots commonly used in video and film production can be described in terms of camera-to-subject distance, camera angle, camera movement, and shot duration.  Long Shot (LS) The long shot orients the audience to subjects, objects, and settings by viewing them from a distance; this term is sometimes used synonymously with the terms establishing  9  shot,  wide  shot, or full shot. A n establishing shot (ES) generally locates the camera at a sufficient distance to establish the setting. Place and time are clearly depicted. A full shot (FS) provides a full frame (head-to-toe) view of a human subject or subjects.  M e d i u m Shot (MS) A medium shot provides approximately a three-quarter (knee-to-head) view of the subject. The extremes in terms of camera-to-subject distance within this type of shot are sometimes referred to as a medium long shot (MLS) and a medium close-up (MCU). The terms two-shot and three-shot define medium shots in which two or three subjects, respectively, appear in the same frame.  Close Shot (CS) o r Close-Up (CU) The terms close shot and close-up are often used synonymously. A close-up refers to the isolation of elements in the shot and normally indicates the head and shoulders of a person. When someone is making an important or revealing statement or facial gesture, a close-up will draw the audience's attention to that event. Close-ups focus and direct attention and create dramatic emphasis. When they are overused, however, their dramatic impact is severely reduced. A very close camera position is sometimes called an extreme close-up (ECU).  2.2.2. Camera Angle  The camera angle is frequently used to establish a specific viewpoint, such as to involve the audience in sharing a particular character's perspective on the action. The goal may be to enhance identification with that person's psychological or philosophical point of view.  10  Point-of-View Shot ( P O V Shot) A point-of-view shot places the camera in the approximate spatial positioning of a specific character. It is often preceded by a shot of a character looking in a particular direction, which establishes the character's spatial point of view within the setting, followed by a shot of that same character's reaction to what he or she has seen. The latter shot is sometimes called a reaction shot. A closely related shot is the over-the-shoulder  shot (OS). The camera is  positioned so that the shoulder of one subject appears in the foreground and the face or body of another is in the background. Another variation on the point-of-view shot is the  subjective  shot, which shows us what the person is looking at or thinking about. Like point-of-view shots, subjective shots offer a nonobjective viewpoint on actions and events, and can enhance audience identification with more subjective points of view.  Reverse-Angle Shot A reverse-angle shot places the camera in exactly the opposite direction of the previous shot. The camera is moved in a 180-degree arc from the shot immediately preceding it.  Low-Angle Shot A low-angle shot places the camera closer to the floor than the normal camera height, which is usually at eye level. A low angle tends to exaggerate the size and importance of the subject.  High-Angle Shot A high-angle shot places the camera high above the subject and tends to reduce its size and importance.  Overhead Shot An overhead shot places the camera directly overhead and creates a unique perspective on the action. This can sometimes be accomplished by a set of periscope mirrors, and overhead track, or by attaching the camera to an airplane, helicopter, or crane. 11  2.2.3. M o v i n g C a m e r a Shots  A moving camera adds new information to the frame and often alters spatial perspective. A moving camera shot can maintain viewer interest for a longer period of time than a stationary camera shot.  Pan Shot A camera can be panned by simply pivoting it from side to side on a fixed tripod or panning device. This shot is often used to follow action without having to move the camera from its fixed floor position.  Tilt Shot A camera tilt is accomplished by moving the camera up and down on a swivel or tilting device. This shot is also used to follow action, such as a person standing up or sitting down. It can also be used to follow and accentuate the apparent height of a building, object, or person.  Pedestal Shot A camera can be physically moved up and down on a pedestal dolly. A hydraulic lift moves the camera vertically up and down within the shot, such as when a performer gets up from a chair or sits down. A pedestal shot allows the camera to remain consistently at the same height as the performer, unlike a tilt shot, where the camera height usually remains unchanged. Pedestal shots are rare, but a pedestal is often used to adjust the height of the camera between shots.  Zoom Shot A zoom can be effected by changing the focal length of a variable focal-length lens in midshot. A zoom shot differs from a dolly shot in that a dolly shot alters spatial perspective by actually changing the spatial positioning of objects within the frame. During a zoom shot the  12  apparent distance between objects appears to change because objects are enlarged or contracted in size at different rates. During a zoom-in objects appear to get farther apart, and during a zoom-out they seem to get closer together.  Dolly Shot A dolly shot is a shot in which the camera moves toward or away from the subject while secured to a movable platform on wheels. It is often needed to follow long or complicated movements of performers, or to bring us gradually closer to or farther away from a person or object.  Trucking Shot In a trucking shot, the camera is moved laterally (from side to side) on a wheeled dolly. The camera may truck with a moving subject to keep it in frame. If the dolly moves in a semicircular direction, the shot is sometimes referred to as an arc or camera arc.  Tracking Shot A tracking shot uses tracks laid over rough surfaces to provide a means of making smooth camera moves in otherwise impossible locations.  Crane or Boom Shot The camera can be secured to a crane or boom so that it can be raised and lowered or moved from side to side on a pivoting arm. This type of shot can create a dramatic effect when it places the subject in the context of a large interior space or a broad exterior vista.  