I bought a new 24" iMac yesterday and.....

[SaSaSushi]

Some see it, some do not...........But it is there, it is the nature of the panels used by Apple.

I think one of the major reasons threads on this topic never die is that someone always has to throw in a flamebait like this one just below pictures of an iMac that clearly does NOT have any gradient at all.

Then idiots like myself never just let it go as I should and we go in circles again.

 

[Leon Kowalski]

I don't know, Leon. I think you and Apple have a difference of opinion on this matter:

Thanks all the same, but I'm inclined to believe the manufacturer's specs;
LG.Philips says a panel should stabilize to within 95% of final brightness
in no more than 2 minutes after power-on....

...if it doesn't, it's BUSTED,

LK

 

[Leon Kowalski]

Extreme left and right edges cropped and put next to each other:

Similar brightness at the extreme edges is a necessary but not sufficient
requirement for uniform brightnesss. The industry-standard metric is the
MAX:MIN brightness ratio -- regardless of screen location. An edge-vs-edge
comparison conceals the most common 24" ALU iMac uniformity problem --
and that silly blue-center-stripe test pattern does the same. For example,
here's a display that passes those tests with flying colors:

LK

 

[SaSaSushi]

Thanks all the same, but I'm inclined to believe the manufacturer's specs;
LG.Philips says a panel should stabilize to within 95% of final brightness
in no more than 2 minutes after power-on....

...if it doesn't, it's BUSTED,

LK

I saw no mention of percentages on the page I linked you to aside from "full brightness" which I will safely assume means 100% not 95%. Thanks for your input though.

 

[SaSaSushi]

Thanks all the same, but I'm inclined to believe the manufacturer's specs;
LG.Philips says a panel should stabilize to within 95% of final brightness
in no more than 2 minutes after power-on....

...if it doesn't, it's BUSTED,

LK

Here is a detailed analysis of the Apple Cinema Displays by a non-Apple color management website in The Netherlands:

http://www.colormanagement.nl/reports/index.php?id=22,0,0,1,0,0

Warmup and stability

The warmup time and color stability remain more or less the same as before for the 20” and the new 30” displays. The display will stabilize after a warmup time of about 90 minutes, which is normal for these types of TFT displays. The 23HD display even stabilizes in about 15 minutes! After the warmup period, the colors remain exactly the same, and will not shift over time (< 1.00 Delta-E (CIE94))!

 

[peejack]

With all this time to be so anal about you computers I wonder if you guys actually do anything productive with then other than try and find 'faults'

 

[SaSaSushi]

With all this time to be so anal about you computers I wonder if you guys actually do anything productive with then other than try and find 'faults'

Well, in my defense I am just wasting all my time in here responding to all the anal nay-sayers.

I am enjoying using my iMac to do so though.

 

[Leon Kowalski]

Here is a detailed analysis of the Apple Cinema Displays by a
non-Apple color management website in The Netherlands:

I have no quarrel with that. For careful/critical color-calibration work,
an extended warm-up period makes sense -- OTOH, I don't believe it's
acceptable to have a noticable gradient for an hour or two every time
the display comes out of "sleep" mode.

LK

 

[czachorski]

Similar brightness at the extreme edges is a necessary but not sufficient
requirement for uniform brightnesss. The industry-standard metric is the
MAX:MIN brightness ratio -- regardless of screen location. An edge-vs-edge
comparison conceals the most common 24" ALU iMac uniformity problem --
and that silly blue-center-stripe test pattern does the same. For example,
here's a display that passes those tests with flying colors:


LK

I can't quite follow what you are trying to say above with your reference to passing some test with flying colors. But I should remind you that the photo you posted (which is from my iMac) was the first one I took and posted, from a proximity of about 2 feet from the screen, which was done before you taught me about the effects of interference when the camera is held at a distance where pixel resolution of the camera matches up with that of the screen. You can very clearly see that interference pattern in the shot you posted in the lower right window.

Here is another one of the same screen with the effect more amplified. So what you have is 3 photos of the same screen, with amazingly different representations of the screen based on how the photo is taken. Fact is, in real life the screen always looks exactly like the photo I posted first in this thread, which is why I posted it. That one was taken after you suggest that I correct an oversatutation problem with the first good photos that I posted, that you thought might be masking any issues.

