p. not so great 2D/TVout quality (c't review)

[thop]

just read the parhelia review in the c't (best german computer magazine) and they say that the 2D quality is worse than with previous matrox cards

i kinda refused to believe it when tecchannel.de stated the exact same thing but c't do their tests very thouroughly and they can be trusted in such things.

here is a small part of that review roughly translated:

At 1280x1024/960 the Parhelia achieved very good results. However at 1600x1200x85Hz the matrox card is not better than most products of other companies. At least matrox offers the same quality on both outputs.
...
But the signal quality of the tv-out leaves much to be desired. Especially the video bandwith (on the composite output) of 2.3Mhz, which is important for the sharpness impression, is only half the bandwith of a matrox marvel g450 and about equal to GF4 cards. Resolutions up to 1600x1200 don't make much sense here.

they also say that BETA 98/ME drivers will be online somewhen in the next 8 weeks.

[MrGaribaldi]

Strange!

I would've thought that with three years in R&D they'd at least have the same quality as the G4x0/G5x0...

Especially since they're touting the imagequality quite a lot...

Anyone asked over at Matrox what's the reason behind this?

[Cheesekeeper]

Hmmm, that's a worry.

I'm sure the illustriuos BetaBoyz can shed some light on this: what's the quality really like?

[Mad Bunion]

okay...so this card is not the cheapest around and doesn't offer the blazing speed that the other manufacturers do....NOW THIS?

I realize that Matrox is trying to corner a particular market with the Parhelia, but even "that" discerning market which expects Matrox's excellent 2D must be thinking about this little bit of info....(if it is true)

hope this is something an updated driver set can fix though...that might be wishful thinking though

[Greebe]

Then they don't have something right. P has by far better 2D/Vid out of any card, period... Shoot it's so good I don't know how they could improve it further.

[Dr Mordrid]

Noshit.....

2D here is VERY CLEAN...and that's coming from the video editing perspective.

Dr. Mordrid

[Mad Bunion]

that's good to hear....actual users disputing the technical results by the reviewer.

just curious, Dr Mordrid...have you ran your Parhelia at the resolution which c't was questioning?

[Ryu Connor]

http://www.tech-report.com/onearticle.x/3747

I disagree as well.

[MrGaribaldi]

Thank you!

I was getting kinda scared here! But now that fear has gone

[thop]

Originally posted by Ryu Connor
http://www.tech-report.com/onearticle.x/3747

I disagree as well.

i don't think the guys from tech-report got that right. the review they linked actually thinks the 2D quality of the parhelia is worse than the G550.

they gave the P 7.3 and the G550 8.2 - various GF4/ATI8500 cards got between 4.7 and 7.6, they did only run the signal quality tests at 1280x1024x85Hz though.

i guess everyone has to see for himself if the quality got worse or better

[Ryu Connor]

Actually I am one of the Tech-Report guys and I disagree with tecChannel's evaluation of the G550.

The Parhelia traces are excellent, reproducible, smooth, and they match for all colors.

The G550 traces show "ringing" at both rise and fall, which is a sign of filter cutoff sharpness -- or rather, lack of it.

I think tecChannel was centering completely on the rise and fall times of the signals and that just doesn't give you a complete picture.

[JohnMK]

You may refuse to believe it, but the weight of the objective evidence shows that 2D quality has worsened in this new video card. At the very least, the quality is going to be rather luck-of-the-draw. You might get a good card one day and a bad card the next. Inconsistency is what we've never known from Matrox, but now we may have to get used to it.

I'm of the opinion though that the quality has worsened and will be consisentely worse with this card. You can trust the German sites. They're excellent engineers and I trust their judgment. Just several months back tecchannel.de did a signal quality comparison of several cards, it's worth a look. It will certainly establish their credentials.

http://www.tecchannel.de/tecdaten/show.php3?catid=148&pageid=989&sort=wertung&sortorder=desc

You see a couple posts in this thread refuting the objective assessments with very subjective opinions. I'll trust the opinions of objective science over that of those whose opinions (subjective, at that) are possibly tainted by their loyalty to a previously excellent video card manufacturer. They simply refuse to believe the evidence. Companies change, I tell you. And it appears that Matrox has changed for the worse. It wouldn't be the first company it's happened to.

http://www.tecchannel.de/hardware/907/25.html

Pertinent section translated courtesy of CiTay from CiTay.de:

The Parhelia's signal quality matrox so proudly talked about turned out to be only average. It doesn't reach the G550's level. Okay, it has an amplitude of 703 mV, close to the optimum of 700. The RGB signals are congruent, too. But the rise- and fall times are worse, and the signal is forming something like a sinus wave, not a square wave like it should.

[Ryu Connor]

I figured they were centering on rise and fall times. The fact is that the VSIS standard states that rise and fall times should be as balanced as possible.

Longer rise and fall times are indicative of limited video bandwidth and can cause loss of spatial resolution. The lack of video bandwidth due to the unbalanced rise and fall times also explain the ringing seen in the rise and fall of it's signal (with regard to the G550).

Consequently the G550 is unable to hit 700mV. That will cause contrast and sharpness difficulties as it can't effectively swing from white to black.

Sorry, but I think tecChannel just totally got this one wrong.

[xortam]

Thanks for the discussion. I had some concern about the IQ based on some of the review comments but they didn't seem overly confident in their analysis. This discussion is enlightening.

[Dr Mordrid]

I also think tecChannel is totally out of whack here as well.

In terms of the VGA what I've been looking at for weeks is just dead flat gorgeous.

I'm also exporting video at 720x480 through DVDMAX/S-video to a JVC-9600 SVHS and it looks better than anything I've used before. When I put the signal on a vectorscope/waveform monitor the signal is clean, strong and properly aligned.

I don't think one could ask for more than that.

Dr. Mordrid