REPORT: Input lag Part 8
Overview of the monitors used in the test.
Since the synchronous image output also had to be tested, it was necessary to select a range of graphics cards which covered various levels of evolution or generations of the graphics card market so that special cases and deviations would be recognised.
We used the following graphics cards: ATI 9600XT, HD2600, HD3870, HD4850 and NVidia GeForce 4 4200Ti, FX5900, 8600GT, 8800GTS. Unfortunately, a GeForce 3 was not capable of supplying two monitors with an image at the same time.
The latest WHQL graphics card driver was installed on a wide range of computers, whereby the driver range tested for ATI stretched from Version 6.2 for the 9600XT using Windows 2000 to 9.5 using Windows XP for all modern cards. For the Nvidia cards, the ForceWare driver 93.71 was used for the Ti4200 and FX5900 models, with the former tested on Windows XP and the latter on Windows 2000. For the newer cards, the GeForce driver was used in version 185.85.
The computer systems used should not play a significant role, but we used a large number of different architecture. Beginning with an old AMD Duron, then the Athlon XP, Athlon64, a dual-Opteron Workstation and finally, the Quadcore from Intel, a wide range of system components was provided, not just in terms of processor technology. This helps to prevent special cases in the otherwise not examined system hardware, the set-up of the systems and software used from causing systematic errors to occur during the test.
Overview of devices used in test.
A large number of stopwatch programs can be found on the Internet, so why is it necessary to program new software for this test?
Stopwatch programs to date have been limited to displaying a counter and do not place particular importance on getting around "normal" display limitations. Thus, a stopwatch is generally fast enough if it can display a new value with every new frame displayed on the monitor and this value also corresponds to the time that has passed.
It is not exactly complicated to program such a stopwatch, but this does not automatically meet the requirements which it should ideally fulfil for a test of the input lag.
Flash has a low target for image refresh rates of 12 fps for historical reasons, but these can be adjusted. Calculations, e.g. sharpening, are always only carried out with this fixed rate. It is possible to set the frame rate to 60 Hz, but this rate fluctuates strongly around the value selected. This can be tested with the TeddyMark benchmark tool, for example.
Here, the extreme differences when different browsers are used to display the Flash videos are interesting. On Internet Explorer 8, the maximum possible frame rate for an area of 600 x 400 pixels is about 65 fps, whilst the same sequence on Firefox runs at a stately 120 fps. The fluctuations in frame rate are over ten percent for both browsers and arise sporadically. Changes to the display size affect the playback speed at the latest if the CPU load is 100 percent, whereby the percentage of fluctuation is retained.
© 2002 - 2017 PRAD ProAdviser GmbH & Co. KG | All rights reserved! | Imprint | Privacy Notice
Registered trademarks and brand names are property of their respective owners.
PRAD assumes no liability for the content of external pages that we link to!
|General information: Our website is not partially displayed correctly with adblocker enabled!|