Automatic lecture recording [Elektronische Ressource] / vorgelegt von Fleming Helge Lampi
229 Pages
English

Automatic lecture recording [Elektronische Ressource] / vorgelegt von Fleming Helge Lampi

-

Downloading requires you to have access to the YouScribe library
Learn all about the services we offer

Description

Automatic Lecture RecordingInauguraldissertationzur Erlangung des akademischen Gradeseines Doktors der Naturwissenschaftender Universit€t MannheimVorgelegt vonM. Sc. Diplom-Inform. (FH)Fleming Helge LampiausHeidelbergKarlsruhe, 2010Dekan: Professor Dr. Felix Freiling, Universit€t MannheimReferent:Dr. Wolfgang Effelsberg, Universit€t MannheimKorreferent: Professor Dr. Ralf Steinmetz, Technische Universit€t DarmstadtTag der mndlichen Prfung: 16. Juli 2010Acknowledgments iAcknowledgmentsI would like to thank a number of people who have supported me during my years of study and promotion.At first, I would like to thank my professors at the University of Applied Sciences in Karlsruhe for their work during my diploma and master studies, emphasizing Prof. Dr. Peter A. Henning for having been the advisor of my two theses and for having me introduced to Professor Effelsberg. I would like to thank Professor Dr. Wolfgang Ef-felsberg for being my doctoral advisor, and inspiring me by many good and fruitful conversations with him and my colleagues. Thanks to all of them for their collabora-tion.Out of my colleagues, I especially like to thank Dr. Nicolai Scheele, Dr. Stephan Kopf, Dr. Thomas King, Hendrik Lemelson, and Dr. Tanja Mangold for their support and all the publications written jointly. In addition, I want to thank Manuel Benz and Patric Herweh for their really good work and Eva Haas for proofreading my disserta-tion.

Subjects

Informations

Published by
Published 01 January 2010
Reads 25
Language English
Document size 10 MB

