Okay, so I'm totally lost and have no idea of what's going on with the class. I was just going to quit and forget about CM and just use the tools I've used in the past. I thought about uninstalling CM since I see no logical reason for me to keep it, since I have no idea what to do with it, but I decided instead to ask a bunch of questions. If I am too far behind the class on this, please say so and I will shut up, leave, and never come back.
Okay, so here's some of my questions......how do you know when to use LAB, CMYK, HSB or RGB? This isn't really explained.
Okay, after that, then my next question is.......where do you place your anchor points on the curves, and how do you know which curve to use? I haven't seen this explained either. We are just told to use anchor points. I still feel like where to place them, how many to use, and which direction to pull them is just a guess. No one seems to know the exact place to put them, so it must be just a guess. If so, then I will stop and never use CM again. I am not good at guessing.
Next is......please don't use explanations where the answer says locate a neutral where you have 255, 255, and 255 (or whatever number you are shooting for). That means nothing to me. Any answer that uses numbers without explaining what they are, how to get them, where you place a hue clock to get them, is worthless to me. I need to know specifics.
Then there is the whole concept of hue clocks. Where do you place them? Why do you place them there? What are you looking at when you do place them? What do you adjust to get them to whatever it is that you are shooting for, and why are you shooting for that number?
Don't even mention masks....I have never used them and see no useful reason to use them. If you need to isolate an area that you want to adjust, then isn't it easier to just use the lasso tool to draw around that area and copy-paste it to it's own layer and adjust that? Masks seem like a waste of time.
These are just a few of the many questions I have. Maybe if I understand the answers to these, I can answer the others for myself.....I don't give that much of a chance though. This whole class is so far over my head that I doubt that I'll understand anything. Photoshop's curves were difficult to understand.....CM is vastly more complicated to understand. I was afraid this would be the case, and it seems my fears were well founded.
Greg, I know you are doing the best you can and that you want to help all of us, but I might be too far behind the curve. If my questions are too stupid, then please say so and I will just fade into the woodwork and never bother you again. CM is a very difficult program to understand, and I can understand why it might be hard to explain it to someone like me. I had hoped that CM might be something I could learn to use, but so far, it seems like that will never happen.....CM is just too complicated.
Questions upon questions....
-
sjordan93436
- Posts: 462
- Joined: Sun Jul 11, 2010 4:23 am
- Contact:
I know Greg can do better.
If you want to learn curves, this is a great place to do it. Small class, no inhibitions. The only way to learn is to:
1. Learn some terms
2. Learn some concepts
3. Have some structured learning.
Then curve, curve, curve.
Don't worry about masks now. They are later.
RGB is the standard that most photoshop photographers use and stay in. Our monitors are RGB, your television is RGB.
CMYK is what everything is printed in. Newspapers, magazines, books. Print guys want CMYK.
For many, they stay in one color space - RGB or CMYK.
Why CMYK? That is how printers work. A black letter is easier to print and cheaper with one ink- black (the K or cmyk).
RGB- three lights. Add them up and you have white. Best way to do TV.
LAB- is .... different. It has some interesting attributes and power. Simplest use... It can make something look colorful without being unnatural. Think of a picture postcard. How colorful they can be.
HSB- is more different. At this point, you probably don't need it.
Simple answer. Camera shoots in RGB. Monitor in RGB. Our inkjet printers want to see RGB, so RGB is most important.
CMYK- adds black. If you need black, more black, more outlines, think of cmyk
LAB- easiest way to add color.
Neutrals- a color with no color. Black, white, grey. Easy enough. You can see them. You know them. Or do you. Our eyes can play tricks. Our camera can play tricks. I have a nice white paper sitting next to me. It is under an old style incandescent light. The paper looks white. The picture looks black and gray. BUT they are yellow! The source of light is yellowish. My eyes convert the yellowish to what I KNOW the paper to be.
Everything is fine until I photograph the paper. I open the file and guess what the paper is yellow, But I know it is white so after a while it is whtie to my mind. A scientific instrument would look at my monitor and say "yellow." My mind says "white", I send the white to print.... Yellow.
How do you solve it?
Well. If this is your profession, and you need to print a book with car colors that people buy cars based on your color, then you spend years and thousands of dollars.
