Dell U2413 24" LED Backlit IPS Monitor and Dell U2412M UltraSharp 24" LED Monitor

[Manfred M]

Windows is ahead of Apple in color management. Although CS4 and up can be set for 10 bit output, Apple machines are limited to 8 bit. Windows offers access to the video card for dynamic range and other functions and Win output can be set for 10 bit. Do not use the HDMI output of any machine for color critical work. HDMI is HD-709A color space which is dynamic range limited and has a compromise green primary. Also you should be aware that when converting color spaces downward in Photoshop and probably any other editor, you can only get relative rendering intent because the profiles used do not have the other intents. If you want to see a rendering intent on your monitor you have to use soft proof.

I'm not sure if your information is completely up to date, Richard. In Photoshop CC, I get a choice of Perceptual, Saturation, Relative Colormetric and Absolute Colormetric rendering intents when converting profiles. I've taken a screenshot to show where this setting can be changed.

Also, I'm not sure where you get 10-bit for Windows, my machine is using 32-bit colour settings.

[ajohnw]

If Richard read one of the reviews I posted he would see that the dynamic range limitation on HDMI can be reset to 0-255. Pass on 1023.

John
-

[ajohnw]

Looking more carefully at the HP monitors it seems to be the LP2480zx that definitely has the 30bit panel. I just love the way they show 0-255 banding but then that's just advertising licence. The others - guess there would be a need to ask them but the 24in is a good price and the calibration gear cost from HP is $250 which for that sort of thing is pretty good. I'd guess they are aiming to give NEC pro monitors a hard time.

John
-

[Nicks Pics]

Hi guys

You all are very knowledgeable about this kind of stuff, I must say a bit over my head at times, but I've also learned a lot, I think. It has been really sharing of you all to spend so much time giving me your opinions. Thank you!

[Richard Lundberg]

You can find the settings, Grumpydiver, always have been able to, but the profiles themselves only have relative. The only way to see rendering on a monitor is to soft proof.

Windows can be set to offer 10 bit per channel. 32 bit has no meaning in RGB.
In Mac OS there is a geeky way to Force RGB which may or may not work. In Windows, you can access the video card and set dynamic range.

[ajohnw]

Just a little correction Richard. In computer jargon the total bits allocated for colour is sometimes used to describe a setting not the per channel figure.

I would see that 32bit colour might be 3x8 plus 8 for the alpha channel. On the other hand as I am primarily a Linux man I have no idea if MS application windows become semi transparent when they are moved around. The MS windows I use when away from home on a laptop is too old to have that. It can be useful at times as windows under the one being worked on can be read.

John
-

[Manfred M]

You can find the settings, Grumpydiver, always have been able to, but the profiles themselves only have relative. The only way to see rendering on a monitor is to soft proof.

Richard; so what you are saying; you get the same rendering intent; but regardless of the selection the user makes; Relative is applied in every case?

[ajohnw]

You can find the settings, Grumpydiver, always have been able to, but the profiles themselves only have relative. The only way to see rendering on a monitor is to soft proof.

Windows can be set to offer 10 bit per channel. 32 bit has no meaning in RGB.
In Mac OS there is a geeky way to Force RGB which may or may not work. In Windows, you can access the video card and set dynamic range.

It would take some one who is very colour management conversant to sort some of this out in terms of what it means in practice.

Intents. This might help as I feel Ufraw is a little clearer on the subject



First get the icons along the +/- line out of the way. They allow me to set how certain aspects of the raw file are handled such as digital linear and HSV and highlight handling. The histogram at the top is what is happening to the raw data. The bottom one is the usual histogram.

Output intent. It has the usual settings. These alter the appearance of the raw histogram but as display intent is set don't really have much effect on the image displayed. That happens what ever is set in the display intent except when that is set to disable soft proofing. What is happening in other cases is that the image is being soft proofed to the display.

The display setting also has the usual set of intents. Change these and it will alter the image on the screen. What happens in practice in my experience depends on the shot.

The output bit depth also has it's implications. For instance a 16 bit TIFF or PNG will still display on a 24bit colour system. In principle at the basic levels of colour management any bit depth is fine even floating point although some wouldn't be suitable for the output software unless it can handle them - usually conversion via an ICC file

++++++++++++

Have you tried googling you machine name HDMI 0-255 or such like Richard? That is if you need to which I assume you don't but some one who wants to use HDMI might. There is a utility about for changing nvidia settings.

On your comment earlier about aRGB being needed for web sRGB you sound like you have been sucked in. When producing sRGB images with the software you are using and many other packages all of the raw information is carried with it all of the time. That data is manipulated and adjusted in the "display" gamut what ever it is. In quotes as it might be a display, printer or even a file.

