Untitled Document

Dan taught a trick in the advanced class last year which I am trying to remember. How to wake up dull vegetation that starts with too little green and perhaps too much yellow. I'm pretty sure it used the channel mixer, but it might have been selective color. It worked when nothing I could do with LAB curves did.

Does this ring a bell for anyone?

Those three day are really like trying to drink from a fire hose. A lot does go in, but a lot doesn't.

John Ruttenberg

___________________________________________________________________________

How to wake up dull vegetation that starts with too little green

Does "look for the weakest channel" in CMYK ring any bells? ;-)

All the best!

Paco Márquez

661 McKinley St.

MIramar

San Juan, PR 00907

787-721-8554

787-587-7384 Cel.

___________________________________________________________________________

John Ruttenberg writes,

On a duplicate layer (not an adjustment layer), let Green=130% Green, -30% Blue. Or, in CMYK, let Magenta=130% Magenta, -30% Yellow.

As long as the subtraction is equal to the addition gray balance won't change. However, under some circumstances it can throw other colors off (like if there is a suntanned person in the image, with much more yellow in her skin than magenta). If this happens, take the document into LAB without flattening it, and use Blend If to exclude (or partially exclude) anything that isn't negative in the A channel.

Dan Margulis

___________________________________________________________________________

Thanks, Dan. That was the trick I was trying to remember and it worked.

--

John Ruttenberg

___________________________________________________________________________

Hello Dan,

When is it logical to correct colors using Channel Mixer versus "doing it by the numbers" using Curves? When in CMYK I usually modify the four channels (if necessary) and that calls for different numbers for different photos, surely your recipe using Channel Mixer cannot suit every situations, so is it intended only for colors that have "too little of one color and not enough of another"?

The fundamental principle behind Channel Mixer is not clear to me, is it easier to maintain a gray balance?

Andre Dumas

___________________________________________________________________________

Andre writes,

Many people use Channel Mixer for various purposes that could also be achieved with other commands. The only application I can think of where it would take much longer to do it any other way is this specific one, trying to make greenery more green without throwing off gray balance.

No, it is only for this specific case. It will not work for other types of image because it involves subtracting one channel from another, that is, lightening the target channel where the source channel is darker. The problem is that this causes the deepest shadows to get lighter faster than the second-deepest shadows do, which kills detail.

In the specific case of natural greenery, the blue channel is ordinarily quite dark with very limited detail. If we use that to subtract it basically lightens the whole area as a block.

If you subtract the same percentage as you add (e.g. Green=+130% green, -30% blue) then neutral areas won't be affected--they start out R=G=B, so adding 30% of one and compensating by subtracting 30% of another won't change anything.

In terms of the philosophy of this move, I see it as being as much of a move away from the art as if we were to change the color of a shirt from blue to red--either way, some unusual action is required to accomplish the change. There are three categories of objects in which we commonly reject the camera's version of color:

1) We remember skies as being bluer and deeper than the camera does;

2) We remember fleshtones as being healthier, not too pale;

3) We remember greenery as being distinctly green, not dull green.

#1 we deal with by blending. As for #2, the difference between the fleshtone we want and the one the camera saw is there, but it's small enough that we can attend to it with curves. In the case of #3, however, we are often looking for a *big* move away from the color the camera saw, so relatively drastic action can be required.

Dan Margulis

___________________________________________________________________________

A few days ago answering this question, Dan Margulis wrote:

I did not understand Where exactly do you "let Green=130% Green", so I played around and would like to know if this was the right way:

1. duplicate coler.

2. goto Channel Mixer;

3. in the Output Channel, select Green;

4. on the Green slider, change 100% to 130%;

5. on the Blue slider, change 0 to -30%.

It seemed to work.

TIA,

Boris Feldblyum

___________________________________________________________________________

Boris writes,

That's the right way.

Dan Margulis

___________________________________________________________________________

Thank you Dan,

Boris

___________________________________________________________________________

Can someone tell me why this move needs to be done on a duplicate layer rather than as an adjustment layer?

Thanks

Dan Dill

___________________________________________________________________________

Nope, no-one can tell you why.

There is no reason.