All this just shows the problems of asking for photographic evidence of anything with the screen. All that matters is that I use the screen every day, and have done all the crazy tests posted here, and have never seen a gradient on this iMac, using the best and only sensors that matter: my own 2 eyes.

 

[stevegmu]

I still don't see it, and I went to the Apple Store yesterday and looked at every iMac they had. Maybe the 'problem' only exists in the 1st batch.

 

[Leon Kowalski]

I can't quite follow what you are trying to say above with your
reference to passing some test with flying colors.

The photo of your screen that I (re)posted, shows a very obvious
backlight non-uniformity such that the center of the screen is far
brighter than either edge. Just compare the white backgrounds of
the left/center/right Finder windows. Of course, when you cover
the center "hotspot" with a wide blue stripe and only compare the
brightness of the extreme left/right edges, the uniformity problem
magically "disappears."

The minor interference (aliasing) in that photo has nothing to do
with it -- aliasing doesn't affect the average brightness over large
areas, it only produces a very localized "ripple" in the average.

The fact that you originally posted that photo as an example of a
"uniform" screen only demonstrates the fact that some folks are
much more sensitive to gradients than others. That's why screen
photos are useful and "mine looks great" testimonials worthless.

LK

 

[czachorski]

The photo of your screen that I (re)posted, shows a very obvious
backlight non-uniformity such that the center of the screen is far
brighter than either edge. Just compare the white backgrounds of
the left/center/right Finder windows. Of course, when you cover
the center "hotspot" with a wide blue stripe and only compare the
brightness of the extreme left/right edges, the uniformity problem
magically "disappears."

The minor interference (aliasing) in that photo has nothing to do
with it -- aliasing doesn't affect the average brightness over large
areas, it only produces a very localized "ripple" in the average.

The fact that you originally posted that photo as an example of a
"uniform" screen only demonstrates the fact that some folks are
much more sensitive to gradients than others. That's why screen
photos are useful and "mine looks great" testimonials worthless.

LK

You keep on running out on those limbs, and I'll just keep cutting them off. The simple solution is just to post a pic of the screen with a solid white background. But before I do, I want to get you on the hook.

The interference problems from using a camera to represent the screen are not uniform across the screen. In the photo you posted, it causes a different pattern in all 3 of the windows on the screen. It's plain as day. In the photo I posted with the severe interference, the effects are drastically varied across the screen, with the bottom and extreme left and right lower corners being made to look much darker by the interference.

This is NOT a matter of "some folks are much more sensitive to gradients than others". There are many, many good screens out there. This ABSOLUTELY IS an issue of people like you continuing to muddy the waters with flawed logic and conclusions, from measurements and photos you know yourself are flawed. When I post an image of a perfect screen on a solid white background, will you shut up?

All of this should be taken in the context of Leon's own 20" white iMac screen, which he has stated is within tolerance and has less than 20% variation. He seems to be hell-bent on saying everyone else's screen is flawed, but seems to have a blind spot for his own screen and measurements. Let's review it for the record:

Leon's screen, claimed by him to be within uniformity tolerances:

Leon's screen, extreme left/right edges together:

My screen, claimed by Leon to have a gradient:

Extreme left / right edge of my screen:

Leon, I think you are going to have to change your signature to:

"I'm so stubborn that a thousand measurements will never convince me I am wrong"

 

[Leon Kowalski]

Leon, I think you are going to have to change your signature to:

"I'm so stubborn that a thousand measurements will never convince me I am wrong"

Maybe you should change your sig to:

"I only use the extreme left and right edges of my screen."

LK

 

[czachorski]

Every modern, moderately-priced digital camera is a perfectly good
luminance meter -- plus an "exposure computer." If you turn off the
"computer" all that remains is the luminance meter. It's dirt-simple
if the camera has manual controls:

1) Set the ISO to a fixed speed.

2) Select aperture priority and choose an f-stop

3) Let the metering system determine exposure time

Luminance is inversely proportional to exposure time -- Physics 101.

http://johnlind.tripod.com/science/scienceexposure.html

...my Canon is more accurate than my old Gossen Luna Pro,

LK

Hey - by the way - I am not following this. From the web site you linked:

Shutter speed controls how long the film is exposed to light. This is the Time Value (TV) portion of the camera's EV setting. Common markings on shutter speed dials found on cameras made since about 1960-1970 are 1, 1/2, 1/4, 1/8, 1/16, 1/32, 1/60, 1/125, 1/250, 1/500, and 1/1000. Some go as high as 1/2000 and 1/4000. These are times in seconds. Many times the markings are just the denominator (60, 125, 250, etc.). How do these change a camera's EV for its exposure setting in terms of the TV?