John
-

[Richard Lundberg]

"On your comment earlier about aRGB being needed for web sRGB you sound like you have been sucked in. When producing sRGB images with the software you are using and many other packages all of the raw information is carried with it all of the time. That data is manipulated and adjusted in the "display" gamut what ever it is. In quotes as it might be a display, printer or even a file."
That is not what I said. I don't care what people who work in sRGB do.
I have never used UFRAW, even though I have it, and I have never seen any expert testimony on whether or not its color spaces have all four intents. If so. fine. I may get around to trying it just because of your example, though, so, good for you.

[ajohnw]

I currently hate Ufraw Richard. It's long over due some changes and finding camera profiles for it is getting difficult. so I can't say that I would advise any one to use it. As raw developers go it is also unusual. No highlight or shadow recovery as such it's all done via 2 manually manipulated curves. The first one shapes the raw data into the output gamut is perhaps the best way to put it and small changes can make a lot of difference to the output. This is similar to what Nikon do with some methods of using their cameras. The other is a conventional curve adjustment.

Without a camera profile it uses another method of translating the raw data but this can result in gradation and colour problems that the profile would prevent = more work than is strictly needed. Sometimes a lot.

The reason I posted a shot of it's control panel is because you should find the same sort of arrangement in Adobe Products. It's very difficult to tie down exactly how this area should be used. Should images be soft proofed to the display for instance in a situation like mine where the user space display is profiled, system too actually. Then there is the question of which output intent should be used. Then there is the fact that some profiles do not offer all intents. Info on intents often reads more like it concerned with gamut changes and little else. That more or less sums it up.

You might find these threads interesting - mostly in relation to comments about intents

Prophoto in NIK, sRGB on export. Problem?

Another thread on Relative vs Absolute Colorimetric intent

sRGB v Adobe RGB

This page doesn't relate to your software but has lots of general info and links It comes up googling ufraw info

http://docs.kde.org/development/en/e...-decoding.html

These don't cover everything though. Some packages take an input gamut or raw and convert to prophoto what ever is being displayed. In this case the output is being soft proofed to the display, it might be better to say intented to the display as the intent will make a difference to the result. The reason for using prophoto is that a colour is never lost. It may have gone out of the display gamut but it's being intented back in or clipped etc and that will change the appearance on the display. What matters is if the person doing it likes the results of the intent really but some are more likely to be effective than others.

I asked some software people what display profile should be used in their PP package. Answer the one that you produced when you profiled your display. The reason I asked the question is that it wasn't clear if the package used it's own colour management within it's own window and bypassed the system colour profile. It still isn't clear so one day I will use a colour sniffer to actually find out if things do change in various places. I stick to setting their package to sRGB. Visually I can't see any difference when the image is viewed with other things but differences can be subtle. I might get an incorrect answer because these people are in to image processing not colour management and use specialised software for that which they have had very little to do with, they just use it.

Your profiling - if you search dispcalgui and install it I am sure you will get better figures than you currently have. My Dell dependent on the weather shows delta E's of under one over a huge set of colour patches. Sometimes a couple are a little over 1. All but a couple are very low. The grey scale errors are all tiny. The best starting point is to use the contrast and brightness settings given for your monitor on tftcentral. If you have any problems with the initial calibration screen ask on here and I will suggest a couple of ways of improving it. The aim is to get the 3 bars dead level with very low error figures. The actual profiling is straight forwards and there is help info on it's web site. The software will work with the vast majority of colorimeters. Many have exactly the same hardware in them. There are only a few variations.

The weather - I doubt if my colorimeter is as accurate as the calibration so would expect some small variations.

John
-

[ST1]

If Richard read one of the reviews I posted he would see that the dynamic range limitation on HDMI can be reset to 0-255. Pass on 1023.

John
-

Hi John to save me searching for the review you wrote wrt HDMI do you have a link to it please

[ajohnw]

I don't Peter but it's mentioned in one of the tftcentral reviews I posted earlier in this thread. Afraid I can't remember which one but suspect that it was the one Nick was contemplating buying with a 16x10 aspect display.

John
-

[JR1]

Move away from the norm and check out professional monitors for editing such as EIZO

[ajohnw]

True Jeremy but the real pro range is rather expensive for many people and as they aren't pro's it can be hard to justify spending that amount and the money might better be spent on camera gear. There is also the point why should I. This is the current report on my monitor. I don't think a more affordable Eizo would be any better.

http://filebin.net/yn4hkvp1nu

It can be better but I am trying a slightly darker screen and Dell do cause a small problem with calibration. They seem to have decided that people who adjust r g b shouldn't see numbers like 0-255 so provide 0-100 which in turns means that calibrating for a target brightness and colour temperature is a bit tricky as the adjustment is too coarse.so the easiest answer is to shift the target colour temperature slightly.