Cheers,

_andrew webb

---------------------------

WebbWorks Words & Pictures

http://www.webbwork.com

___________________________________________________________________________

Dan Dill writes,

The Channel Mixer move sometimes makes alterations in undesired areas of the image. In that case, the layered file moves into LAB, where the green areas are easily isolated using Blend If sliders, a move that is unavailable in RGB. This can eliminate or tone down the non-green areas, while allowing us to further fine-tune how the green areas are affected (e.g. letting the greenest areas become more saturated faster).

Adjustment layers do not survive the move between colorspaces, so a standard layer is required.

Dan Margulis

___________________________________________________________________________

Well shut my mouth.

I stand corrected.

Cheers,

_andrew webb

---------------------------

WebbWorks Words & Pictures

http://www.webbwork.com

___________________________________________________________________________

Brilliant.

----- Original Message -----

From: Dan Margulis

___________________________________________________________________________

Hello Dan,

I have a question about photos of the greens of nature, the greens that we encounter in a forest, leaves, ferns, mosses etc., either in shadow or lit by a brilliant sun, early spring or summer etc. I'm excluding photos taken with filters or artificial light.

My question: fluorescent greens without any magenta such as L64,a(59),b66 and L82,a(59),b66, or bluish greens also without any magenta L82,a(65),b38 and L68,a(58),b15. Are these really impossible in the above environment? Not that you ever said they were, but I encounter these kind of greens frequently in other people's photos and to my eyes they seem somewhat improbable, so I am wondering ...?

I've checked your writings and I can't find anything anywhere about magenta being an essential ingredient in the greens of nature. You do say that greens that have more cyan than yellow are impossible but we see those in some photographs.

Many thanks,

André Dumas

___________________________________________________________________________

Andre writes,

Yes. "Fluorescent" is a good term. Natural greens do not have this attribute. In fact, we remember greens as being greener than they actually were (this is why it's so common to resort to LAB to pump them up) and even so we never get close to such values. Back when Hexachrome was being promoted heavily the vendor used to pump up the greens to these values to prove how big of a gamut their extra green ink offered. Like so many of the efforts of today's inkjet vendors, all it did was make the subject look radioactive.

"To sum up, when trying to improve sky, the professional thinks of yellow first. To correct faces, we concentrate on cyan, and for plants and other greenery, we focus on magenta."

--PP4E, p. 82

In PP5E, which I hope you have received by now, there are several examples of greens that readers have commented are rather loud, yet they all contain significant magenta, as follows:

The lower right cover shot (Figure 19.5), the green leaves in the foreground are typically 75c25m95y35k; the duller hedge in the background typically 70c40m95y35k.

The prairie shot with yellowish grass by Darren in Figure 4.6, 50c20m90y3k.

The Irish scene of Figure 3.13, grass is typically 45c10m80y.

So, even these bright greens all contain magenta, and most of them black as well. About the only time you would even think of losing all magenta in natural greenery is when printing on newsprint. The dirty paper contaminates the greens just as magenta and black would, so less of these colors are needed. Possibly in the Emerald Isle image of Figure 3.13, you might go to 0m in a very small area on newsprint. But you could not have a large area because the magenta is responsible for detail in the green and without it the grass would look like AstroTurf.

Dan Margulis

___________________________________________________________________________

André,

You're being buried by ProPhoto...........until Dan gets back to you I'll offer this...

When I put your L*A*B numbers in I don't get greens, more like carol and oranges. What do you think I'm doing wrong ???

Greens would have magenta to give the color any dimension. The "M" is the gray component to add depth. Almost all the greens I see in nature have some "M tones to create the light to darker tones.

It's the worst in early morning light. Here you have an added color cast from the rising sun and even more neutral shadows pick up "M". Another place you see "M" in green is all the camo-greens that try to simulate naturally occurring (woodland) greens.

On the other hand if you want to be very, very specific about the hue when the "C" is higher than "Y" you don't have a commonly named green.

Lee Clawson

2/\V/\7 Studio

___________________________________________________________________________

Thanks Dan,

That sentence is the one that is most meaningful AFAIC. In landscape photography details are important for me so magenta in the greens is an essential ingredient. I think that you are also saying that, in nature,greens without any magenta are extremely unlikely (?)

Photographers like Bryan Kramer (sometimes)prefer strikingly beautiful fluorescent greens and that is also very nice. Different purposes need different treatment, it depends on the impression you wish to convey to the onlooker.