2.3. Voice-driven Animation Voice recognition technology is currently used in automobiles, telephones, personal digital  13  assistants (PDAs), medical records, e-commerce, text dictation and editing. Like the mouse, keyboard, and the trackball, voice recognition can enhance the control of a computer and improve communication. We discuss three animation systems which resemble our own work by making use of voice recognition technology. The interactive movie system [8] has the capability of creating a virtual world with various kinds of realistic scenes and computer characters. People can enter cyberspace and experience the stories of the virtual world through interactions with the characters and the environments of the world. The interactions include voice and gesture recognition. However, only exaggerated gestures rather than minute gestures could be recognized because of low lighting in the area where the system played. As a result, the system ran almost exclusively on voice recognition and players were limited to available modalities that allowed only simple interaction. This system differs from ours in that the progress of the story is pre-defined and mostly controlled by the system; the user is only acting as a participant who can interact with the system at some branch points, where the results from voice or gesture recognition will determine the future development of the story. In our system the user takes the role of the director and controls the entire storyline of the animation. V I R T U A L T H E A T E R [9] provides a collaborative virtual environment for industrial tele-training. Instead of working with physical objects, these are represented by virtual objects that can then be placed in a virtual environment accessible to many users. The users, represented by avatars, can manipulate and interact with the objects as in the real world, gaining valuable experience and training before using the real equipment. The environment can be fully controlled in a wireless fashion where speech commands and user gestures are used as input, allowing for a more comfortable interaction between the human participants. For example, the user may simply say pre-defined commands such as "Go to the table" for the avatar to perform, or may change views by saying "Table Top View", and so on. However, the system is limited to a particular use, e.g. training operators of A T M switching equipment in this case. The pre-defined high level actions make it impossible to create any free-style animation with this system.  14  Perhaps the most closely related work to ours is that of [10]. This presents the design and development of a Voice User Interface (VUI), which maps predefined voice commands to a set of behavioral motions for humanoid avatars using Extensible 3D (X3D) graphics. It uses Virtual Reality Modeling Language (VRML), Humanoid Animation (H-Anim) Standard and Java Speech API. The V U I maps between human voice commands and a set of animations of the avatar, thus providing voice access to the animation application. Graphical User Interface (GUI) is also provided as an alternative interaction. Their work demonstrates the possibility of  integrating  speech-recognition  technology  into  Networked  Virtual Environments  (Net-VEs). However, it cannot record voice dialogues, and the avatar is restricted to do only a small number of actions, such as standing, walking, and jumping. There is no control over the camera shots either.  2.4. Summary In this chapter we have provided an overview of voice recognition technology, directing, and previous work in animation based on these. Our system shares the same motivation as most applications in this category: to speed up and simplify the human-computer interaction using natural spoken languages. We have developed a computer animation system with a number of unique features, which will be described in the following chapters.  15  Chapter 3  System Overview Our voice-driven animation system is divided into a number of functional blocks: speech recognition, voice command analysis, motion controller, and camera controller. The role of each component is elaborated in this chapter.  3.1.  System Organization  The organization of our animation system and the path of user input and workflow through the system are illustrated in Figure 3.1.  Speech Recognition The Microsoft Speech API is integrated into our system as the speech recognition engine. It takes and processes the user speech from the microphone. If the speech matches one of the pre-defined grammar rules, the speech engine will trigger a recognition event and pass the recognized rule id to the analysis module.  Voice Command Analysis This module analyzes the recognized voice command received from the speech recognition engine and then extracts the associated parameters, such as the walking speed of the character, and the zooming distance of the camera. These parameters are then passed to different controllers according to the type of the voice commands.  16  Speech Input  1Z. Speech Recognition  Motion Capture Database  Voice Command Analysis  7^ ±1 Mouse Input  Motion Controller  Camera Controller  Sound Controller  Voice Input  Path Planner  Character Animation  Figure 3.1: The architecture of the Voice-Driven Animation System  Motion Controller The motion controller determines which motion the character will be performing and then loads the corresponding motion data from the motion capture database. The controller is also responsible for the transition of consecutive motions to make the animation as smooth as possible. Additional mouse input may be provided by the user to indicate the destination of some movement which cannot easily be explained by speech input.  17  Path Planner After the motion and the destination have been decided by the motion controller, the path planner takes into account the location of obstacles, and generates a collision-free path to the destination along which the character will be moving. When a character is asked to follow another character, the path planner is also used to find out the shortest path to perform the following action.  Camera Controller Each camera movement supported by our system is implemented by the camera controller. The controller makes the necessary changes of the position and shooting directions of the camera whenever a camera-related voice command is issued by the user. A timer is used in the controller to produce the continuous zoom in/out effect as well as the pan shot over a specific point in the scene.  