2007WFP gradient banding

[imported_greyhound]

I tried spronkey's test with the L2000c via VGA - all looked good, no banding noticed. I'll try DVI later. Sorry, no digital camera here!

 

[Heartbreaker]

Originally posted by: VV1920

Originally posted by: cmge
i just got my 2007wfp as well im still a little confused with the banding issue... is this good or bad?... i compared my dell 2007wfp with my friends samsung 204B

IMHO your banding isn't that severe, cmge; just a little bit in the green row.

Obviously this is just a computer screen capture, not an actually picture of what the screen shows. This tells us nothing. Well except what jpg can do to a perfect gradient. This is why I said I would shoot mine in RAW, and upload png files.

What next, people start showing their burnt pixels by doing a screen cap in windows?

You need to take an actual picture with a camera, to show screen defects. No wonder I am a skeptic.

If you want to demo how good/bad your issue is:

1: DVI input required. VGA can also be included as a comparison, but DVI is where the bulk of the issue lies. Please include that.

2: Consistent gradient source. Gradlin is good because most have used it so far, small program that fills the screen. No parameter, no alternate modes, just does the trick. esc when you are done: http://xtknight.atothosting.com/tools/gradlin-v0.2-fs.exe. This makes it easy to compare.

3: A real camera picture of the screen displaying a consistent image. Preferably shot in RAW, converted to Tiff/PNG. Show a crop of the darker end


Now here is an example of a well done photo, I bet it was DSLR:

http://members.iinet.net.au/~json/stuff/blue-grey.jpg
http://members.iinet.net.au/~json/stuff/red-green.jpg

 

[B4RK0D3]

Originally posted by: josh6079
My truck makes a clicking noise, and it has some rust forming on it. Oh, and the paint job isn't exactly the same shade everywhere, etc.

Honestly guys, shut up and play. Put your microscopes back in your pockets, forget what name is on the monitor (since some people don't even want to acknowledge the fact that Dell can do some good things), and just start spending more time doing what you wanted to do with a computer. If you don't like what you have, return it and try another monitor until you do.

I don't want to offend anyone, but dam, everyone agrees that every monitor has its issues. So what is the point in nit picking at every little issue on one monitor if you're going to find another issue in another monitor? IMHO, the Dell is probably the one of the best bangs for the bucks in terms of monitors (The VeiwSonic is probably a little better here though since it is 130~150 dollars cheaper).

I know that I shouldn't even bother replying to this. But, I don't know about you, but ~ $500 is a lot of money for me and I think I have a right to expect quality. Especially beyond the quality of displays that are considerably cheaper. Also, I'm a graphic designer. A display is one of the single, most important aspects of working with digital imaging. What if I'm designing a print layout and I can't tell if a transition has steps in it, or if it's the monitor? Also, if I'm working on touching up professionally-shot photography and I see banding in part of an image...I would have no way to determine if it's in the photo, or if it's my ****** display that I have to train my eyes to use? There are just standards now, and this is no small deal.

I think that this thread is very refreshing and helpful. This was even linked to on the DELL.com technical support forums, and is getting attention from moderators there. This is quite a large issue, so go stand a soapbox somewhere else because you obviously don't know what you're talking about.

 

[cmge]

Originally posted by: spronkey
First, cmge - are those actual photos? (if so, what on earth amazing camera do you have !?!??!?!) If so, that looks like some severe banding. The dell is definitely worse, but the samsung doesn't seem to be doing very well either :S If not you'll need to take photos of the screen and not screenshots in your operating system.

my mistake and i do apologize for the confusion... the pictures were screenshots.. ... i will try and some pics later with my digicam.... and ill post it again for some opinions.. ill try it with both AGP and DVI setup and post my test setup as well...

 

[link26]

looks like dell is gonna have a bunch of lcd's at the outlet for a discount.........

 

[spronkey]

Originally posted by: cmge
my mistake and i do apologize for the confusion... the pictures were screenshots.. ... i will try and some pics later with my digicam.... and ill post it again for some opinions.. ill try it with both AGP and DVI setup and post my test setup as well...

If those pictures were screenshots, i'm now concerned about the gradlin program (if that's what you used), because there is clear banding in them that shouldn't be there.

