On Saturday 24 April 2004 7:37 am, Fred Miller wrote:

Is there any software that will calibrate a monitor available for Linux, so that you can more closely match what is seen on screen to what is printed?

Thanks,

Fred

Hi Fred I think that the answer will be that even if there is such a utility, you are likely to be disappointed with the outcome, unless you have a really bad printer. Just for fun, I tried in Win NT, years ago, to set the screen background to the same colour as the wall in my office. It was difficult, but on 24 bit colour I got very close, although a difference of 1 or 2 on a colour co-ordinate in 24 bit colour was noticeably wrong or different. Half an hour later, it looked all wrong. At night under artificial light, it was well wrong. I don't think NT or the monitor were drifting either. The explanation which satisfies me is that the monitor emits light at set absolute levels. The wall reflects whatever light it is given as relative levels on all 3 colour coordinates - relative to the colour balance of the incoming light which changes from hour to hour and between direct and indirct natural light and artificial light. Through evolutionary adaptation, we are adapted to varying incoming light and compensate for it, so that we see things which reflect light according mostly to the reflection co-efficients of the reflecting surface. ie, we filter out the variations in incoming light. With an experiment like mine with the wall and the monitor, something must give, and the monitor is perceived to be wrong, when 1] it isn't and 2] there is no way to put it 'right'. all the best Vince

At 03:32 PM 4/24/2004 +0100, Vince Littler wrote:

On Saturday 24 April 2004 7:37 am, Fred Miller wrote:

...

Is there any software that will calibrate a monitor available for Linux, so that you can more closely match what is seen on screen to what is printed?

Thanks,

Fred

Hi Fred

I think that the answer will be that even if there is such a utility, you are likely to be disappointed with the outcome, unless you have a really bad printer.

Just for fun, I tried in Win NT, years ago, to set the screen background to the same colour as the wall in my office. It was difficult, but on 24 bit colour I got very close, although a difference of 1 or 2 on a colour co-ordinate in 24 bit colour was noticeably wrong or different. Half an hour later, it looked all wrong. At night under artificial light, it was well wrong. I don't think NT or the monitor were drifting either.

The explanation which satisfies me is that the monitor emits light at set absolute levels. The wall reflects whatever light it is given as relative levels on all 3 colour coordinates - relative to the colour balance of the incoming light which changes from hour to hour and between direct and indirct natural light and artificial light.

Through evolutionary adaptation, we are adapted to varying incoming light and compensate for it, so that we see things which reflect light according mostly to the reflection co-efficients of the reflecting surface. ie, we filter out the variations in incoming light. With an experiment like mine with the wall and the monitor, something must give, and the monitor is perceived to be wrong, when 1] it isn't and 2] there is no way to put it 'right'.

all the best

Vince

Vince, I must disagree with you. Magazines and books with color are put together with computers (mostly Macs) and for those computers there _is_ adjustable software, which will make the monitor reproduce the source material accurately, and will cause the print engine (whatever it is) to print the same colors, all within some reasonable tolerance. (You probably have to work in a room with invariable lighting.) Such software is far from free--the last time I looked (around 10 years ago) it was in the $500 and up range. I don't know whether it has come down since, or not. Since I am not in the printing business, I never pursued this at all, but there is surely s/w now for Windows which will do essentially the same thing, since so many people are using digital cameras now. I have seen some rather good digital photos reproduced on color printers. Epson seems to have the edge in that regard. However, if you jam up the ink nozzles on an Epson, you may have to throw it out. Whereas, the HP ink-jets replace the nozzles with the ink. BTW, there will always be some error, especially in the monitor, since the monitor runs on an RGB (red/green/blue) system, at least if it's a CRT, and a good modern printing system, such as a print shop or an advanced ink-jet would use, uses CMYK, (cyan/magenta/yellow/black) which doesn't map exactly to RGB, and of course, the monitor has no true black at all. AFAIK, LCD displays also use RGB, but may have a better black due to the nature of the screen. DO NOT try and adjust color with the monitor. Use the software. However, in spite of the objections above, almost all monitors have a "factory default" position, as, I believe Windows does, to get back to square 1 if you goof it up too badly. --doug

On Saturday 24 April 2004 10:32 am, Vince Littler wrote:

On Saturday 24 April 2004 7:37 am, Fred Miller wrote:

...

Is there any software that will calibrate a monitor available for Linux, so that you can more closely match what is seen on screen to what is printed?

Thanks,

Fred

Hi Fred

I think that the answer will be that even if there is such a utility, you are likely to be disappointed with the outcome, unless you have a really bad printer.

No.....printer is fine. One of my lenses is the Canon 24-70L USM. Anyone who knows glass knows the cost of that puppy and the exceptional resolution, color rendition, and contrast. The Epson Stylus Photo 2200 will reproduce the resolution on 13" stock.....course it take Gimp an extraordinary amount of time to process it. I only have dSLR (Canon) now - no other format.

Just for fun, I tried in Win NT, years ago, to set the screen background to

[snip]

filter out the variations in incoming light. With an experiment like mine with the wall and the monitor, something must give, and the monitor is perceived to be wrong, when 1] it isn't and 2] there is no way to put it 'right'.

I don't perceive that we're going to get to 100% accuracy or near it anytime soon, no matter what OS and software. I'm just disappointed that we're not a LOT closer to it in Linux, and I think we should be. Best, Fred -- "The only secure Microsoft software is what's still shrink-wrapped in their warehouse..." (Forno)