André Dumas

___________________________________________________________________________

Thanks Lee,

I like the way you put it, "magenta adds dimension" that's a very good description for the unwanted color (even Dan didn't think of such an elegant term). Regarding the L*A*B* numbers, Dan suggested in his "Canyon book" that negative numbers be enclosed in parenthesis so (59)would mean minus 59 and more commonly, in other books would be written L64,a-59,b66. Could this explain why you had problems with the orangy color?

The discussion re ProPhoto/sRGB is still interesting, I hope to see a complete document on these interchanges (exchanges?) real-soon-now.

André Dumas

___________________________________________________________________________

André writes,

Yes, in fact even in our imagination (which gives us greener greens than the camera records) there would be magenta.

True enough. I think you would have to agree that the greens on the lower right cover of PP5E are somewhat louder than you would expect to find in nature (granted that you did not seen the original scene, but you are allowed to use your experience in forming this view). Nevertheless, IMHO the color is pleasing and clients would be unlikely to object.

That color, vivid as it is, still contains substantial magenta. Could I have gone even farther? Possibly. That's why LAB is so attractive, it lets us decide how big of a whopper to tell about color without damaging detail.

If you go too far, however, and blow out all the magenta, detail vanishes. The piece of paper that you're printing is flat. The only sense that you have that the greenery is three-dimensional is that certain parts of it are more green than others. This illusion is created by adding magenta, which is the direct opponent of green and thus has a profound effect when added to any green areas.

The maximum you can go IMHO is to have some relatively small area go 0m0k, allowing magenta and black to move in quickly in less green areas. Doing otherwise merely makes the greens look radioactive, which almost nobody likes.

Dan Margulis

___________________________________________________________________________

Folks,

Since finishing PP5E, I've been putting together some of my own photo albums so that my wife and I can remember some of the interesting places we've been over the years. I've worked on several hundred images, mostly mine, some taken by others.

This has been an interesting learning experience, because in recent years when I was correcting files it was usually because somebody else had not been successful with them or because in looking at the image I wanted to try out a particular Photoshop technique. This time, the images are more "typical". I'm working on them because I want to make them look better, not because they pose any particular technical challenge. Also, I'm not shooting for perfection--I'm trying to limit each correction to five minutes, although in practice they often take much longer because I correct them several different ways to test various techniques.

Anyhow, I've now found some better ways of doing certain things than described in PP5E. There are three such areas so far. I'll share them with the list but am not yet ready to discuss two of them because I'm still testing them against other alternatives to see exactly how much better they really are. However, I'm ready with the following, which affects almost anyone who shoots nature and is looking for vivid greens.

If the original has pretty good greens to start, particularly if they are differentiated greens (plants of different species) then the best option is still LAB to drive them apart while intensifying them. Problems can develop when the original greens are too brownish, though.

At various times on this list, and in pp. 231-234 of PP5E, I've discussed using Channel Mixer to intensify greens, an idea first suggested to me off-list by Jerry P'Simer. My previous view was that the method could create problems in non-greens and that it should be used primarily as a desperation measure when areas that *should* be green don't measure as green at all, such as the tree in PP5E's Figure 10.10A. It would have been troublesome to take that into LAB because the tree was so neutral that it didn't measure A-negative. I showed how to use Channel Mixer to make it just green enough that we could now move the file into LAB and make a pretty good picture out of it.

It turns out that there are a couple of safeguards that transform this method into something that can be used much more frequently. I now tend to use it as a start in any image where I'd like to put more pop into greens, not just where the greens start out nearly hopeless. I still will use LAB to differentiate greens where necessary afterward. Here goes.

Problem: We often remember greenery as being considerably more saturated than the camera saw. Clients also clearly prefer greener, rather than duller, lawns, trees, and other plant life, regardless of what is found in the original capture.

If the affected area already starts out somewhat green, then we have LAB, Hue/Saturation, and/or Selective Color, among other things, to intensify it. But some "greens" are so desaturated on capture that Photoshop doesn't recognize them as greens at all, but rather dark yellows.

Solution: for maximum flexibility, here's the workflow:

1) Starting with an RGB file with poor greens but no obvious color cast, on a duplicate layer (not an adjustment layer) Image: Adjustments>Channel Mixer. Choose Greens and enter Green +140%, Blue -40%.