Unless you compressed them too much but it was vertical banding and JPEG from memory ...

Hmm.

 

[Heartbreaker]

I did that test myself. Screencap gradlin to jpg makes a bit of a mess. It is jpg causing the problem. Screen cap to PNG is perfectly smooth.

Gradlin is fine. This is just a very rough test for JPG.

 

[LurkingInNC]

Originally posted by: link26
looks like dell is gonna have a bunch of lcd's at the outlet for a discount.........

Maybe not.

Dell has had a special which includes the 2007WFP monitor for longer than I would have anticipated:
Dimension 5150
From $1,229
Now from $599

Any cynics think they may recognize the problem and are clearing out existing faulty batches of the 2007WFP through this deal?




I dunno...

 

[xtknight]

The way the fullscreen program works is that it divides the screen width by 256 and rounds down that number (unit pixels must be an integer). That will be the width of each step of color. When it hits the end, if the width isn't evenly divisable by 256, then the last color will just be repeated the rest of the way. I think this is the best way to test it but if you have any other ideas let me know. The program is still governed by the GDI gamma ramp (like every other window on your screen except overlays).

 

[Heartbreaker]

I agree with xtknights method. You want to give each transition some size to make problem visible. Merely doing a 255 pixel wide transition will hide issues because they are only 1 pixel width in size.

The only change I would make is get rid of the stepped bands altogether. That may actually confuse some people. Then you cant just say every band should be a smooth continuous gradient, instead of every second.

Just for fun I would replace stepped bands with RG, RB, BG, combo so continuous bands of:

0-255 Red
0-255 Green
0-255 Blue
0-255 Red-Green
0-255 Red-Blue
0-255 Blue-Green
0-255 Red-Blue-Green (the greyscale you alread have).

These are all the pure tones and all the pure 1:1 mixes.

Note that where you see banding in programs also has it spread out. Areas like sky have vast sections primarily one color then a step.

 

[xtknight]

OK, that should be pretty easy to do. I also have a CMY version which I believe (but am not sure),

1. Is more indicative of real life performance; and,
2. Makes calibration easier because CMY are brighter than RGB; and,
3. In the end, is calibrating the same color components, just that CMY is subtractive.

I'd just be curious if the banding is as bad in CMY mode or if it is actually worse. Just to feed my own curiosity. This doesn't invalidate the RGB scale tests.

(here)

 

[xtknight]

By Red-Green do you mean the gradient starts out at red and ends up greenish, or do you mean both red and green are turned on (yellow) and incremented from 0-255? I assume the latter but I just want to make sure.

 

[Heartbreaker]

I meant the latter. Both Red and Green on, and incremented 0-255. Essentially A combo of your RGB and CYM charts, removing the steps and adding in the CYM. All continuous.

I don't know which is worse. I wont get my banding 2007 till next week (if you can believe Dells arrival date). I suspect it will be similar. CYM looks lighter so maybe it won't be as noticable. OTOH abbertations showing up in one color will probably show up in two under CYM since they are each combos.

 

[xtknight]

Here: http://xtknight.atothosting.com/tools/gradlin-v0.3-mix-fs.exe

 

[Heartbreaker]

Exactly. Simple, all continuous tones. Big enough to spot abberations.

Now you can simply say every row should be smooth continuous tone.

If something gets through this withought banding, I would suspect it would do very well in real images/games etc.

The idea is to have the toughest test case, once you pass that you don't have to think about it anymore.

 

[Ryan Norton]

So what's the 2007WFP look like under v0.3? I'll be getting home in 1-2 hrs and will have another ****** pic from my point and shoot, but someone with a nice DSLR should really give it a shot.

 

[imported_greyhound]

For those of us who plan to keep their defective monitors, it would also be nice to hear from people comparing input quality how VGA looks in general applications (not just banding), so we can decide whether to run in VGA or DVI mode as a general matter.