Sound Controller This controller consists of the sound recorder and sound player. The sound recorder records any voice input from the user other than the voice commands. The recorded sound is played back as the voice of the character by the sound player during the animation in order to create an audible dialogue between the virtual characters.  3.2.  Microsoft Speech A P I  We evaluated two mainstream voice recognition products that are currently readily available. IBM ViaVoice Dictation SDK, based on many years of development by I B M , and the Microsoft Speech API, also known as SAPI. Both of them provide speech recognition and speech synthesis engines, as well as a complete development toolkit and documentations. They each also have their unique features, such as the O L E automation support from  18  Microsoft, and the specialized vocabularies support from IBM. Our only requirement was that of being able to recognize a set of voice commands made up by common words. We chose to use the Microsoft Speech API as our voice recognition system in this project because of its functionality and because it is freely available.  3.2.1. A P I O v e r v i e w  The Microsoft Speech API (SAPI) provides a high-level interface between an application and speech engines for recognition and synthesis. SAPI implements all the low-level details needed to control and manage the real-time operations of various speech engines, and thereby reduces the code overhead required for an application to use speech recognition and text-to-speech. This makes the speech technology more accessible and robust for a wide range of applications [11]. The two basic types of SAPI engines are text-to-speech (TTS) systems and speech recognizers. T T S systems accept text strings and synthesize spoken audio using synthetic voices. Speech recognizers convert human spoken audio into text strings and files.  1  Application  Application  API S A P I Runtime  | DDI Recognition Engine  TTS Engine  Figure 3.2: SAPI Engines Layout [11]  In our system, we only use the Speech Recognition Engine. There are two types of utterances to be recognized by the Speech Engine. The first one is dictation, which means the  19  Speech Engine will attempt to recognize whatever the user is saying to the microphone. The speech recognition on such large vocabularies is extremely hard without a lot of training from the user, because the computer doesn't know what to expect from the user's speech without any given context. The other type of recognizable utterance is referred to as "command and control grammar", which means the application can tell the engine in advance what kinds of voice commands are expected. The speech engine will then try to match the speech with one of those commands during run time of the application. This is the approach we have adopted because it requires less user training while it has a higher recognition rate than dictation.  3.2.2. C o n t e x t - F r e e G r a m m a r  The command and control features of SAPI 5 are implemented as context-free grammars (CFGs). A C F G is a structure that defines a specific set of words, and the combinations of these words that can be used. In basic terms, a C F G defines the sentences that are valid, and in SAPI 5, defines the sentences that are valid for recognition by a speech recognition engine. The C F G format in SAPI 5 defines the structure of grammars and grammar rules using Extensible Markup Language (XML). The X M L format is an "expert-only-readable" declaration of a grammar that a speech application uses to accomplish the following: •  Improve recognition accuracy by restricting and indicating to an engine what words it should expect.  •  Improve maintainability of textual grammars, by providing constructs for reusable text components (internal and external rule references), phrase lists, and string and numeric identifiers.  •  Improve translation of recognized speech into application actions. This is made easier by providing "semantic tags," (property name, and value associations) to words or phrases declared inside the grammar.  20  The CFG/Grammar compiler transforms the X M L tags defining the grammar elements into a binary format used by SAPI 5-compliant speech recognition engines. This compiling process can be performed either before or during application run time. Since our system does not need to modify the grammar at run time, the compiled binary format is loaded statically before the application run time in order to get the best performance.  3.2.3. G r a m m a r R u l e E x a m p l e  Figure 3.3 shows an example of the grammar rules used in our system, defined in X M L .  <RULE ID="VID_TurnCommand" TOPLEVEL="ACTIVE"> <0>turn</0> <RULEREF REFID="VID_Direction" PROPID="VID_Direction"/> <0>by</0> <0><RULEREF REFID="VID_Degree" PR0PID="VID_Degree7></0> <0>degrees</0> </RULE> <RULE ID="VID_Direction" > <LPROPID="VID_Direction" > <P VAL=" VID_Left">left</P> <P VAL="VID_Right">right</P> <P VAL=" VID_Around">around</P> </L> </RULE> <RULE ID="VID_Degree" > <LPROPID="VID_Degree" > <P VAL=" VID_Ten">ten</P> <P VAL=" VID_Twenty">twenty</P> <P VAL=" VID_Thirty">thirty</P> <P VAL=" VID_Forty">forty</P> <P VAL="VID_Forty_Five">forty-five</P> <P VAL="VID_Fifty">fifty</P> <P VAL="VID_Sixty">sixty</P> <P VAL=" VID_Seventy">seventy</P> <P VAL="VID_Eighty">eighty</P> <P VAL="VID_Ninety">ninety</P> </L> </RULE> Figure 3.3: Grammar rule example  21  According to this grammar rule, if the user says "turn right by 70 degrees", then the speech engine will indicate the application that rule name "VID_TurnCommand" has been recognized, with the property of child rule "VID_Direction" being "right", and the property of child rule "VID_Degree" being "seventy". The recognized information is then collected by the system to generate the desired animation. Right now our system only supports the above degrees that are involved in the turning actions. These are the most common degrees for an ordinary user. More special degrees can be easily added to the grammar rules if necessary.  3.3.  