Since it is not, you move your mouse of the paper and look at the numbers.
If it is RGB, then the numbers should be equal. If the red is the same level as the green and blue, it is neutral. If you look at the numbers, and the Y is higher, then you have a choice:
print a yellowish print
ask your program to auto fix it
hope that someone else can fix it
OR fix it yourself.
Is the sky blue? or cyan?
Is the greenery dark green or yellow green?
Is the face (this can kill you) too red? too yellow? too green?
If any of these items are off, how do you fix them? Right now, if a photo image comes with a blue overall color. It looks unnatural. You want to fix it. How do you fix it?
(I can suggest some easier, less complicated, LESS accurate ways to do that)
Hue clocks- are simply a graphical representation of those ugly numbers. No more, no less.
If you want to learn curves, this is a great place to do it. Small class, no inhibitions. The only way to learn is to:
1. Learn some terms
2. Learn some concepts
3. Have some structured learning.
Then curve, curve, curve.
Don't worry about masks now. They are later.
RGB is the standard that most photoshop photographers use and stay in. Our monitors are RGB, your television is RGB.
CMYK is what everything is printed in. Newspapers, magazines, books. Print guys want CMYK.
For many, they stay in one color space - RGB or CMYK.
Why CMYK? That is how printers work. A black letter is easier to print and cheaper with one ink- black (the K or cmyk).
RGB- three lights. Add them up and you have white. Best way to do TV.
LAB- is .... different. It has some interesting attributes and power. Simplest use... It can make something look colorful without being unnatural. Think of a picture postcard. How colorful they can be.
HSB- is more different. At this point, you probably don't need it.
Simple answer. Camera shoots in RGB. Monitor in RGB. Our inkjet printers want to see RGB, so RGB is most important.
CMYK- adds black. If you need black, more black, more outlines, think of cmyk
LAB- easiest way to add color.
Neutrals- a color with no color. Black, white, grey. Easy enough. You can see them. You know them. Or do you. Our eyes can play tricks. Our camera can play tricks. I have a nice white paper sitting next to me. It is under an old style incandescent light. The paper looks white. The picture looks black and gray. BUT they are yellow! The source of light is yellowish. My eyes convert the yellowish to what I KNOW the paper to be.
Everything is fine until I photograph the paper. I open the file and guess what the paper is yellow, But I know it is white so after a while it is whtie to my mind. A scientific instrument would look at my monitor and say "yellow." My mind says "white", I send the white to print.... Yellow.
How do you solve it?
Well. If this is your profession, and you need to print a book with car colors that people buy cars based on your color, then you spend years and thousands of dollars.
Since it is not, you move your mouse of the paper and look at the numbers.
If it is RGB, then the numbers should be equal. If the red is the same level as the green and blue, it is neutral. If you look at the numbers, and the Y is higher, then you have a choice:
print a yellowish print
ask your program to auto fix it
hope that someone else can fix it
OR fix it yourself.
Is the sky blue? or cyan?
Is the greenery dark green or yellow green?
Is the face (this can kill you) too red? too yellow? too green?
If any of these items are off, how do you fix them? Right now, if a photo image comes with a blue overall color. It looks unnatural. You want to fix it. How do you fix it?
(I can suggest some easier, less complicated, LESS accurate ways to do that)
Hue clocks- are simply a graphical representation of those ugly numbers. No more, no less.
-
mikemeister_admin
- Posts: 4927
- Joined: Fri Sep 20, 2013 8:29 pm
Hey there Steve! Thanks for replying. I understand the difference between CMYK and RGB.....LAB, not so much. In my first post I meant, how does one know whether to use LAB, RGB or CMYK when making adjustments? How can you look at a picture and say, okay, I need to adjust this in LAB, as opposed to making adjustments in RGB?
With the hue clocks, how do you know where to place them? In one example, a hue clock was placed on the side of a boat in order to adjust the colors in the sky. That made no sense to me.
Then once you've placed a hue clock, what do you adjust to make the clock read whatever it is your trying to get it to read? By the same token, let's say the clock reads R=135.....B=190.....G=175 what does that tell you, and what are you trying to get them to read? You mentioned that to remove a color cast, the clocks should all read the same, so does that mean I should try to get them to all read 135?....190?....or 175? What action do you use to adjust the numbers to get them to read whatever it is that you want them to read?