Explanation: The procedure of subtracting part of a channel from another is usually not advisable because it costs detail in darker areas (the darkest areas of the channel are subtracted more heavily than the lighter parts). In natural greenery, however, the blue channel is usually almost solid, so subtracting doesn't harm detail in the target channel.

The procedure does not harm gray balance, because in a gray, all three RGB channels have equal values, so adding 40% of the green while subtracting 40% of the blue results in the same color. In a naturally green or dark yellow area, though, the green channel starts out *much* lighter than the blue, so adding 40% of green while subtracting 40% of blue gives a lighter (greener) result in the green channel.

40% is chosen because I have not seen any examples where a higher value was appropriate, and we can always reduce it later. An adjustment layer should not be used because it rules out the possibility of taking the layered document into LAB.

2) After applying the above Channel Mixer command, choose Edit>Fade: Lighten.

Explanation: Without this step or something that does the same thing, blues such as skies tend to go purple because the green channel is getting darker. The Lighten command restricts the effect to areas where the green channel is getting lighter, as they will be in any areas that are greenish in nature.

The Fade command is used instead of changing the layer mode because that command is needed in step 4. You can retain the flexibility of letting the skies go partially purple by adding another layer to the process. However, I have yet to encounter an image where I wanted any part of this effect, so I don't mind eliminating it altogether with the Fade command.

3) Carefully examine the top vs. the bottom layer to see whether there is an undesirable change in any color other than green. The most likely culprit is that certain browns may get too yellow. If you see this happening and consider it a problem, make a mental note of it but move on.

4) Change the top layer's mode to Color and decide whether the image now looks better. If not, cancel the command.

EXPLANATION: The Channel Mixer move makes green areas greener but it can also make them too light. Color mode restores the original darkness while retaining the added greenness. Sometimes this looks better and sometimes not, in which case you return to Normal mode. I'd guess I use Color mode for about 2/3 of my images and Normal mode for the other 13.

5) Because the settings of 140% and -40% are excessive for most images, reduce the opacity of the top layer to taste. If it's a *big* reduction in opacity, you may wish to re-think your Color vs. Normal decision of Step 4.

6) If you decided in Step 3 that colors other than greens are being damaged by the move, use Convert to Profile to convert the document to LAB, *without* checking the Flatten Image box--the two layers must remain separate. Now use layer Blend If options to eliminate the offensive colors.

EXPLANATION: The colors that are being hurt are presumably browns and reds. Such colors are always A-positive, whereas greens are A-negative. You can therefore restore the bottom layer in those colors by bringing the Blend If slider in to exclude anything that's A-positive (more magenta than green).

Dan Margulis

___________________________________________________________________________

Hi Dan.

I have been using the Channel Mixer Command or it's predecessor, Scitex Transcol, developed for the original Scitex systems, for over 20 years. I still find it to be the most powerful CMYK color correction tool for most problem images where Reds, Greens, or Blues are particularly off the mark. Your description for the process you are using to correct greens in a RGB image sounds interesting. I have not yet tried this, but I don't see why it wouldn't work.

I am very comfortable correcting in CMYK using Channel Mixer because of my experience. I do a considerable amount of product correction for advertising and the bulk of what I do is in CMYK and there is no reason to change color spaces. It is powerful because of the black plate and the ability to use it as a lever to control gray balance. There is no such means of leverage in a RGB image (especially in dark areas) aside from the use of blending options, so I still prefer to stay in CMYK. I'll have to admit that I don't use Channel Mixer much on RGB images other than the occasional luminance correction where I can take advantage of the blend modes like Darken or Lighten.

I have never found it to be difficult to control all colors using channel Mixer in CMYK. I also don't consider it to be "desperation" to make radical changes in color when it is necessary. ;) Still, one of my favorite uses for Channel Mixer in CMYK is the removal of severe color casts. As and example, I went back to one of your old books and used the image "bluecastwoman" and did a 2 minute correction. I used a curve correction to boost the black plate for greater leverage. I used a Channel mixer correction in color mode to remove the cast and then a second curve correction to control contrast. I did use a mask to hold back part of the correction in the woman's slacks because the correction was so strong in the yellow plate that it was necessary to prevent over saturation. Her slacks, IMO, appeared to be about the only thing is the original image that was not heavily casted.