M o t i o n Capture Process  Motion capture is the recording of 3D movement by an array of video cameras in order to reproduce it in a digital environment. We have built our motion library by capturing motions using a Vicon V6 Motion Capture System. The Vicon system consists of the following: • Vicon DATASTATION  i  • A host WORKSTATION P C • A 100 base T TCP/TP Ethernet network connection • 6 camera units mounted on tripods with interfacing cables • System and analysis software • Dynacal calibration object • Motion capture body suit and marker kit  The first step of motion capture is to determine the capture volume - the space in which the subject will move. The capture volume can be calculated based on the size of the space we have available, the kind of movements going to be captured, and the number and type of cameras we have in the system. The space of our motion capture studio is about 9*9*6 meters, with 6 cameras placed in the corners and borders of the room. The type of the cameras we use  22  is MCam 2 with 17mm lens operating at 120Hz, which has 48.62 degrees in the horizontal field of view and 36.7 degrees in the vertical field of view, according to [12]. Therefore, the capture volume is approximately 3.5*3.5*2 meters, as shown in Figure 3.4.  MCam 2 17mm lens 120Hz 3.5*3.5*2m Capture Volume  3.8 m  Figure 3.4: Capture Space Configuration  The capture volume is three-dimensional but in practice is more readily judged by its boundaries marked out on the floor. The area marked out is central to the capture space and at the bottom of each camera's field of view to maximize the volume we are able to capture. To obtain good results for human movements, cameras are placed above the maximum height of any marker we wish to capture and point down. For example if the subject will be jumping then cameras would need to be higher than the highest head or hand markers at the peak of the jump. Cameras placed in this manner reduce the likelihood of a marker passing in front of the image of another cameras strobe, as shown in Figure 3.5. Camera calibration is another important step of the preparation process. There are two main steps to calibration. Static calibration calculates the origin or centre of the capture volume and determines the orientation of the 3D Workspace. Dynamic calibration involves movement of a calibration wand throughout the whole volume and allows the system to calculate the relative positions and orientations of the cameras. It also linearizes the cameras.  23  Camera 2 (above Camera 1 Field of V i e w )  Camera  1  Figure 3.5: Placement of cameras [13]  At the beginning of the motion capture process, the actor with attached markers is asked to perform a  Range  of Motion  for the purpose of subject calibration. The subject calibration  analyses the labeled range of motion captured from this particular actor, and automatically works out where the kinematic segments/joints should be in relation to the range of motion, using the kinematic model defined by the subject template. It also scales the template to fit the actor's proportion, and calculates statistics for joint constraints and marker covariances. The result of this process is a calibrated subject skeleton which can then be fit to captured maker motion data. During the capture of the motion, the actor moves around the capture space which is surrounded by 6 high-resolution cameras. Each camera has a ring of L E D strobe lights fixed around the lens. The actor, whose motion is to be captured, has a number of retro-reflective markers attached to their body in well-defined positions. As the actor moves through the capture volume, light from the strobe is reflected back into the camera lens and strikes a light sensitive plate creating a video signal [14]. The Vicon Datastation controls the cameras and strobes and also collects these signals, along with any other recorded data such as sound and  24  digital video. It then passes them to a computer on which the Vicon software suite is installed for 3D reconstruction.  3.4.  The Character Model  We use Vicon iQ software for the post-processing of the data. We define a subject template to describe the underlying kinematic model, which represents the type of subject we have captured, and how the market set is associated to the kinematic model. The definition of the subject template contains the following elements [15]: A hierarchical skeletal definition of the subject being captured. Skeletal topology, limb length and proportion, in relation to segments and marker associations. A kinematic definition of the skeleton (constraints and degrees of freedom). Simple marker definitions, name, color etc. Complex marker definitions, and their association with the skeleton (individual segments) Marker properties, the definition of how markers move in relation to associated segments! Bounding boxes (visual aid to determine segment position and orientation) Sticks (visual aid drawn between pairs of markers) Parameter definitions, which are used to define associated segments and markers from each other, to aid in the convergence of the best calibrated model during the calibration process  Our subject template has 42 markers and 44 degrees of freedom in total. It consists of 25 segments, arranged hierarchically as shown in Figure 3.6. Motions are exported as joint angles time series, with Euler angles being used to represent 3 D O F joints.  25  Symbol  Joint Type  Degrees of Freedom  •  Free Joint  6  0  Ball Joint  3  +  Hardy-Spicer Joint  2  -  Hinge Joint  1  A  Rigid Joint  0  Figure 3.6: The character hierarchy and joint symbol table 26  The subject template is calibrated based on the range of motion captured from the real actor. This calibration process is fully automated and produces a calibrated subject, which is specific to our particular real world actor, as shown in Figure 3.7.  Figure 3.7: The calibrated subject  The calibrated subject can then be used to perform tasks such as labeling, gap filling, and producing skeletal animation from captured and processed optical marker data.  27  Chapter 4  Implementation Details of our implementation are provided in this chapter. Our voice-driven animation system is developed under Microsoft Visual C++ .Net IDE (Integrated  Development  Environment). We use M F C (Microsoft Foundation Class library) to build the interface of our system and OpenGL for the rendering of the animation.  