When I see a curve with one or more anchor points or control points, how do you know where to place them? You can move the cursor over the picture and you can see on the curve where that color is on the curve, but how do you know what part of the picture you need to set the control point on? As you move the cursor over the picture, the "thick" area on the diagonal line on the curve moves up and down the line, so how do you know where to stop the "thick" area on the diagonal line in the curve. It seems that you can move the cursor over the picture and that "thick" area can be anywhere along the diagonal line.
Then, once you've selected some random point along the diagonal line in the curve window, how do you know whether or not you need another control point?....or two more?....or many more? And if however you determine that you need more than just the one control point, how do you know where to place them?
This has been what is confusing to me and why I said that I don't have the basic knowledge to understand the class. When I read some of the other posts, and I see an explanation that says, "Set a highlight hue clock and a shadow hue clock...." I don't know what that means. What's a highlight hue clock and how do you set it and where do you set it? That isn't explained, it just says to set it as if we should already know how to do that. Since I don't and no one else asks, I assume that I am way behind everyone else.
It seems that most explanations are like that.....telling us to do something when I have no idea how to do it. When I see an explanation like that, I stop right there and I can't go any farther because I don't know how to do that before going on to the next step. When this happens on each step, by the time I've read the whole explanation, I am so far behind that it isn't worth bothering to read any more. I haven't understood the first four steps, so why continue reading?
This is what I'm up against trying to understand CM, and except for two or three things that I do understand, the rest comes across as too complex for me. I guess I'm just too slow with all of this since every else seems to understand it. To explain it in minute detail would bore the class, and I didn't want to do that to the others in class, so I figured I'd just lurk and watch, but I figured it can't hurt to ask some questions first to see if I can understand more about CM. If I can't understand CM, then I'll just give up on it and go back to my old methods of editing pictures. I may not get the same results as if I had used CM, but it'll be the best I can do.
With the hue clocks, how do you know where to place them? In one example, a hue clock was placed on the side of a boat in order to adjust the colors in the sky. That made no sense to me.
Then once you've placed a hue clock, what do you adjust to make the clock read whatever it is your trying to get it to read? By the same token, let's say the clock reads R=135.....B=190.....G=175 what does that tell you, and what are you trying to get them to read? You mentioned that to remove a color cast, the clocks should all read the same, so does that mean I should try to get them to all read 135?....190?....or 175? What action do you use to adjust the numbers to get them to read whatever it is that you want them to read?
When I see a curve with one or more anchor points or control points, how do you know where to place them? You can move the cursor over the picture and you can see on the curve where that color is on the curve, but how do you know what part of the picture you need to set the control point on? As you move the cursor over the picture, the "thick" area on the diagonal line on the curve moves up and down the line, so how do you know where to stop the "thick" area on the diagonal line in the curve. It seems that you can move the cursor over the picture and that "thick" area can be anywhere along the diagonal line.
Then, once you've selected some random point along the diagonal line in the curve window, how do you know whether or not you need another control point?....or two more?....or many more? And if however you determine that you need more than just the one control point, how do you know where to place them?
This has been what is confusing to me and why I said that I don't have the basic knowledge to understand the class. When I read some of the other posts, and I see an explanation that says, "Set a highlight hue clock and a shadow hue clock...." I don't know what that means. What's a highlight hue clock and how do you set it and where do you set it? That isn't explained, it just says to set it as if we should already know how to do that. Since I don't and no one else asks, I assume that I am way behind everyone else.
It seems that most explanations are like that.....telling us to do something when I have no idea how to do it. When I see an explanation like that, I stop right there and I can't go any farther because I don't know how to do that before going on to the next step. When this happens on each step, by the time I've read the whole explanation, I am so far behind that it isn't worth bothering to read any more. I haven't understood the first four steps, so why continue reading?
This is what I'm up against trying to understand CM, and except for two or three things that I do understand, the rest comes across as too complex for me. I guess I'm just too slow with all of this since every else seems to understand it. To explain it in minute detail would bore the class, and I didn't want to do that to the others in class, so I figured I'd just lurk and watch, but I figured it can't hurt to ask some questions first to see if I can understand more about CM. If I can't understand CM, then I'll just give up on it and go back to my old methods of editing pictures. I may not get the same results as if I had used CM, but it'll be the best I can do.