I converted the image to RGB for my host site because it does not allow the use of CMYK images...

http: //img.photobucket.com/albums/v460/jerry717/Ch02_womanbluecast.jpg

I presume that what you are not showing us here is the ability to correct Reds and Blues using the Channel mixer command in RGB? Certainly, more difficult. I look forward to seeing what you come up with.

Regards,

Jerry P'Simer

___________________________________________________________________________

Jerry:

Wow! Looks great.

Obviously I have a lot to learn about the Channel mixer.

One thing puzzles me: what do you mean "when Reds, Greens, or Blues are particularly off the mark"? As opposed to when yellows or magentas, etc. are off the mark? As opposed to a general color correction?

Your example is that of a heavy color cast but I wouldn't see how the reds, greens, or blues are particularly off the mark here. However, there's no disputing the quality of the result: very nice indeed.

Thanks,

Ron Kelly

___________________________________________________________________________

On Jan 16, 2007, at 12:11 AM, Ron Kelly wrote:

Pure colors in a CMYK image are somewhat easier to correct using conventional methods. Cyan, magenta or yellow can also be corrected using Channel Mixer as well, but plate blending is usually not necessary. As in Dan's description of correcting greens in RGB, blues and reds can be corrected in the same manner. I am comfortable in doing this in CMYK. However, I have doubts about doing the same in RGB because there is no black plate to use as a fulcrum to control gray balance in the dark end of an image. Greens are easier. As Dan States... "The procedure does not harm gray balance, because in a gray, all three RGB channels have equal values, so adding 40% of the green while subtracting 40% of the blue results in the same color". This is true for RGB but more careful planning is needed to control gray balance in CMYK because of the weakness that exists in cyan. Equal values of CMY do not produce gray.

Cast removal is just an additional use for channel mixer. The example image lacks yellow, hence the blue cast that exists throughout the image. Curves cannot correct something that isn't there to begin with. That is when the addition of value is needed through the use of some sort of plate blend. Channel Mixer works well in CMYK because of it's ability to do several types of blends in all plates simultaneously while using the black plate as a fulcrum to control gray balance and total ink limits in the dark end of the image.

Following is a screen shot of the Channel mixer setting that I used in color blending mode to correct color and produce values that could be curved afterwards. A radical remix... but necessary to produce the desired result. I first boosted the black from it's maximum value of 68% to 87% while holding it's current quarter tone value on a curve (an S shaped curve). The final result is certainly subjective but the point is that I could conceivably make the image look like anything that I want it to look like using this method, providing the colors look believable of course. ;)

http: //img.photobucket.com/albums/v460/jerry717/Mixer.jpg

I followed up by writing a conventional curve to finalize the result and then removing some of the correction through a mask (a path combined with the data of the current black plate) of the woman's slacks. I used three control points using a 3x3 eyedropper setting and corrected strictly by the numbers.

For those interested, I redid the correction as a psd and uploaded it to the colortheory files section in a folder called BlueCastWoman.

regards,

Jerry

___________________________________________________________________________

I wrote a big long introduction yesterday about how thrilled I am to see this forum. I love Dan's books, and think Photoshop LAB Color is the most powerful Photoshop book I've read to date. I am very enthused about being able to post in a forum with as much clout as is assembled here.

My post was, however, rejected because I didn't sign it.

My question was a tangent to the Channel Mixer/green enhancement suggestion. In the aerial image of Downtown Tucson I provided under Files>Tucson06_Downtown, I noticed that my trees were lacking saturation. We sell this imagery and in the desert scenes that we have sometimes the green ends up looking lackluster throughout. So this isn't a one-time fix but something to look forward to for the long haul.

With LAB a thick yellow haze was removed, and the colors were brought to this point. Now I'm trying to create some more green in the trees. The grass at Tucson High School to the north-east is visible, and just saturating the greens with a curve to the A channel makes it look imaginary. This is a fix that we try to apply to large sets of TIFF files in a batch to avoid doing them all by hand.

Any suggestions?

Thanks in advance,

Ben Allen

Landiscor Aerial Information

___________________________________________________________________________

Ben, I'm not a photographer but I think you need to use polarizing filters first, and then Dan's excellent Lab methods.