Automatic Lecture Recording
Inauguraldissertation
zur Erlangung des akademischen Grades
eines Doktors der Naturwissenschaften
der Universit?t Mannheim
Vorgelegt von
M. Sc. Diplom-Inform. (FH)
Fleming Helge Lampi
aus
Heidelberg
Karlsruhe, 2010Dekan: Professor Dr. Felix Freiling, Universit?t Mannheim
Referent:Dr. Wolfgang Effelsberg, Universit?t Mannheim
Korreferent: Professor Dr. Ralf Steinmetz, Technische Universit?t Darmstadt
Tag der m?ndlichen Pr?fung: 16. Juli 2010Acknowledgments i
Acknowledgments
I would like to thank a number of people who have supported me during my years of
study and promotion.
At first, I would like to thank my professors at the University of Applied Sciences in
Karlsruhe for their work during my diploma and master studies, emphasizing Prof. Dr.
Peter A. Henning for having been the advisor of my two theses and for having me
introduced to Professor Effelsberg. I would like to thank Professor Dr. Wolfgang Ef-
felsberg for being my doctoral advisor, and inspiring me by many good and fruitful
conversations with him and my colleagues. Thanks to all of them for their collabora-
tion.
Out of my colleagues, I especially like to thank Dr. Nicolai Scheele, Dr. Stephan
Kopf, Dr. Thomas King, Hendrik Lemelson, and Dr. Tanja Mangold for their support
and all the publications written jointly. In addition, I want to thank Manuel Benz and
Patric Herweh for their really good work and Eva Haas for proofreading my disserta-
tion.
Furthermore, I would like to thank Jochen Braitinger for supporting me in getting
detailed information about the work-flows in broadcast stations, perfecting my educa-
tion concerning media production.
I would like to thank my family, in particular my aunt Alma Lampi and my mother
Helga Lampi, for supporting me during all the years. Finally, I like to thank my girl
friend Marion Wehde for her support, her patience and her love during all our years
and my studies. Even more, I thank her very much for proofreading my diploma the-
sis, my master thesis and my dissertation and her reliable strong support even in diffi-
cult situations.
Dissertation Fleming Lampi, Computer Science IV, University of Mannheimii Abstract
Abstract
Lecture recording has become a very common tool to provide students with additional
media for their examination preparations. While its effort has to stay reasonable, only
a very basic way of recording is done in many cases. Therefore, watching the result-
ing videos can get very boring completely independent of how interesting the original
topic or session was.
This thesis proposes a new approach to lecture recordings by letting distributed com-
puters emulate the work of a human camera team, which is the natural way of creating
attractive recordings.
This thesis is structured in six chapters, starting with the examination of the current
situation, and taking its constraints into account. The first chapter concludes with a
reflection on related work.
Chapter two is about the design of our prototype system. It is deduced from a human
camera team in the real world which gets transferred into the virtual world. Finally, a
detailed overview about all parts necessary for our prototype and their planned func-
tionality is given. In chapter three, the implementation of all parts and tasks and the
incidents occurring during implementation are described in detail.
Chapter four describes the technical experiences made with the different parts during
development, testing and evaluation with a view to functionality, performance, and an
proposal towards future work. The evaluation of the whole system with students is
presented and discussed in the fifth chapter.
Chapter six concludes this thesis by summing up the facts and gives an outlook on
future work.
Dissertation Fleming Lampi, Computer Science IV, University of MannheimZusammenfassung iii
Vorlesungsaufzeichnungen sind mittlerweile ein h?ufig verwendetes Mittel, um Stu-
dierende mit zus?tzlichem Material f?r ihre Pr?fungsvorbereitungen auszustatten.
Dabei muss der ben?tigte Aufwand im Verh?ltnis bleiben, so dass oft nur eine sehr
grundlegende und einfache Art der Aufzeichnung realisiert wird. Die daraus entste-
henden Resultate zu betrachten kann sehr langweilig werden, unabh?ngig davon wie
interessant das urspr?ngliche Thema oder die Vorlesung war.
Die vorliegende Dissertation schl?gt einen neuen Ansatz f?r Vorlesungsaufzeichnun-
gen vor, in dem ein verteiltes Computersystem die Arbeit und Vorgehensweise eines
menschlichen Kamerateams nachahmt, um auf diese Weise attraktive Aufzeichnungen
herzustellen.
Diese Arbeit ist in sechs Kapitel gegliedert und beginnt mit der Betrachtung der aktu-
ellen Situation und der gegebenen Vorgaben. Das erste Kapitel schliesst mit einer
Betrachtung von verwandten Arbeiten.
Kapitel zwei beschreibt das Design des Prototyp-Systems, das von einem menschli-
chen Kamerateam in der realen Welt abgeleitet und in die virtuelle Welt transferiert
wird. Es endet mit einer detaillierten ?bersicht ?ber die f?r den Prototyp notwendigen
Teile und ihrer geplanten Funktionalit?t. Kapitel drei beschreibt im Detail die Imple-
mentierung aller Teile und Ihrer Aufgaben, sowie die Besonderheiten, die w?hrend
der Implementierung aufgetreten sind.
Kapitel vier beinhaltet die technischen Erfahrungen die mit den einzelnen Teilen des
Systems w?hrend ihrer Entwicklung, ihrer Testphasen und der Beurteilung ihrer F?-
higkeiten gemacht wurden, insbesondere mit einem Fokus auf deren Funktionalit?t
und Performanz sowie einem Vorschlag f?r zuk?nftige Implementierungen. Die Eva-
luierung des gesamten Systems mit Studierenden wird in Kapitel f?nf detailliert be-
schrieben und diskutiert.
Kapitel sechs beschliesst diese Arbeit indem es die Fakten zusammenfasst und einen
Ausblick auf zuk?nftige Arbeiten gibt.