4.1. User Interface Figure 4.1 shows the interface of our animation system. On the left, a window displays the animation being created. On the right is a panel with several groups of GUI buttons, one for each possible action. At the top of the button panel is a text box used for displaying the recognized voice commands, which is useful for the user to check whether the voice recognition engine is working properly. The action buttons are used to provide a point of comparison for our voice-driven interface. Each action button on the interface has its corresponding voice command, but not every voice command has the equivalent action button, because otherwise it would take excessive space to house all action buttons on a single panel, which makes the interface more complicated. Many of the action buttons are used to directly trigger a specific motion of the character, while some of them have associated parameter to define a certain attribute of the action, such as the speed of the walking action, and the degree of the turning action. These parameters.are selected using radio buttons, check box or drop-down list controls. As we can see from the interface, the more detailed an action is, the more parameters and visual controls have to be  28  associated with the action button.  Recognized Command  Cameia Fmnt  Top  Long  Medium |  |  Shouldel j  Ctase |  Panning j  Side  |  J  Moving Action Speed Normal  Walk  C  Fast  C  Stow Jump[  Run  j  Turning Action Tuin Riflht  Turn Left f" Simple Turn  )  jjj  Degrei  Curve T urning Action Curve Left  Curve Right J  Speed Control Faster  I  Stower  Stop  J  Pick Up | Draw Box  Wander  ]  Throw  |  Follow  |  End  |  Reset  j  5pecirieo«cnon Start  |  Record Sound  Record Action Start •File  j  Replay j  Save  Figure 4.1: Interface of our system  4.2.  Complete Voice Commands  Figure 4.2 gives the complete voice commands list that is supported in our system. There are a total of 23 character actions, 5 styles of camera movement, and 10 system commands. Some of the voice commands have optional modifiers, given in square brackets. These words correspond to optional parameters of the action, or allow for redundancy in the voice commands. For example, the user can say "left 20" instead of "turn left by 20 degrees". By omitting the optional words, the former utterance is faster for the speech engine to recognize.  29  I  Replay i  Select Character  ••••  Open J  Start  j  Two J  Both  |  [Character Action Commands]  [Camera Action Commands]  walk [straight/fast/slowly]  front/side/top [cam]  [fast/slow] backwards  close/medium/long shot  run [fast/slowly]  zoom in/out [to close/medium/long shot]  jump [fast/slowly]  [over]  shoulder  cam/shot  [on  character  [turn] left/right [by] [10-180] [degrees]  one/two]  turn around  panning cam/shot [on character one/two]  left/right curve [sharp/shallow] wave [your hand] stop waving/put down [your hand]  [System Commands]  push  draw box/obstacle/block  pick up  box/obstacle/block finish  put down  [select] character one/two  pass  both [characters]  throw  action  sit down  start/end recording  stand up  replay  shake hands  reset  applaud  save file  wander around here  open file  follow him don't follow/stop following walk/run/jump [fast/slowly] to here [and say/pick up/sit down...] when you get there, say... walk to there and talk to him...  Figure 4.2: Voice commands list  Voice commands can be expanded by creating new grammar rules for the speech engine, adding new motions into the motion capture database, and mapping the new voice commands with the corresponding new motions. Similarly, more tunings and controls over the existing motions can be added to our system, as long as the variations of the motions are also imported into the motion capture database. While the adding of new motions may require significant change and redesign of the graphical user interface, its impact on the voice-driven interface is much smaller, and the change is visually transparent to the users.  30  4.3.  Motions and Motion Transition  The motions of character in our system are animated directly from the motion capture data. We captured our own motions using Vicon V6 Motion Capture System. The capture rate was 120 frames per second, and there are a total of 44 degrees of freedom in our character model. By default, the animated character is displayed in the rest pose. Transitions between different actions may pass through this rest pose, or it may be a straight cut to the new action, depending on the types of the original and the target action. The allowable transitions are illustrated using the graph shown in Figure 4.3.  Adding compatible actions  Figure 4.3: Motion Transition Graph  In Figure 4.3, we divide the actions into several groups according to their attributes. Actions such as walk and run belong to continuous actions. Transition among these actions is done by a  31  straight cut, which may result in a visible discontinuity in the motion. This choice was made to simplify our implementation and could be replaced by blending techniques. Actions such as pick-up, put-down are called discrete actions. These actions cannot be interrupted. Transition among these actions is straight forward, because every action needs to complete before the next one can begin. Transition between continuous actions and discrete actions must go through the rest pose, which serves as a connecting point in the transition. Several additional actions such as talk, turn, and wave, can be added to the current actions where they are compatible with each other. For example, talking can be added to the character while he is performing any action, and turning can be added to any continuous actions since it only changes the orientation of the character over the time. Waving is another action that can be added to the continuous actions because it only changes the motion of the arm, while the motion of other parts of the body remains unchanged. A special motion clip is dedicated to blending the transition to and from the waving arm, which makes the motion less discontinuous and more realistic.  4.4.  Path Planning Algorithm  The character can perform two high level actions, namely walking to a destination while avoiding any obstacles, and following another character as it moves through out the space. These features are implemented by using simple path planning algorithms, as described below.  