Laurentiu Todie

___________________________________________________________________________

Ben Allen writes,

The Channel Mixer move won't work with this particular image because it doesn't follow the first instruction in the recipe--there can't be a color cast before applying Channel Mixer. There are not too many known colors in the picture, but an inference can be drawn from the roofs of the buildings. When the Info palette is set to read LAB numbers, most of those roofs are A-negative--more green than they are magenta. For roofs, if anything, we'd expect the opposite. This identifies a greenish cast at least in the light half of the image. Applying the Channel Mixer move would make the cast worse.

Since the cast apparently is only found in the lighter parts of the image it is easiest to correct it with RGB curves. Then, the Channel Mixer move will work. The greener areas, such as the football field, can be controlled by converting to LAB after the Channel Mixer move, and using Blend If sliders that partially exclude the greenest areas of the top layer.

Dan Margulis

___________________________________________________________________________

Dan,

I have been wondering if the Channel Mixer will become obsolete in CS3 and will be replaced by the new black and white command. Have you looked at it yet, and if so, do you have any thoughts on it?

Jane Edwards

___________________________________________________________________________

Jane, I have not looked at the beta yet, but if it is similar to Lightroom, then this would not make the channel mixer 'obsolete' as a tool, although it may become an obsolete method for creating monotone images for some folk, just as the method made famous by Russell Brown became favoured over channel mixer for others.

Channel mixer is more useful for colour work, although not very intuitive and may require blending modes, blend if, layer masks or other things to limit the effect to wanted areas. And of course, no discussion of channel mixer would be complete without noting how apply image and calculations can also do similar and different things and vice versa.

Regards,

Stephen Marsh.

___________________________________________________________________________

Thanks, Stephen,

I didn't know about Russell Brown's technique, but I'm sure I can find it on his website.

A lot of people are hesitant to put beta software on their production machines, but Deke McClelland has done some training videos for PS CS3, which lynda.com has made available for free. The one on the Black and White command movie is 5:39.

http://movielibrary.lynda.com/html/modPage.asp?ID=327

Jane Edwards

___________________________________________________________________________

Jane Edwards writes:

I have not downloaded CS3 and so haven't played around with the command enough to give an opinion. My first take, subject to change, is that it's the most intriguing new feature in the release.

Channel Mixer, Apply Image, and Calculations all seem to me to duplicate one another to a much greater extent than any of them duplicate B&W. The artificial "channel" structure is apparently different, so I suspect that certain images will handle better with it than by using the other three commands, but others will handle worse.

With respect to the current exercise, though, (using Channel Mixer to create greener greens) AFAIK the new B/W command would not be helpful.

Dan Margulis

___________________________________________________________________________

Jerry (or FTM anyone comfortable with Ch.Mxr), is it possible for you to explain your methodology when using channel mixer in general and for colour correction? I've opened the tool but not used it. Any guiding information would be tremendously appreciated.

Howie Nordström

howienordstrom.com

___________________________________________________________________________

Howie, I am not as comfortable with this tool as Jerry, a few years ago we had a long discussion on and offlist about this and it was obvious to me that from his Scitex systems days that Jerry was way ahead of me on this tool and it's use for CMYK colour correction (I was never a true Scitex operator [played a bit] and the Purup prepress system that I used before the Desktop revolution did not have a comparable function). For Jerry to give a useful example, a simple image would be needed and he would have to go step by step through the process, as Jerry's channel mixed production files are too complex for study.

The use of channel mixer for full colour work is highly dependent on the image and the edit in question. Some images have a channel structure (colour appearance and K plate) that makes this command more workable than other images or GCR conversions of the same image appearance (a key point, as Jerry mentions multiple times, the K channel is often used as a lever to affect the CMY channels).

Here is a common example of a channel mixer move that I make on a regular basis, that is not using the K plate as a lever. A digicam image with a critical skintone may be too saturated in magenta and the cyan values too low. A channel mixer approach may be to simply mix in the R channel output a 80%R and 20%G to boost the weak R (cyan) channel. There are other approaches to the same issue, but depending on the tonality of an image, the channel mixer command may produce a more pleasing result than some other methods. In this simple RGB editing case, Apply Image and Calculations could do similar with say darken blend mode using similar source/target channels.

Where CM really comes into it's own for photographic full colour images is in affecting GCR and gray balance in a full colour image. For example, one may wish to have a legacy custom CMYK Heavy GCR separation, but wishes the gray balance to conform to TR001 (SWOP v2) aimpoints. Channel mixer appears to be one of the easiest ways to alter the gray balance to a new condition (if not altering the legacy Custom CMYK LAB ink/stock values beforehand to deliver the required gray balance). Some final touches may be needed with curves for some tonal transitions, but channel mixer should do 95% of the job over the entire tonal range.