 

[astromoose]

here is the convo i just had with dell chat

5:42:11 PMDeepinder_01110990 I have checked my records for this issue. We are working on this and are developing the Drivers which will resolve this issue.
5:42:34 PMDeepinder_01110990 Also I am providing you a Link for the latest updated driver for the Monitor, which will assist you in resolving this issue.
5:42:45 PMYou display drivers?
5:43:03 PMDeepinder_01110990 Exactly.
5:43:06 PMSystemDeepinder_01110990 pushes page, http://support.dell.com/support/downloa...3372&formatcnt=1&libid=0&fileid=148796
5:43:21 PMDeepinder_01110990 Please follow the Installation Instructions.
5:43:50 PMYou it seems like more of an electronics problem than a driver issue
5:44:55 PMDeepinder_01110990 Eric, I do agree with you.
5:45:16 PMDeepinder_01110990 However, Dell Is doing the need full for this issue.
5:45:38 PMYou need full?
5:46:05 PMDeepinder_01110990 I mean we are doing everything possible to resolve the issue

obviously a low level tech, but at least they are aware of the problem, i specifically asked if they planned to resolve the issue in a future revision and he just said they are doing everything they can. whatever

 

[mpc7488]

I apologize up front for the quality of the pictures, but it is difficult to get a clean shot this close off of a monitor.

Pic 1: Gradlin Image 1

Pic 2: Gradlin Image 2

Full screen: Full Gradlin shot

And Xtknight's version 0.3 of Gradlin: Xtknight 0.3 Gradlin

It's easier to see banding on Xtknight's version, it's visible on the darker 1/3 of the image. I would still classify it as mild. guidryp, I saw the shots you posted from iinet.net.au, those are what I'd call really bad, mine looks nothing like that unless my setup is different.

Open to suggestion on other shots or tests.

*Edit: P.S. I know it's not a test, but here's a screenshot of Guild Wars with a lot of sky in it. At least it's pretty. Guild Wars outside

 

[yacoub]

My 0.3 Gradlin results are slightly more banded than your photo shows. Mine like yours shows up more readily in the darker 1/2 to 1/3 of the screen. I will try to get a picture of it.

 

[yacoub]

Originally posted by: mpc7488
Panel: 2007WFP, Made in Mexico, factory adjustments except phase changed from 73 to 69 (DisplayMate test had much better results at this level, no 'rolling' of pixels).

Then you're not using DVI, because in DVI there are no settings to set like Phase. That's VGA only, in my understanding.

 

[yacoub]

Anyone: What is the "Gain" setting for R,G,B? What does it affect? What part of a panel adjusts this setting? I can't seem to find any sort of change in gradients or color when adjusting Gain.

And what are the Offsets?

 

[mpc7488]

Originally posted by: yacoub

Originally posted by: mpc7488
Panel: 2007WFP, Made in Mexico, factory adjustments except phase changed from 73 to 69 (DisplayMate test had much better results at this level, no 'rolling' of pixels).

Then you're not using DVI, because in DVI there are no settings to set like Phase. That's VGA only, in my understanding.

Correct. The Phase setting is only valid for when I had it set up with VGA. That was only a few hours when I was setting up the Dell box, all screens and such are from DVI.

 

[yacoub]

Well in DVI, I just tried messing with Custom Color settings, and in the Service menu I tried adjusting the PC Normal colors when I had the panel set to PC Normal. Adjusting the PC Normal RGB settings there makes the banded gradients flicker or fluctuate, but never really makes them more or less apparent. What this tells me is that the gradient issue is unrelated to color levels, and is either a hardware issue (maybe some sort of interference on the DVI input on the display) or is related to something I cannot adjust (since I also tried briefly messing with the Gain and Offsets - the only other adjustments - and they didn't seem to have any effect either).

This is not good - it means it's likely a hardware issue that cannot be fixed with settings, but likely requires a physical hardware replacement by Dell.

Hopefully they figure out the problem and recall the monitors and will implement the fix on them. That would be the 'right' thing to do in this situation, since the product is not performing up to reasonable expectations for a brand new product. We shall see though, if they get around to a proper resolution like that.

 

[Heartbreaker]

I just did a quick compare here of one of your shots with one of the more prominent examples take with a better camera (DSLR I assume). Yours is on top. None of the links I provided were mine. I don't have an LCD monitor yet. I sold my 2405, which had no banding.

http://i.pbase.com/o4/04/606404/1/59300729.compare.png

Overall. I see essentially the same thing, though less prominent on yours. Possible different modes used on the monitor and the better camera might play a part as well. Anyway I would hardly consider this proof that the problem doesn't exist.