4.4.1  Obstacle Avoidance  In our system, obstacles are rectangular blocks in 3D space. The size and locations of the blocks are specified by the user and are given to the path planner. At each time step, the path planner checks the positions of the characters and detects whether there is a potential future collision with  32  the blocks. If there is any, then the path planner will compute another path for the characters to avoid hitting the obstacles. Since all the obstacles are assumed to be rectangular blocks, the collision detection is very straight forward. The path planner only needs to check whether the character is within a certain distance  from  the  boundaries  of  the  blocks  and  whether  the  character  is  moving  Figure 4.4: Collision detection examples  towards the blocks. A collision is detected only when these two conditions are both satisfied. Figure 4.4 illustrates an obstacle as a shaded rectangle, and the collision area as a dashed rectangle. The small shapes represent characters and the arrows represent their respective directions of motion. In this case, only the character represented by the circle is considered having a collision with the block because it is the only one that meets the two collision conditions. When a potential future collision is detected, the path planner adjusts the direction of the walking or running motion of the character according to the current position of the character and the destination. The character will move along the boundary of the block towards the destination, until it reaches a corner. If there is no obstacle between the character and the destination, the character will move directly towards the destination. Otherwise, it will move onto the next corner, until it can reach the destination  directly. Figure 4.5 shows examples of these two cases. The  circles represent the characters and the diamonds represent the destinations. Note that the moving direction of the character needs to be updated only when the collision is detected or the character has reached a corner of the blocks.  33  /  t  /  0  7\'  Figure 4.5: Path planning examples  4.4.2  F o l l o w i n g Strategy  Another high level action that the character can perform is to follow another character in the 3D space. At each time step, the path planner computes the relative angle between the current heading of the pursuer and the direction to the character, as shown in Figure 4.6. If 6 is greater than a pre-defined threshold, the character will make a corrective turn to follow the target character.  (xl.yl) Figure 4.6: Following example  If the other character has changed his action during the movement, say, from walking to running, the following character will also make the same change to catch up with him. The character follows the path along which the other character is moving, and also enacts the same action during the movement.  34  4.5.  Camera Control  The camera movements in our system modify the camera position and the look-at point. There are three kinds of static cameras available, including the front camera, the side camera and the top camera. Each of them can be used in one of three settings, including the long shot, the medium shot, and the close shot. While each setting has its own camera position, the look-at point of the static camera is always on the center of the space. There are three  types of moving camera shots in our system.  The first is  the  over-the-shoulder shot, where the camera is positioned behind a character so that the shoulder of the character appears in the foreground and the world he is facing is in the background. The look-at point is fixed on the front of the character, and the camera moves with the character to keep a constant camera-to-shoulder distance. The second moving camera shot provides a panning shot. With a fixed look-at point on the character, the camera slowly orbits around the character to show the different perspectives of the character. Regardless of whether or not the character is moving, the camera is always located on a circle centered on the character. The last moving shot is the zoom shot (including zoom in and zoom out), where the camera's field of view is shrunk or enlarged gradually over the time with a fixed look-at point on either the character or the center of the space. A zoom shot can be combined with over shoulder shot or panning shot simply by gradually changing the field of view of the camera during those two moving shots. The other types of camera shots that we discussed in Chapter 2 are not implemented in the current version of our system. We have only implemented the above camera shots because they are the most common ones used in movie and video directing, and they should be enough for ordinary users of our system.  35  4.6.  Action Record and Replay  Our system can be used in online and off-line modes. When creating an online animation, the user simply says a voice command to control the character or camera, and then our system will try to recognize this voice command and make the appropriate change to the character's action or the camera movement in real-time according to the instruction from the user. When creating an animation in off-line mode, the user gives a series of instructions followed by an "action" command to indicate the start of the animation. We use a structure called an "action point" to store information related to each of these instructions, such as the type and location of the action. Each time a new action point is created, it is places onto a queue. When the animation is started, the system will dequeue the action points one by one, and the character will act according to the information stored in each action point. The animation created in online or off-line mode can also be recorded and replayed at a later time. During recording, we use a structure called record point to store the type and starting frame of each action being recorded. The difference between a record point and an action point is that a record point can be used to store the information of online actions, and it contains the starting frame but not the location of each action. If some part of the animation is created off-line, the corresponding action points will be linked to the current record point. All the record points are put into a queue. Once the recording is done, the animation can be replayed by taking out and accessing the record points from the queue again. Since the record points are designed to be serializable objects, they can also be stored in a file and opened to replay the animation in the future.  4.7.  Sound Controller  We have integrated the Wave Player & Recorder Library [16] into our system as the sound controller. This allows spoken dialogue to be added to an animation, as may be useful for  36  storyboarding applications. The use starts recording his or other people's sound using a spoken "talk" or "say" command. Each recorded sound clip is attached to the current action of the character. During the recording of the sound, the voice recognition process is temporarily disabled in order to avoid any potential false recognition by the speech engine. When the recording is done, the user clicks a button on the interface panel to stop the recording, and then voice commands will be accepted again.  