With these being in graditions of 2x, how do you measure 20%? Or am I not following this correctly???? On my camera, i have the following units:

f2.8
1/30
+/- -1.0
ISO Auto

Do I want to lock everything down, and let the 1/30 type number varying?

 

[Leon Kowalski]

With these being in graditions of 2x, how do you measure 20%?

IF your camera has only 2x steps of shutter speed, you can't do it directly.
(But you could do it by adjusting ISO or aperture and doing a little math.)

My camera has several 15% shutter speed steps (e.g., 1/13 vs. 1/15), so
direct measurements in the 20%-ish range are fairly straightforward --
and it's possible to resolve 10%-ish differences by playing with aperture
and ISO settings in combination with shutter speeds.

Do I want to lock everything down, and let the 1/30 type number varying?

Yes. If you lock down aperture and ISO, 1/60 is twice as bright as 1/30.

LK


"Stand firm in your refusal to remain conscious during algebra.
In real life, I assure you, there is no such thing as algebra."
- Fran Lebowitz

 

[czachorski]

Maybe you should change your sig to:

"I only use the extreme left and right edges of my screen."

LK

I'm having a busy day, but that cracked me up. Next time I have a chance, I would like to try the camera as a light meter, and will take a pic of the white screen.

 

[czachorski]

Let me make sure I understand the contentions around here: are you saying that all alum iMacs have gradients outside of the 20% spec? Are you saying that all of them have the 2.5x brightness ratio? I wouldn't be surprised if some are in the 2.5x range, because I have seen good photos of them here, but I believe that they are the minority. I believe that most would have be less than 20%, including mine.

I played around with the setting a little on the camera, and it appears to go from 1/50th in the middle to 1/40th on the edges on the brightness monitor setting, and 1/25 to 1/20 on the dimmest setting. Amazingly, that is exactly 20% correct? So that is right on spec, no? Which would make perfectly good sense to me that it is the spec, because that is a brightness variation that you can only notice when you put a light meter on it.

Here is a photo:

 

[Leon Kowalski]

Are you saying that all alum iMacs have gradients outside of the 20% spec?

I can't say anything about "all" however I can say:

- I've never seen one that appeared to be nearly as good as 20%.

- I have personally measured five samples; my own ex-iMac and four other
samples at my local Apple store. Mine, with a 2.5:1 ratio (150% difference)
was the best of the lot -- and the worst was 3.5:1 (250% difference).

- I have examined about six others (at a second apple store and a reseller)
and they all appeared similar to mine. (I don't always carry a light meter.)

I believe that most would have be less than 20%, including mine.

Judging by your photo with the 3 Finder windows, yours is far more uniform
than my 2.5:1 disaster -- but the center-to-edge variation is well over 20%.
DigitalColorMeter.app suggests about 35% to 40% center-to-edge difference
in the brightnesss of the white Finder background areas. OTOH, I'd tend to
put more faith in your exposure measurements; DigitalColorMeter is a good
tool, but a jpeg-processed image is less reliable than a direct measurement.

I played around with the setting a little on the camera, and it appears to go from
1/50th in the middle to 1/40th on the edges on the brightness monitor setting,
and 1/25 to 1/20 on the dimmest setting. Amazingly, that is exactly 20% correct?

Not quite. The industry-standard way of calculating and reporting luminance
ratios is MAX/MIN, so 50/40 or 25/20 both work out to 25% non-uniformity.
Not a big deal; either way, your display is far better than my 24" ex-iMac.

So that is right on spec, no?

I'm not sure where you got the idea that 20% was a "spec." I'm not aware
of any 20% specification for displays. I did state that my white iMac had
"no greater than 20%" variation -- but that was just an upper-bound based
on the resolution/accuracy of my measurement, not a specification.

Here is a photo:

There's no information in that photo -- the screen's exposure is "pegged" at
99%-100% red & blue and 97%-ish in green. The camera's auto-exposure
system is workin' fine, it's trying to set an average exposure of "18% gray",
but with a bright white screen and dark background, it has no choice but to
overexpose the screen. Try a "Solid Gray Medium" background -- or meter
the screen only (if your camera has a spot-metering mode).

