Timeliness Counter Measures I

1. Control the System

• Hardware and OS.

• Often not practicable.

• May require industry support.

2. Test, Profile & Benchmark

• Instrument all software layers.

• Graphics is not enough.

• Avoid test interference.

3. Adopt Principles

• Learn from Embedded World.

• OS and HW often specialized.

Observation

• Laborious method prone to error in the identification of the turning points in motion.

• Temporal resolution: Multiples of video’s FPS.

• Only few prerequisits

• Quality increase using high-speed camera.

Results

• Superior results compared to manual frame counting with 500Hz camera.

1. Display

• Display moves with user, e.g., HMD.

2. Head

• Display fixed w.r.t. user, e.g., CAVE.

3. Object

• Fixed view and screen.

Properties

• No calibration requirements (as only the change in gradient expressed by each signal is considered) and the

• High sampling rate (up to 44KHz) considerably higher than display refresh.

Idea

• Derives from manual frame counting.

• Identify a distinctive pattern of motion and then identify at which frame this began.

Observation

Identifying a distinctive pattern becomes a problem of feature detection in the single dimension.

Goals

• Determine usability of input systems.

• Provide standardized comparison.

• Play little VR Games.

• Measure speed & accuracy.

Pros

• Indication found that benchmarks provide an improvement over the analysis of technical specifications.

• Measures what matters: effects on users.

• Includes objective numbers.

• Combines overall quality aspects.

• Good for all I/O tests up to the application.

Cons

• Excludes actual application.

• Needs extensive runs.

• No easy isolation of detrimental sources.