4.8. Summary In this chapter we have presented implementation details of our system. The character motions in our system are animated directly from the motion capture data. Transitions between different motions may pass through the rest pose, or it may be a straight cut to the new motion, depending on the types of the original and the target action. Path planning algorithm is used in obstacle avoidance and the following action of the character. Online and off-line animation can be created and stored in different data structures. Finally, spoken dialogue can be added to an animation via sound controllers in our system. In the following chapter, we will present voice-driven animation examples created using our system, and compare the results with a traditional graphical user interface.  37  Chapter 5  Results This chapter presents a selection of results generated with our voice-driven animation system, including some of the individual motions, and two short animation examples created using the online and off-line modes. In order to illustrate the animation clearly, we use figures to show discrete character poses that are not equally spaced in time. A comparison between our voice-driven interface and a GUI-based interface is also given at the end of this chapter.  5.1.  Voice-Driven Animation Examples  The animation examples below are created using voice as the only input, with one exception which is the off-line animation, where the user has to use extra mouse input to draw the block and specify the destinations of the movement of the character.  5.1.1  Individual Motion Examples  Figure 5.1 and 5.2 illustrate two character motions available in our system, i.e., sitting down on the floor, and pushing another character, respectively. The voice commands for these two motions are simply "sit down" and "push him". For the pushing motion, if the two characters are not close to each other at first, they will both walk to a midpoint and keep a certain distance between them before one of them starts to push the other one.  38  Figure 5.1: Sitting down on the floor  Figure 5.2: Pushing the other character 39  5.1.2  Online Animation Example  Figure 5.3 illustrates an animation example created online by using our system. The phrase below each screenshot is the voice command spoken by the user. When a voice command is recognized, the action of the character is changed immediately according to the command, which can be seen in the example images below.  40  "Stop waving"  "Turn around"  5.1.3  Off-line Animation Example  Figure 5.4 shows an example of creating off4ine animation combined with online camera controls. During the off4ine mode, the camera is switched to the top cam viewing from the above, and additional mouse input is needed to specify the locations of the block and the action points. Each action point is associated with at least one off4ine motion specified by the voice command. When the "Action" command is recognized, the animation is generated by going through the action points one by one. At the same time the user can use voice to direct the movement of the camera, which is illustrated in Figure 5.4.  43  "Zoom Out"  Figure 5.4: Off-line animation example  44  5.1.4  Storyboarding E x a m p l e  Figure 5.5 shows an example of using our system as an animation storyboarding tool. In this example, the user wants to act out a short skit, where two people meet and greet each other, and then one of them asks the other person to follow him to the destination. At first they are walking. After a while the leading person thinks they should hurry up, so they both start running together. Finally they arrive at the destination after going through all the bricks and walls on the way along. Instead of spending a lot more time on drawing a set of static storyboard pictures to describe the short skit to others, the user of our system can use his voice to create the equivalent animation in less than a minute, which is more illustrative and more convenient. Unlike the previous examples, the voice commands that are used to create this example are not displayed here. The sentences below some of the figures in this example are spoken dialogues which are added to the animation by the user.  "Nice to meet you!" "Nice to meet you too!"  45  "Come with me please. I will show you the way to the destination"  46  "Here we are!"  Figure 5.5: Storyboarding example  5.2. Comparison between GUI and VUI We have conducted an informal user study on a group of ten people to compare our voice user interface (VUI) with the traditional graphical user interface (GUI). None of the users have experience in using voice interface before, but all of them have at least five years experience with graphical user interface. Before using our system for the first time, the users need to take a 10-minute training session by reading three paragraphs of text to train the speech engine to recognize his or her speech patterns. Once the training is over, the users are given 10 minutes to get familiar with the voice user interface by creating short animation using voice commands. Then they have another 10 minutes to get familiar with the graphical user interface. After that, each of them is asked to create one online and one off-line animation using both interfaces according to the scripts that are given to them. The scripts consist of a series of actions which are listed along the X axis of Figure 5.5 and 5.6. Cumulative time is recorded as the user creates each action in the script using GUI and VUI. The average time of all the users performing the same task is plotted on Figure 5.6 for the online animation and Figure 5.7 for the off-line animation.  47  Figure 5.6: Average time of creating online animation using both interfaces  As we can see from the two figures, using voice user interface (VUI) costs less time than using graphical user interface (GUI) to create the same animation, whether online or off-line. We can also note that, when using GUI to create complex actions, such as "Turn right by 120 degrees" in Figure 5.5 and "Walk fast to here and pick it up" in Figure 5.6, it takes considerably more time than creating simple actions such as "Wave" and "Draw Box", due to the fact that complex actions involve searching and clicking more buttons in the graphical user interface. While using voice commands to create the same animation, it doesn't make much difference on the time taken between simple and complex actions, because one voice command is enough for either simple or complex action.  