LK

 

[czachorski]

I gotta tell you, if all we are talking about this whole time is 25% brightness variation, which is only noticeable when you put a light meter on the screen, I gotta believe that 99% of folks would be thrilled with such a screen. If we are talking about 250% variation, I believe we have put the myth to bed that all iMacs have such a defect. (but I believe a small percentage certainly do).

If you got one with 250% variation, take it back immediately. It is a variation from left to right that is more varied than turning the screen brightness from the brightest setting to the lowest setting on my iMac. That would be awful, and I can guarantee you would be startlingly obvious on every day use and you would NOT need a light meter to know its there. If you got one like mine, within about 25% variation, you practically need to take a digital picture or put a light meter on it to know there is any variation - you definitely can not see it in every day use. My guess is that the vast majority of iMacs are like that. I am basing it on the high count of folks in this thread and this thread who are thrilled with their iMacs, compared to the relatively few threads and small trickle of folks posting with gradient issues now.

Now I am seeing your "some are more sensitive than others comment", Leon. I think what we have is a few number of people who are getting really bad ones (250% gradient) posting their problems, and then a small, sensitive, vocal, minority who are very sensitive to a 20-25% brightness variation who are looking at perfectly good iMacs and concluding they all have bad screens, and then posting it on-line and further stirring things up with FUD.

I will personally not believe a single word of those who post "all iMac screens are bad" or "every iMac screen I have seen at the store is bad", unless they can post results measured with the camera tools Leon has posted, showing that all the iMacs have significantly worse than a 20-25% variation in brightness.

Good luck!

 

[tsd]

Weird

Wait, wait. Are there actually people out there that think the 24" gradient issue is just an urban legend? Wow.

I've had two Apple repair technicians come here to do on-site repairs on my iMac because of the screen. They've agreed about the problem (a dark gradient on the right side of the screen), replaced the display, logic board and power supply. The replaced parts looked different from the OEM, but still bad. The OEM display also had dark bands going across the screen horizontally. The replacement display ONLY had the gradient. So, the repair tech went ahead and ordered a new one, which was just installed two days ago. He and I agreed that it looked better, but only when at full brightness. I use a solid grey medium desktop (I'm a graphic designer), and you can really notice the gradient issue on it. It's especially bad when the brightness is down below 50%.

So, now I've got to drive an hour to give my iMac to an Apple genius before they'll replace the whole machine.

I should mention that the first display issue (dark bands) didn't show up until two months of use.

Oh, I should also mention that two other people that haven't heard about these 24"iMac screen issues have seen my gradient and commented on it.

So why no screen shots? I still use a 35mm film camera, and haven't cared enough to prove it to you to take the time to shoot, process, print, scan and post photos. Also, the repairs have happened pretty quickly, so there's not much point in posting pics of a screen that's about to get replaced for the third time.

 

[czachorski]

Here's a better image. I thought I had the exposure turned down on that previous photo by -1, but looking at it this morning it was +1, which is probably why it is so over-saturated.

I've never used DigitalColorMeter before today. What a neat little handy app. I'm not sure how to use it to make brightness comps? When I run over the photo I just posted, the the highest numbers I can get in the center are:

R: 82
G: 84
B: 92

On the right edge (the darkest edge):

R: 66
G: 68
B: 67

If I just divide all those numbers, I get:

R: 1.24
G: 1.23
B: 1.37

Is that how you make the calcs? I can generate a lower value as I approach the menu bar near the top of the screen, but those values are affected by the shadow that OS-X puts across the entire menu bar. So it looks like a 25% - 37% variation from center to edge. That's pretty consistent with what I got with the camera as a light meter. So on my iMac it is not an left to right gradient, but a slight hot spot in the middle.

So let me understand - are the detractors here saying that such performance is a defect? Because I would never notice such a thing in normal use, and I would never send this iMac back or attempt to exchange it with a straight face or without expecting resistance from Apple. As a matter of fact, I would bet that I could put the grey screen on this iMac and send 99% of people in front of it, and no one would be able to tell with their naked eye there is any brightness variation at all. It is so slight, and the human eye's 15% conical focus makes it almost impossible to see unless you take a photo of it and put a colormeter on it, or use a light meter.