The tool is easier to use in CMYK. Although the moves are similar in RGB, there is no K channel, which is a major limitation. Another issue is that the command only goes up to 200%, so lighter tones may not reach their target value and darker tones may be overblown. There are numerous layer masking and blend if approaches to working with CM for colour correction.

Regards,

Stephen Marsh.

___________________________________________________________________________

--- "Jane_ Edwards" wrote:

Here is a direct URL (Russell appears to have done some housekeeping):

http: //www.outbackphoto.com/workshop/photoshop_corner/essay_21/essay.html

http: //www.photoshopforphotographers.com/pscs2/download/movie-07.pdf (page 4 of the PDF)

I am sure you have seen/heard of it before, it uses two Hue/Saturation adjustment layers to do a similar thing as the new command that you are talking of in CS3.

Thanks Jane and I have now downloaded the CS3 beta.

The black and white tool is obviously good for B+W conversions, but it is also very applicable to full colour images too. By blending the adjustment layer in luminosity mode (and or using blend if or masks) and playing with the sliders, one can affect the individual luminosity of various hues in the full colour image - which was a big part of the later stages of Dan's earlier Professional Photoshop books.

Regards,

Stephen Marsh.

___________________________________________________________________________

Howie, I think of the Channel Mixer as a color correction tool for use in the CMYK color space. To understand my methodology would require a bit of background in the way that I think about color. I began my career by serving an apprenticeship as a process artist back in the days when color seps were photographed on a commercial process camera on continuos tone film, corrected using "light" [more commonly known as dry etching as opposed to chemical etching] and then screened out at multiple sizes and screen rulings. The correction process involved masking and blending of channels very similar to the processes used today in the digital realm. This has been the foundation of my experience. Channel [plate] blending is really not a new (exclusive to digital) concept. I am neither a writer nor a teacher, but I will do my best to explain how I visualize the CM command in hopes of answering your question.

In CMYK, I visualize the CM command as a four x four matrix that allows for the mixing of all channels simultaneously based on existing channel structure. The first time that I encountered it, it was immediately intuitive for me because of my prior experience in analog retouching, but way more powerful. Stephen Marsh stated in an earlier post that it is not a very intuitive function and he is correct in regard to the current interface of CM. The following diagram may help you to visualize the CM in the manner in which I see it and hopefully shed some light as to what it is actually doing.

http: //img.photobucket.com/albums/v460/jerry717/matrix.jpg?t=1169327874

The matrix above is representative of the CM before any alterations have been applied. The number 1 represents a 100% value of the current channel. Using the same matrix format, the following is an exact simulation of the CM correction that I applied in the example image that I submitted in my first post.

http: //img.photobucket.com/albums/v460/jerry717/modmatrix.jpg? t=1169327077

The values placed into the various boxes represent the values that are used by each specific channel within each targeted channel to determine the value of each new channel. The Black plate was not altered because it was corrected first using curves. The "Color" blending mode was also used, so altering the black plate in this matrix would have no effect in photoshop. But, as a blending tool in the other channels, black becomes a pivotal force in controlling shadow end gray balance, and in this case, allowed me to set a new value for the total Ink limit to 300% and to also set a new gray balance that eliminated most of the heavy blue cast of the original. Gray balance is also controlled through the precise mixing of the CMY channels. In Dan's example for intensifying greens, the values used would have no impact on gray balance as per his explanation. My main goal here was to significantly alter the current state of gray balance to drastically new values for final curving.

The other points about CM that Stephen made in answering your post are right on the mark. Stephen is also correct in saying that it would probably take some very simple images using a step by step outline of the thought processes used in using the CM command effectively. I use the CM as a strictly "by the numbers" tool based on the current state of any given image. I use the numbers in each channel to modify the existing numbers in other channels to create new values to meet the current goals of the desired correction. I work from the principle that "the best way to correct any image is to use the image itself" The Channel mixer command is perfectly suited to this philosophy. There are so many uses for CM that it would be impossible to go into detail in a single email. Hopefully the above information will at least give you a somewhat better insight into the functionality of the tool in regard to "how" it works.

Regards,

Jerry P'Simer