Dissertation Fleming Lampi, Computer Science IV, University of Mannheimiv Content
Content
1. Introduction............................................................................................................1
1.1. Basic Idea.......1
1.2. Starting Point .................................................................................................2
1.2.1. Traditional Lectures Today....3
1.2.2. Room for Improvement..........................................................................4
1.3. Constraints .....................................5
1.3.1. Types of Recording................6
1.3.2. Aesthetic Considerations .......................................................................8
1.3.3. Financial Constraints.............9
1.4. Related Work ...............................................................10
1.4.1. Surveillance Recording........................................................................11
1.4.2. Meeting Recording...............12
1.4.3. Documentary Recording......13
1.4.4. Presentation ........................................................................14
1.4.5. Lecture Recording................17
1.4.6. Additional Related Work.....21
2. Design of the Distributed System ........................................................................25
2.1. Analyzing the Real World...........25
2.1.1. Determining the ?Ingredients?.............................................................25
2.1.2. Details of a Real Camera Team...........................28
2.1.3. Important Constraints...........................................................................35
2.2. Determining the Parts of our System ...........................................................37
2.2.1. The Director .........................................................38
2.2.2. The Cameraman...................39
2.2.3. Sensor Tools.........................................................41
2.2.4. The Sound Engineer.............................................46
2.2.5. The Lighting Technician......................................48
2.2.6. The Audio/Video Mixing Console.......................................................48
2.3. System Overview .........................................................49
2.3.1. Virtual Director?s Main Ideas..............................51
2.3.2. Virtual Cameraman?s Main Ideas ........................................................53
2.3.3. Sensor Tools? Main Ideas....................................53
Dissertation Fleming Lampi, Computer Science IV, University of MannheimContent v
2.3.4. Virtual Sound Engineer?s Main Ideas..................................................54
2.3.5. AV Mixing Console?s Main Ideas.......................55
3. System Implementation ................................57
3.1. Director Module...........................................................57
3.1.1. Tasks to Fulfill.....................60
3.1.2. Implementation Details........................................................................63
3.2. Cameraman Module.....................74
3.2.1. Tasks to Fulfill ................................................................74
3.2.2. Implementation Details........76
3.3. Sensor Tools Module...................87
3.3.1. WLAN Indoor Positioning System......................................................88
3.3.2. Question Management Software..........................91
3.4. Sound Engineer............................................................98
3.4.1. Tasks to fulfill......................................................98
3.4.2. Implementation Details......................................101
3.5. Audio-Video Mixer/Recorder....................................105
3.5.1. Tasks to Fulfill ...................................................................................107
3.5.2. Implementation Details......110
4. Technical Experience.........................117
4.1. Experience with the director module ................................117
4.1.1. Evaluation of the Virtual Director.....................117
4.1.2. Testing Setup ................................119
4.1.3. Simulation Results.............................................121
4.1.4. Overall Performance ..........................................122
4.2. Experience with the cameraman module...................126
4.2.1. Performance of the image processing algorithms..............................126
4.2.2.of the camera controlling ............................129
4.2.3. Overall Performance ..........................................................................132
4.3. Experience with the Sensor Tools module.................133
4.3.1. Experience with the Question ? Answer Interaction Control ............133
4.3.2. Experiences with the Event Reporting...............................................135
4.3.3.with the Audio Streaming..............136
4.3.4. Overall Performance ..........................................138
Dissertation Fleming Lampi, Computer Science IV, University of Mannheimvi Content
4.4. Experiences with the AV Mixer/Recorder.................................................138
4.4.1. Experience with the AV Decoding....................138
4.4.2. Experiences with the Output Trigger.................139
4.4.3.with the Video Processing.............................................140
4.4.4. Experience with the AV Output.........................141
4.4.5. Overall Performance ..........................................142
4.5. Experience with the Sound Engineer module............................................142
4.5.1. Audio Mixing and Mastering.............................................................142
4.5.2. Overall Performance..........146
4.5.3. Improving the AV Mixer/Recorder ...................................................146
5. Evaluation with Students ................................149
