Image Quality & System Design Considerations Stuart Nicholson Architect / Technology Lead Christie
SIM University - Objectives 1. Review visual system technologies and metrics 2. Explore connections between technologies, system capabilities and customer requirements 3. Exchange ideas how to best solve current and future requirements in visual (and sensor) simulation 3
Customer Specifications & System Design Customer Specifications Visual / Sensor System Specifications SIM University System Design Technologies Products 4
Outline Disruptive Technologies and Visual System Specifications The Changing Landscape Specifications for Image Quality Current specs - Theory vs. Application Spatio-Temporal Resolution / Sample Rates Alternative System Approaches to Optimize Image Quality 5
Visual Systems: from Analog to Digital Model Board Database Camera Image Generator Processing Processing Display (CRT) Display (Digital) Images: Link L-3 website 6
Specifying Visual System Requirements Proposed Criteria includes: 1. Ensure the system meets the training tasks 2. Develop with an understanding of the technology and test methods Recognizing applicability to current technology But without constraining or dictating a particular technology or design Meeting a clearly defined and measurable test criteria 3. Account for user s visual perception system 7
Static vs. Dynamic Resolution MTF Curve for moving bar pattern How do we account for this in our requirements/specs? How to improve? Interdependence between spatial/temporal resolution? Relative impact on image quality? 8
IG MTF Advertised vs Actual As advertised As Calculated As measured? How does this relate to training value? 9
Visual Systems: from Digital to Perceptual Real-time man-in-theloop visual simulation (not TV or Cinema) Database Image Generator Pixel Data per Frame Processing Display (S/H) Human Visual System Retinal Image Perceived Image Gaze Control System 10
Motion-Induced Artifacts Motion-Induced Blur Interaction between eye movements and display sample and hold filter Proportional to image motion and hold time (update rate) Spatio-Temporal Aliasing Function of image content (frequency and motion) and update rate Provides false or poor velocity cues Proportional to image motion AND resolution 11
Motion-Induced Blur Dr. Barbra Sweet, presented at Christie Sim University, Orlando 2012 12
Motion-Induced Blur 6 arcmin Resolution, 60 Hz Dr. Barbra Sweet, presented at Christie Sim University, Orlando 2012 13
Motion-Induced Blur 6 arcmin Resolution, 120 Hz Dr. Barbra Sweet, presented at Christie Sim University, Orlando 2012 14
Motion-Induced Blur Resolution vs Update Rate Resolution of 6 deg/sec Motion, 60 Hz vs 120 Hz vs 240 Hz Blur (temporal resolution) at any image speed improves with update rate Dr. Barbra Sweet, presented at Christie Sim University, Orlando 2012 15
Motion-Induced Blur Motion vs Update Rate 2 arcmin Resolution, 60 Hz vs 120 Hz vs 240 Hz Increasing update rate allows increased image motion before blur drops effective resolution Dr. Barbra Sweet, presented at Christie Sim University, Orlando 2012 16
Methods to Reduce Motion Induced Blur Perceived Motion Induced Blur can be reduced by shorter draw (hold) times But with reduced max brightness potential Flicker limited & increased image break-up / artifacts Hold times of 5.8 msec may provide results comparable to CRT displays Task Validation of Display Temporal-Resolution Measurements, I/ITSEC Conference 2007 17 Optional
Blur Perception Testing Task Validation of Display Temporal-Resolution Measurements, I/ITSEC Conference 2007 18 Optional
Dynamic Image Quality & Update Rates Problem: Christie AccuFrame addresses smearing, but: At the expense of peak brightness, image break-up and limited by flicker Does not address temporal aliasing Increased Update Rates address both perception of smear and spatiotemporal aliasing, but: 120Hz not compatible with all current Databases and IG channels One Solution: Composite multiple IG channels to achieve 120Hz update rate Leverage system architecture with dual-input modes Demonstrated at I/ITSEC 2011 and SIM University 2012 (WUXGA) Demonstrated at I/ITSEC 2012 (WQXGA) Demonstrated at SIGGRAPH 2013 (4K) with Christie TruLife TM platform 19
120/60/120 Hz Demo Side-by-Side (2) Matrix StIM WQs, edge-butted running 60Hz/120Hz (1) Matrix StIM WU running 120Hz (5) Vital X IG channels each running at 60Hz Host application runs at 120Hz or Host runs 60Hz and IG interpolates based on time stamp 120Hz output capability without changes to IG Reduced Spatio-Temporal Aliasing Noise maximizes training value 20
120/60/120 Hz Demo Host @ 60 (or 120 Hz) IG Chan @ 60 Hz 2560 x1600, 60Hz IG Chan @ 60 Hz 2560 x1600, 60Hz Matrix StIM WQ 2560 x1600, 120Hz IG Chan @ 60 Hz 2560 x1600, 60Hz Matrix StIM WQ 2560 x1600, 60Hz IG Chan @ 60 Hz 1920x1200, 60Hz IG Chan @ 60 Hz 1920x1200, 60Hz Matrix StIM WU 1920x1200 120 Hz 21
Impact of Update Rates 22
Impact of Update Rates 23
Conclusions Requirements and specifications tied to existing technologies can limit the potential solutions Image quality requirements should reflect the training objectives Blur can be reduced with lower display hold-time Blur and Spatio-temporal aliasing can be reduced with higher system update rate Higher updates rates can be achieved with current IG systems Breakup can only be minimized Update rate Reducing resolution Best approach may be a combination of methods and system level solutions 24
Thank you. Questions?