Bhojan Anand Karthik Thirugnanam Jeena Sebastian Pravein G Kannan Akhihebbal L Ananda Mun Choon Chan and Rajesh Krishna Balan ID: 219731
Download Presentation The PPT/PDF document "Adaptive Display Power Management for Mo..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Adaptive Display Power Management for Mobile Games
Bhojan Anand‡, Karthik Thirugnanam†, Jeena Sebastian‡, Pravein G. Kannan‡, Akhihebbal L. Ananda‡, Mun Choon Chan‡ and Rajesh Krishna Balan†‡ National University of Singapore (NUS) and † Singapore Management University (SMU)
1Slide2
Problem: Display draws significant phone power
Key Challenge: No loss in end user experienceKey problem and Solution2Display45-50%
Network
35-40%
CPU
4-15%
Measured on HTC Magic while streaming a Youtube VideoSlide3
Dynamically adjust image brightness and LCD display backlight levels
50-70% display power savings with no significant user experience impactKey problem and Solution3Slide4
Key problem and Solution
BackgroundSystem DesignEvaluationDiscussion/Future Work 4Talk OutlineSlide5
LCD displays have two components:
Power consumed mostly by the BacklightThus brightening the image, and darkening the backlight saves power. 5Background - LCD DisplaysBacklight - Provides light, and consumes powerLCD Panel - Filters light based on image to be displayedSlide6
Method 1: Naively dim the display
Creates visible artifacts (flicker, brightness loss, etc)Especially noticeable in high frame rate applications6Background: Saving Display PowerSlide7
Method 2: Compensate with increased brightness
Linearly apply same transform to entire imageLeads to saturated images7Background: Saving Display PowerSlide8
Method 2: Compensate with increased brightness
Non linear approaches prevent saturation but cause contrast lossOur solution uses this approach intelligently8
Background: Saving Display PowerSlide9
Gamma Correction, or gamma, is a tone mapping function used to brighten scenesVery Low Saturation relative to linear
Low computational overhead 9Non Linear Gamma CorrectionBefore After: Gamma 2Slide10
10
Effect of Gamma on Image QualityOriginal
Image after Gamma Increase
(gamma=2)
Image after Gamma
Increase
and
backlight
reductionSlide11
Games are popular and resource intensive
Extremely high frame ratesFlicker and brightness changes very noticeable to users We use two representative gamesQuake III – Commercial First Person Shooting (FPS) game Planeshift – Massively Multiplayer Role Playing Game (MMORPG)
11
Test ApplicationsSlide12
Key problem and Solution
BackgroundSystem DesignEvaluationDiscussion/Future Work 12Talk OutlineSlide13
Ultimate goal: Save significant power with no loss in end user experience
Challenge 1: Understanding the relationship between the backlight intensity, gamma, image brightness, and the power consumedChallenge 2: Identifying human thresholds for brightness compensationChallenge 3: Dynamically applying the solution 13System Design: Key ChallengesSlide14
Relationship found to be linear. No other major contributing factors
14Challenge 1: Backlight vs PowerHTC HeroLaptop (W500)HTC MagicSlide15
Perceived brightness kept constantNon linear compensation necessaryUseful gamma range is 1 to 4
15Challenge 1: Backlight vs GammaSlide16
Obtained via small user study5 postgraduate students
Each user shown a range of imagesCovered a full range of brightnessFor each image, users had to boost gamma to obtain two quality thresholds Described in next slideTool provided boosted gamma at .1 intervals with automatic backlight compensation16Challenge 2: Human ThresholdsSlide17
Conservative: Image quality comparable to originalAggressive: Image quality is affected but acceptable
17Challenge 2: Two ThresholdsSlide18
18
Challenge 3: Runtime AlgorithmStartCalculate Average Brightness of last X Samples. Is there a change?
Mode + Brightness -> Gamma & Backlight
Leave Settings as it is
Sleep Thread for Y ms
Yes
NoSlide19
Test platformMost of the evaluation on LaptopPrototype Mobile implementation available (demo)
Objective Analytical Experiments Power measurements Measured power saved in different modesPerceived User ImpactLarge scale user study (60 users) with Quake IIIMeasured perceived quality loss in different modes19Evaluation MethodologySlide20
20
Evaluation: Three Test ModesUnmodified
Conservative
Aggressive
Two bounding modes tested but omitted for simplicity Slide21
A recorded trace used to measure powerEvaluation: Power Savings
21Slide22
Large Scale User study60 Singapore Management University undergrads.34 Male and 26 Female students with differing background and game experiences
Participants trained on an unmodified version of the gameThey then played the 3 different versions of the gamePlay order randomized with recalibration at every step22User Study MethodologySlide23
Users rated each version by 6 criteriaCovered different quality dimensions
Evaluation: User Study Results23
Bad
Good
Strongly Agree =
Strongly Disagree =
Neutral =Slide24
Difference between Aggressive and Conservative significant
24Evaluation: Aggressive vs ConservativeSlide25
Difference
betw. Conservative and DefaultEvaluation: Power vs Perception25Slide26
Conservative (Dynamic Conservative)High QualityPerceived quality comparable to default.
Significant Savings – 49% Aggressive (Dynamic Aggressive)High Power Saving – 68%Acceptable Quality26Evaluation: ConclusionsSlide27
OLED DisplaysDo not use BacklightPower consumption depends on displayed content
Algorithm needs to be rethoughtOther Limitations & Future WorkInitial User studies were small scale, and in a controlled environmentPower Measurements could be more accurate, esp for mobile phoneEvaluation of mobile implementation27DiscussionSlide28
General systematic approach to save power on LCD screens
Identified key parameters determining the quality vs power savings tradeoffImplemented and tested with 60 end usersSystem achieves significant power savings with minimal overhead and quality loss28ContributionsSlide29
Thank you
Any Questions 29