5.1. Evaluation Description...............149
5.1.1. Evaluation Design..............................................149
5.1.2. Description of the sample and its participants...154
5.1.3. Operationalization of the constructs..................154
5.1.4. Presentation of the evaluation method...............................................159
5.2. Evaluation Results .....................................................159
5.3. Discussion of the evaluation ......................................................................165
6. Summary............................................167
6.1. The Virtual Camera Team.........167
6.2. Implementation Experiences......................................................................168
6.3. Evaluation Experience...............169
6.4. Rating the Prototype ..................................................................................170
6.4.1. Aesthetic Approach............170
6.4.2. Affordable .........................................................................171
6.4.3. Space- and Time-Saving Approach...................171
6.4.4. Successful Prototype..........172
6.5. Outlook ......................................................................................................173
6.5.1. Improving the Current Prototype .......................................................173
6.5.2. Extending the Current 173
6.5.3. Transferring Automatic Lecture Recording to other environments...173
7. Appendix............................................................................................................173
7.1. Configuration Files....................173
Dissertation Fleming Lampi, Computer Science IV, University of MannheimContent vii
7.1.1. XML file of the FSM used in our prototype ......................................173
7.1.2. Example Configuration File of the Cameraman173
7.2. Sourcecode Snippets ..................................................................................173
7.2.1. Function ?FSM.GetTimerInterval?....................173
7.2.2. Procedure ?AVMixer.startgettingFrames? ........................................173
7.2.3. Function ?DefineCompressionLine?.................173
7.2.4.?Sample2DB? .....................................173
7.2.5. Function ?DB2SampleValue?............................173
7.2.6.?WaveExtrema?..................................173
7.2.7. Function ?getFactor?..........................................173
7.3. Original Evaluation Papers........173
7.3.1. German Pre-Test Form......................................173
7.3.2. German Questionnaire .......................................173
7.3.3.Post-Test Form.....173
8. Bibliography ......................................................................173
Dissertation Fleming Lampi, Computer Science IV, University of Mannheimviii List of Figures
List of Figures
Figure 1: Invoice snippet of a live production .............................................................10
Figure 2: Part of a Storyboard of "Kaffee oder Tee" of the SWR ...............................29
Figure 3: Example of wrong recording positions according to the "Line of Action"..30
Figure 4: of correct to the "Line of Action".31
Figure 5: Schematic view of equipment for studio production (based on Schmidt,
2005) ....................................................................................................................33
Figure 6: The job of a cameraman as a work-flow......................41
Figure 7: Interaction Diagram of Lecturer and Questioner .........................................44
Figure 8: Systems and Communications Channels Overview.....49
Figure 9: Draft of the Prototype System ......................................50
Figure 10: Graph of an Exemplary FSM for Standard Lectures..................................52
Figure 11: Three exemplary exposures (Benz, 2007)..................................................77
Figure 12: Skin color detection example (Benz, 2007)...............78
Figure 13: Images of the steps in background subtraction (Benz, 2007).....................78
Figure 14: Skin detection and region joining example (Benz, 2007). .........................79
Figure 15: Results of two backlight compensation algorithms (Benz, 2007)..............79
Figure 16: Comparing annotation visibility of color and gray scale images. ..............80
Figure 17: Distance of two points in the RGB color space..........................................81
Figure 18: Frame arrangement tests of abstracted protagonist (Benz, 2007). .............83
Figure 19: Lecturer and Questioner aligned in a shot - counter-shot scenario. ...........83
Figure 20: Motion-triggered automatic zoom-out followed by automatic zoom-in
(Benz, 2007).........................................................................................................84
Figure 21: Information exchange from director to camera via the cameraman and back
(based on Benz, 2007)..........................................................................................84
Figure 22: Estimated area marked and seat of student confirmed. ..............................90
Figure 23: Standard interface of the questioners' client...............................................92
Figure 24: Popup on the Lecturer?s computer showing announced questions. ...........94
Figure 25: Basic question - answer interaction, amended with client screen shots.....95
Figure 26: GUI of the QM server, client on seat 77 asking.........................................98
Figure 27: Ideal silence (top), minimal noise as silence (middle), and this minimal
noise after normalization (bottom).....................................100
Figure 28: Part of a 440Hz sine tone at 48 kHz (top) and at 8 kHz (bottom)............104
Dissertation Fleming Lampi, Computer Science IV, University of Mannheim