carefully. (seriously) and maybe in conjunction with brightness/contrast and color balance adjustment layers. those 3 adjustments work pretty well together for what you’re trying to do.
Several years ago I wrote a tutorial for manually matching colors between two images. It may still provide some useful tips for you.
Scroll down, read the part in red first.
<
http://www.adobeforums.com/webx/.2ccdfae0>
You can measure the colour in the info panel and reproduce that in the Color Picker if it’s a simple image. No good for a photograph though!
Depends, John.
If you’re able to create a reasonable selection via Color Range, feathered masking, etc, then by matching one specific color, you’ll also be correcting for that entire range of hue. It might have been needed anyway…or you use a color filter to correct the whole image, if it’s needed.
As we all know, each image presents its own specific challenges and solutions.
In my experience, no other tool in Photoshop can completely ruin an image so quickly than Hue/Saturation.
The reason for this is that the affected pixels are not feathered in any way. IOW, if one pixel is slightly blue and its neighbors are neutral or slightly yellow, then this single pixel can stand out like a sore thumb after a transformation. This is especially bad in RGB mode, since a shift in hue will also cause a shift in luminance (this doesn’t happen in Lab mode).
What you may very easily end up with is an extremely grainy or patchy image. (btw, the Camera Raw h/s works the same way, so going the TIF > ACR route is no better in this respect).
So I always safeguard by doing Hue/Saturation adjustments on a separate layer set to Color blend mode, and I also gaussian blur that layer a couple of pixels (depending on the subject) prior to performing the adjustment.
Even so, I always approach Hue/Saturation with the greatest care, and as a last resort. It’s a sleeping dog.
If you’re referring to a few points of yellow being visible on a printer, but not a CRT, that’s just a fact of life. Use the info palette in critical situations.
Curvemeister
No, not printer dots, real image pixels.
The reason I got into this is that I had a scanner that would give very good color balance, except for the greens, which would be shifted significantly towards cyan. Using levels or curves would fix that, but mess up the other colors, so I started using hue/saturation.
Then after a while I noticed that some of those pictures looked incredibly grainy, and zooming in (and a little experimenting) revealed the culprit: In a given area, hue/saturation had picked some pixels but not all. It was either/or. Because of the luminance shift in RGB mode, the images looked grainy. I’m talking print reproduction file sizes; 2 to 3000 pixels wide.
The real problem I suppose is color noise, but that is also a fact of life. I suppose what I’m looking for is a way to reduce color noise without introducing colored fringes, or a hue/saturation that goes after color component; not just absolute color.
BTW, this experience is what taught me to never, ever, work directly on the original. Always on a copy or copy layer.
I was referring to the fact that even a CRT cannot distinguish RGB(255,255,255) from RGB(255,255,253), though these colors may be distinguished on a printer. This seemed to be one of the main points in the article you posted the link to in your previous post.
HSB is a noise prone color space. You might have better results tweaking the green half of the a curve in Lab mode. If you want to retain the Hue/Sat step, it is easy to remove chroma noise by blurring the a and b channels.
You might have better results tweaking the green half of the a curve in Lab mode
Good point, and thanks for that tip. I’ve been somewhat reluctant about using curves on the chroma channels in Lab mode, as I’ve found the results very hard to predict. But it may be the best way to fix those greens after some practice, and maybe I could even make an action of it to give me a good starting point. See, I have mostly nature images, and natural-looking foliage and vegetation is essential.
In any case, as an alternative option, I’ll pop over to the ACR forum and see if I can cet some comment from Jeff on the ACR hue/sat controls.
Hi, as far as matching colours of images, I would never use Hue saturation, In fact The hue saturation sliders are really a very crude means of adjustment, even as a layer. I would use the APPLY IMAGE command under Image.
Chris.
Curvemeister,
Just to let you know you kicked me in the right direction re my scanner colors.
If I stop thinking in terms of RGB color space and start thinking in terms of Lab space, everything falls into place. The b channel is in fine shape, but the a channel needs some minor adjustments. It was like opening a dirty window.
Thanks again 🙂
Glad to hear that using Lab helped!
