View Full Version : Can anyone explain the Tech Report image quality tests?

8th September 2002, 13:55
From the Tech Report tests:

Parhelia Measurements (http://www.tecchannel.de/hardware/907/29.html)

G550 Measurements (http://www.tecchannel.de/hardware/907/27.html)

From what I understand(using the crude Babelfish translator), they claim the G550 has better 2D image quality because of the ff. reasons,

1) Rise and fall times of 1.72ns and 2.06ns respectively(compared to 2.64ns and 2.75ns for the Parhelia). What does this mean exactly?

2) The signal resembles closer to a square wave rather than the Parhelia's sine wave. Correct me if I'm wrong, but isn't the signal coming out from the video card supposed to resemble pure sine waves since it now belongs in the analog domain(all the way to the monitor output)?

Yes, I've heard a lot of subjective testimonies of how beautiful the Parhelia's 2D image quality is. However, I'd like to find out more about this objective test. Many thanks.

8th September 2002, 15:50
1. Shorter rise and fall times make a better signal. Here the G550 that tecChannel tested obviously was better than that of the Parhelia board they tested. However the rise and fall times should also be "symmetric", i.e. not much different from each other. The Parhelia does a bit better in this field.

2. No. In fact a square wave would be the ideal waveform, the longer the rise and fall times the less "sharpness" (to put it in easy words). However you won't achieve a real square wave and those CE filters only make things worse (sometimes MUCH worse as can be seen with NVidia).

I would not pay too much attention about postings on the web about the individual signal quality of gfx-cards as there is IMO generally too much zealotry. Just compare yourself and you'll get real results (of course only if you are sensible to those sometimes subtle differences). E.g. a lot of fanATIcs (Rage3D, anyone) said the R8500 to have great signal quality, however, I found the 2D of my old Radeon VIVO much better than the one of my current R8500 (which is btw confirmed by the tecchannels results). My G400 beats the old Radeon VIVO.

Zealotry can be found everywhere, be it with the AMD/Intel "wars", with NVidiots, with fanATIcs and as well with Matrox fans. I have seen a G450 and contrary to what some might say this had worse 2D than my G400.

Another thing to think of when it comes to 2D signal quality are those manufacturing tolerances. Only very few people have the luck to actually be able to test more than one G550/Parhelia/Radeon/GF4 based card so those results stay subjective (the same as with overclocking results. The fact that I can get my R8500LE to 300/300 does not mean that this is possible for everyone else).
And last not least newer high-quality monitors do have extra circuitry that can correct bad VGA signals and those might work quite well with the nearly perfect sine of the Parhelia (transforming it back to a more square-like signal), so that might explain the discrepancy between those measurements and the real reported results.

Ryu Connor
8th September 2002, 19:57
TecChannel is the author of that test.

Tech Report is a different entity and we have a different view upon the results presented than what TecChannel concluded.


10th September 2002, 06:55
Ryu, you got all the links on that page messed up.

Ryu Connor
10th September 2002, 09:18
They added new results since that news piece was "printed." So, yeah. The direct links are no longer accurate.