I'm not worried about the errors shown as I know from other profiles I wont currently see the difference and the report has a lot more coverage than most give. The colour swatches shown in the report are split in 2 in an attempt to show what went in and what came out. Bad errors are easily seen. The less extended report comes up with better figures. This one is designed to push things some what.

I suspect I will try profiling for my ambient conditions next and look at a number of my images to see if I can detect any difference and try to get an even better calibration. I'm also using an xyz plus lut profile and will probably try a straight ICC file instead. Firefox may not handle xyz profile well as it doesn't check which are r g b and should.

John
-

[Richard Lundberg]

http://ufraw.sourceforge.net/Colors.html
This page says that ufraw uses the same profiles as Adobe. This means that all conversions from a wider gamut to a smaller gamut are in relative colorimetric rendering intent. Unless you install a receiving profile that has other rendering intent tags in the form A2Bn and B2An. ColorThink can inspect profiles for intent tags. The ICC has a Windows profile inspector which will do the same thing. ICC has released a couple of new sRGB v4 profiles with perceptual.

[ajohnw]

I just tried it with an Adobe profile and it just reports it as corrupted. It likes icc file not dcp which is what adobe use. I would need to delve deeper but I don't think intents matter on this file as it's aim is scale the colours from the camera correctly so in real terms there isn't a gamut. I also believe that the correct output intent at this stage is absolute colourmetric. I believe Colin mentioned this at some point but I often fiddle with this as see what it does. The end of one of the threads I posted links too ends on a comment about absolute colourmetric.

Adobe provide several dcp files for all of the cameras that they cover. Usually their own standard plus mimics of some of the camera modes. Another source of profiles is an application called Photivo. On an E-M5 for instance there's was the best I found. The gradation in blacks was excellent as was the whites. You could find them by installing it and looking around in it's file structure.

I'm pretty sure that the software that I linked too for screen calibration will produce a profile that has all intents. But as one of the other links points out elsewhere they may or may not.

I believe there is also a windoze camera profile type kicking about too and it and the dcp files can be converted to icc types.

One thing I do like about ufraw is it's flexibility but many would find the curves aspect difficult. There is some help on that as there are curve packs about for the 1st one going left to right.

Some one else has recently produced a raw converter with more conventional controls. Bed time so I will try and dig a link out tomorrow. Done by a crew that I have a lot of respect for.

John
-

[Nicks Pics]

John, I looked up the downloadable icc profiles from TFT Central. which you have helpfully mentioned, but I don't lnow how to choose between the different ones available for the same monitor, I was expecting the icc profiles, having been made by a calibrator, to be close to "accurate", but seeing that some of the numbers for brightness, contrast, and R, G, and B can be rather dissimilar yet for the same monitor is slightly confusing. If they are different, how can they all be "accurate", or which one would be accurate? Thank you if you have any suggestions about this.

[Richard Lundberg]

<I just tried it with an Adobe profile and it just reports it as corrupted. It likes icc file not dcp which is what adobe use. I would need to delve deeper but I don't think intents matter on this file as it's aim is scale the colours from the camera correctly so in real terms there isn't a gamut. I also believe that the correct output intent at this stage is absolute colourmetric.>
Just tried what? RAW conversion? And what profile inspector did you use?
There's no intent when doing a RAW conversion to ProPhoto.

[ajohnw]

You will only get the best profile for your monitor by using a colorimeter yourself. Of the profiles that are available on tftcentral pick the ones they did. It may be a pack containing sRGB and aRGB. Usually the only reason for changing contrast and brightness settings is during calibration to set target white points and brightness before profile is generated so profiles will vary according to what some one wanted when they profiled and calibrated their monitor. Tftcentral always aim at 6500K, 150 cd per what ever it is and gammas of 2.2. People use all sorts of numbers.

I tried Ufraw on a raw file with an Adobe camera profile. It rejected the profile. Last time I took a camera profile from Photivo is was perfectly happy with that.

The other raw converter I mentioned is here

http://www.mm-log.com/dlraw/download

I don't know what state it's in as haven't installed it myself yet. It will be a piped system. It makes notes of the changes that have been made and runs them each time one is altered. That usually means that the actual final save of an image will take longer than they did in preview. I hope this one will accept adobe profiles. If not I suspect converters are around.

John
-

[Richard Lundberg]

I have inspected the Dell U2413 gamut with several profile inspectors and they all show the gamut volume of the Dell factory aRGB to be larger than the the Adobe RGB 1998 standard profile by about 5 percent or so. Obviously the reason Dell does not brag about this is that a tiny sliver of the Adobe standard is not covered. I can't see it but my guess is it's either on the blue green side or the in the yellow.
The red is clearly much larger. I suppose that's the reason that early complaints about wide gamut had to do with saturated reds.