Thread: Printing Images
View Single Post
  #3  
Old 28-02-2007, 01:49 PM
okiscopey's Avatar
okiscopey (Mike)
Rocky Peak Observatory

okiscopey is offline
 
Join Date: Aug 2006
Location: Kandos NSW
Posts: 536
Leon,

As you may have gathered from ghsmith’s post, to get truly accurate colour between screen and print is not a simple matter. I work in a graphics department alongside techo photographers, but even so we had to get in third-party help to do our workstations and Epson printers, and to teach us how to set Photoshop up to orchestrate the whole thing. We found that any CRT older than about three years is usually not able to be properly calibrated, and there can be graphics card limitations. We had to junk some of our 21-in CRT monitors and replace them with mid-level calibratable LCD’s. Not only that, we had some of our ceiling fluoro tubes replaced with colour matching 5000K ones so we were viewing our prints under the industry standard illumination.

In case you think I’m trying to put you off, it’s more to put the whole thing in context, and also to introduce another workable (but by no means perfect) way of doing things.

If you have a standard test chart (see below), you can print it and then use your image editor's colour controls to try and make the image on the screen look like the print. You can then write down or save these settings and apply them on a temporary basis to any other image just before you do a print. This is what we did for years, but it only works up to a point … if you get the ‘shadows’ right, the ‘highlights’ are off, if you get reds looking right, greens look wrong, and so on, but it’s a lot better than doing nothing.

Just a couple of notes on this approach (apologies if you know some of this already):

1. Even an old CRT will have much bigger range of contrast than the print … you have to allow for this and not expect too much.

2. You need to shield direct light from falling on the screen. Always do your ‘about-to-print’ colour corrections under the same, subdued room illumination.

3. Your test print needs to be viewed in a patch of bright light at the same time as adjusting the monitor (this can be difficult to arrange!). It’ll look very different under fluorescents than tungsten light. It’s your choice, but a mixture of both is probably a good idea. Try to avoid daylight, which is too variable in brightness and colour.

4. Your CRT may be too ‘blue’. If you have any control over the colour temperature of the monitor, try pegging it down from around 8000K to 4000 or 5000 so that the monitor ‘whites’ match the white unprinted area of your nicely-illuminated test print.

5. It’s usually best to turn your monitor’s contrast up to maximum, or near-maximum, and brightness to mid-way. Then stick something over the adjustment controls so they don’t get changed. If they do, your 'saved' Photoshop (or equivalent) settings won't mean anything anymore.

6. let the monitor warm up for at least 20 minutes before doing any colour-critical work. Also, if by some miracle you get a pefect match by this 'back door' method, don't expect it to match in a year's time - the CRT is slowly changing.

7. The printer may be trying to control things, so you may need to look at the printer options to find out what modes are available.

Colour is a fascinating and complex subject, and whilst there’s a lot of science in it, it’s also very subjective. If you colour correct an image before dinner, it’ll look wrong when you come back to it afterwards.

You can become very critical and paranoid about the whole thing! That’s not me of course, but when I see those ‘carrot people’ on friends’ TVs which have the colour turned right up, I feel like throwing something at the screen.

I can e-mail you a standard test chart if you PM me.

If this has helped you, then I'll feel a lot better about the $3.80 postage costs you had to bear for those filters recently!
Reply With Quote