BTW - I did the camera-as-a-light-meter test on my 3 year old Apple Cinema Display. It goes from 1/25 in the middle to 1/20th on the edge, for also a 25% variation. What do you know. This is a screen that is as perfect and beautiful as they come, and it has a 25% variation using these test. I think that is telling you something when the iMac performs at a similar level as the ACD.

 

[czachorski]

Wait, wait. Are there actually people out there that think the 24" gradient issue is just an urban legend? Wow.

If you are referring to me, the answer is: NO. The gradient is not an urban legend. A very small portion of iMacs absolutely do have a severe gradient. The urban legend is that all iMacs have it. The overwhelming vast majority are probably less in the 25% brightness variation range, just like mine and the Apple Cinema displays, and probably most other LCD panels in the world.

 

[0sync0]

So let me understand - are the detractors here saying that such performance is a defect? Because I would never notice such a thing in normal use, and I would never send this iMac back or attempt to exchange it with a straight face or without expecting resistance from Apple. As a matter of fact, I would bet that I could put the grey screen on this iMac and send 99% of people in front of it, and no one would be able to tell with their naked eye there is any brightness variation at all.

I agree that it wouldn't be a problem for most people.

Do you do any work on your iMac that involves adjusting luminance levels on photos or graphics?

I haven't seen these screens but I wouldn't want one with 25% luminance variation for my photo editing. I was thinking of getting my first Mac but now I'm rethinking that because of this issue.

 

[czachorski]

I agree that it wouldn't be a problem for most people.

Do you do any work on your iMac that involves adjusting luminance levels on photos or graphics?

I haven't seen these screens but I wouldn't want one with 25% luminance variation for my photo editing. I was thinking of getting my first Mac but now I'm rethinking that because of this issue.

I don't do any luminance adjusting. I do work with a lot of photos and video and can't tell any difference in luminance across them, nor have the 25-30 people who have looked at photos or videos with me on this iMac.

With the Apple Cinema Display testing at 25% as well, I'm guessing that this range is pretty normal amongst LCDs. Perhaps you should broaden you statement to include LCD screens, not just iMacs.

 

[Leon Kowalski]

I gotta tell you, if all we are talking about this whole time is 25% brightness
variation, which is only noticeable when you put a light meter on the screen,
I gotta believe that 99% of folks would be thrilled with such a screen.

What's "acceptable" versus "unacceptable" is largely a matter of personal
opinion. OTOH, the "typical" performance of commodity LCD monitors can
be deduced from the numerous tests/reviews published on sites such as:
Xbitlabs.com, tomshardware.com, tftcentral.co.uk, etc.

Here's Xbitlabs' opinion specifically on the subject of brightness uniformity:

"As for the maximum deviation, the range is more lenient: if it doesn’t
exceed 15% - good, 20% - acceptable, over 20% - poor."

http://www.xbitlabs.com/articles/other/display/lcd-testmethods_4.html#sect0

Read Xbit's test methods article and form your own opinion. They don't say
whether their grading system is based on the range/distribution of results
over large numbers of monitor reviews -- or on subjective "good" vs. "bad"
judgements by actual users.

If you got one with 250% variation, take it back immediately.

I tried. However, "immediately" turned into a 1-month running battle with
Apple tech support -- escalated to Apple engineering. They had (and have)
the photos from my picasaweb galleries plus measurements on four other
24" iMacs in the Apple store's showroom (ranging from 2.5:1 to 3.7:1).

I finally got a 100% refund, but Apple refused to admit that my iMac was
defective; they also refused to reveal their own specs (if any) regarding
acceptable visual quality of iMac displays.

It is a variation from left to right that is more varied than turning the screen
brightness from the brightest setting to the lowest setting on my iMac.

I made exactly that point, in writing, to Apple tech support and engineering,
and documented it with photographs and meter-readings. Apple's (verbatim,
emailed) response was:

"Previous reports of luminescence and brightness non-uniformity
were found to be within the expected range for iMacs."

That was back in September/October, and I haven't bothered to stalk the
Apple stores since. Maybe the newer iMacs are better, maybe not -- but
based on recent posts here and the Apple forums, they certainly haven't
eliminated the problem entirely. Reports of two or three unsuccessful
exchanges are still common.

...illegitimi non carborundum,

LK