48  GUI ----- VUI  Cumulative Time(Sec)  Top  Draw Box  Box Finish  Walk fast to here and pick it up  Run slowly Jump to here to here and and sit down throw it  Action  Figure 5.7: Average time of creating off-line animation using both interfaces  When making the above comparison between GUI and VUI, we have not taken into account the extra time spent in the voice recognition training session. Although it takes more initial time than using graphical user interface, the training session is needed only once for a new user, and the time will be well compensated by using the faster and more convenient voice user interface in the long run.  49  Chapter 6  Conclusions We have presented a voice-driven interface for creating character animation based on motion capture data. The current implementation of our system has simplified and improved the traditional graphical user interface and provides a starting point for further exploration of voice-driven animation system. Our system can help users with little knowledge of animation to create realistic character animation in a short time. As this research is still in its first stage, our system has a number of limitations and directions for future work.  6.1. Limitations The use of voice recognition technology has brought some limitations to our animation system. First of all, each user needs to take a training session for the speech engine to memorize his or her personal vocal characteristics before using our system for the first time. Although it is feasible for someone to use our system without taking the training session, it will be harder for the speech engine to recognize his or her speech and thus results in lower recognition rate. A constraint of using any voice recognition system is that our system is best to be used in a quiet environment. Any background noise may cause false recognition of the speech engine and generates unpredictable results to the character animation being created. Currently our system only has a small number of motions available in our motion capture database. While capturing and adding new motions to the database is a straightforward process, it will require a more sophisticated motion controller, especially for motions that require interactions between multiple characters. Some of the interactions may not be easy to express only in spoken  50  language. In these cases, a vision-driven system with graphical user interface may be a better choice for the users. The transition between several actions in our system is now done by a straight cut, which may result in a visible discontinuity in the motion. It can be replaced by using blending techniques to make the transition more smooth and realistic. Presently, the environment in our system only contains flat terrain and rectangular blocks as obstacles, and the character is animated as a skeleton for the sake of simplicity. More types of objects could be added to the environment in the future to reconstruct a life-like setting, and skinning could be added to the skeleton to make the character more realistic.  6.2. Future Work One straightforward direction for future work is to add new motions to the existing motion capture database to enrich the capability of our animation system. It would be interesting to add new motions of characters interacting with other characters or objects in a more complex environment. As there are more complicated motions available, there might also be a need to redesign the grammar rules for the voice commands, so that the user can direct the movement of the character more efficiently. Our current system implements a simple path planning algorithm for collision detection and obstacle avoidance based on the assumption that all the obstacles are rectangular blocks in the 3D space. With the addition of new objects of various shapes in the future, we, will also need to employ more sophisticated path planning algorithm, such as potential field method and level-set method. Another possible direction for future work is towards using the timeline as another dimension of the motion. When creating an animation, the user would be able to tell the character exactly when an action should happen. It could be in an explicit form, e.g. "Walk for 5 seconds"; or in an implicit form, e.g. "Run until you see the other character coming". The addition of  51  timeline will make the animation more flexible and is a way of attributing additional "intelligence" to the characters, which can simplify the required directives.  52  Bibliography [I] Online Computer Dictionary for Computer and Internet Terms and Definitions. [2] Maurer, J., "History of Speech Recognition" 2002 [3] Golomb, J., Speech Recognition History, 2001 [4] Spectrographs Voice Recognition, 25th February 2001 [5] History of speech recognition [6] Bolt, R. "Put-That-There": Voice and Gesture at the Graphics Interface, Computer Graphics, 14(3):262~270, 1980. Proceedings ofSiggraph 1980 [7] Kindem, G , Musburger, R.B., Introduction to Media Production: From Analog to Digital, Focal Press Publications, 200J [8] Nakatsu, R etc. Interactive movie system with multi-person participation and anytime interaction capabilities, Proceedings of the sixth ACM international conference on Multimedia: Technologies for interactive movies, Bristol, United Kingdom, Pages: 2 - 10, 1998 [9] Oliveira, J.C. etc., Virtual Theater for Industrial Training: A Collaborative Virtual Environment, Proceedings of 4th World Multiconference on Circuits, Systems, Communications & Computers (CSCC 2000), Greece, July 2000 [10] Apaydin, O., Networked Humanoid Animation Driven by Human Voice Using Extensible 3D (X3D), H-anim and Java Speech Open Standards, Master's Thesis at Naval Postgraduate School, Monterey, California, 2002 [II] Microsoft Speech SDK 5.1 Documentation, 2004 [12] Vicon V6 Manual, Vicon Motion Systems Limited, 2002 [13] Vicon Manual - Preparation VI.2, Vicon Motion Systems Limited, 2002 [14] Vicon Manual VI.2, Vicon Motion Systems Limited, 2002 53  [15] Vicon iQ 1.5 Tutorial 1: Subject Calibration, Vicon Motion Systems Limited, 2004 [16] Beletsky, A . etc., Wave Player & Recorder Library http://www.  54  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items