DGA-6000 ULTRA-HD VIDEO GENERATOR / ANALYZER USER MANUAL
B-7
Introduction to Video Motion Testing
Quantitative and Qualitative Testing
Motion patterns are an essential tool for assessing video performance. The AccuPel DGA-6000 motion
patterns are digitally generated in real-time using mathematical algorithms that produce perfectly uniform
spatial frequency characteristics for the measurement and evaluation of displays and video processors. The
digital patterns are generated using the native sample rates of each user-selected video format so there are
no stored patterns to resampling or scale, which would degrade the spatial frequency response and
introduce artifacts into test patterns.
Experienced video engineers will recognize the spatial frequency characteristics and various non-linear
artifacts of particular video processing algorithms and digital processing techniques used in displays and
video processors. Video processing implementations can be tuned to minimize those artifacts and the loss
of spatial frequency response (resolution) using the supplied zone plates and motion patterns. Detailed
quantitative analysis of motion-adaptive and motion-compensated video signal processing requires an
extensive understanding of those algorithms in the spatio-temporal frequency domain, which is beyond the
scope of this User Manual.
However, quantitative measurements of effective horizontal and vertical display resolution versus motion
rate and direction can easily be made to characterize and compare displays, and to optimize a display’s
video processing settings (such as sharpness, detail enhancement, smooth-motion modes, deinterlacing
modes, and overscan). In addition, qualitative observations will reveal linear and non-linear processing
artifacts that can be used to optimally adjust or rank the quality of video processing in displays and
standalone video processors. Qualitative analysis is readily accessible for video enthusiasts and product
evaluators, and with experience can be correlated to typical video image problems observed in common
source material.
Frequency Domain Concepts
Spatio-temporal frequency domain –
Video signals represent discrete two-dimensional images (fields or
frames) that can only change at fixed periodic intervals. Hence, video signals are actually a function of three
dimensions – horizontal (x) and vertical (y) position (which define an image within a field or frame), and
discrete instances in time (t). Consequently, detailed analysis of video signals and video signal processing
can get complex because signals and moving images are functions of a 3-dimensional spatio-temporal
frequency domain.
Spatial Frequency vs. Temporal Frequency -
Temporal frequency refers to the number of magnitude
cycles that occur in a defined interval of time. Analog audio signals are a common example of temporal
frequency. The amplitude of an audio signal may swing between high and low excursions 1000 times per
second. In that case the signal frequency is 1000 Hz (1 kHz), where Hz (Hertz) is the unit cycles per second.
Video signals and pictures are defined by the composition of frame (or field) images and how they change in
time from frame to frame (or field to field). Individual field or frame images can be characterized by spatial
frequencies, which are the number of periodic changes (cycles) in brightness intensity that occur over a
distance (spatial) interval, rather than a time interval. Video images occupy a 2-dimensional plane, so spatial
frequencies exist in two dimensions. In some situations the horizontal spatial frequency and the vertical
spatial frequency can be considered independently, but in other instances those orthogonal frequency
components can be combined into a radial frequency at a defined angle. However, to fully understand video,
spatial and temporal image changes must be analyzed together in the 3-dimensional spatio-temporal
frequency domain.
Digital Video Structure & Sampling Frequency
- The temporal sampling frequency of the video structure
is the picture rate. New temporal samples (fields or frames) occur at the picture rate, which typically might
be 59.94 frames per second, 50 fields per second, 24 frames per second, and so forth. The term “picture”
refers to either “fields” or “frames”, which is understood by context.
For digital video formats the spatial sampling frequencies are the horizontal sampling frequency defined by
the total number of pixels per picture width (Pixels/PW), and the vertical sampling frequency defined by the