Art,
These are very good questions and deserve a solid answer...
Please give me the time to get some things written out for you. If you prefer I can do a private 1 on 1 session with you. I did something similar with Rhonda when I fixed her Pins issue. I think talking might be better than writing but I'll give both a shot. One of the reasons for the shared screen session is the trouble with the typing...as they say a picture is worth a thousand words.
I'll start writing answers and if you could please consider a conference call I am willing to make that happen.
Greg
These are very good questions and deserve a solid answer...
Please give me the time to get some things written out for you. If you prefer I can do a private 1 on 1 session with you. I did something similar with Rhonda when I fixed her Pins issue. I think talking might be better than writing but I'll give both a shot. One of the reasons for the shared screen session is the trouble with the typing...as they say a picture is worth a thousand words.
I'll start writing answers and if you could please consider a conference call I am willing to make that happen.
Greg
-
mikemeister_admin
- Posts: 4927
- Joined: Fri Sep 20, 2013 8:29 pm
Art,
These are very good questions and deserve a solid answer...
Please give me the time to get some things written out for you. If you prefer I can do a private 1 on 1 session with you. I did something similar with Rhonda when I fixed her Pins issue. I think talking might be better than writing but I'll give both a shot. One of the reasons for the shared screen session is the trouble with the typing...as they say a picture is worth a thousand words.
I'll start writing answers and if you could please consider a conference call I am willing to make that happen.
Greg
Hi Greg! Well, there are several reasons why I don't want to do a shared screen session. First is because I don't want to install any software on my computer to do this. I know this is a minor issue, but I don't like installing programs that "share" my computer. Also, I don't know what's involved with a shared session, but I don't have a camera or microphone connected to my computer.
The second reason is the main one.....I have a real problem remembering things. I don't know why this is, but it started some years ago. If you were sitting here in my computer room and spent hours showing me how CM works, I'd forget it by next week. If instead, it's written here in the forums, I can always go back and re-read it to refresh my memory. By refreshing my memory over and over, I can remember things a lot better. If I had a one time session, I wouldn't remember what you taught me, and I'm sure that would frustrate you....and me.
This memory thing is really getting bad, and I don't know what's causing it. I'm only 57 years old, but this has been ongoing for sometime. Heck, my daughter called me and discussed her upcoming surgery. Two weeks later, she called me on the phone to tell me how her surgery went, and I asked her why she hadn't told me she was having surgery. She said that we discussed it two weeks ago, but I don't remember the phone call or the discussion.
If I can't remember something that important, a shared screen session about CM would be probably be forgotten in a few days.....no matter how hard I'd try to remember it.
As for using the written medium to explain things, I know that things can be hard to explain. I don't expect you to spend a lot of time trying to explain things in order to help me to catch up with everyone. I'm sure you have better things to do with your time. What I was hoping was that when you (or anyone else) explains things, to just explain them in more detail so that what's being explained is crystal clear.
Don't say, "Set a hue clock to get a neutral color." Instead say, "We need to set a hue clock so that we can get a neutral color which is gray.(or whatever a neutral color is....I still don't know.) "To set a hue clock we need to xxx." "Okay, now place the hue clocks sample point here. We need it to be here for this reason." Once we have the hue clock set, then we need to adjust xxx so that the numbers in the clock read xxx. We need the numbers to read xxx for this reason."
See how that explanation leaves no room for misinterpretation? To just say "Set a hue clock to get a neutral color" assumes that we already know how to set a hue clock.....we already know where to place the hue clock's sample point, we already know what a neutral color is, we already know what numbers we need to get in the hue clock, and we already know what to adjust to get those numbers. If we don't know just one of those things, we are lost.
Art,
I’m going to try to respond to your questions one at a time and give as much explanation as possible so please bear with me if this gets lengthy.
Your first question, “How do you know when to use LAB, CMYK, HSB or RGB?”, this really is a matter of choice. Many times I'll open image and I'll start looking at it in LAB. The reason I choose LAB Is for me; currently it is the fastest and easiest workspace to do quick image analysis. LAB is an acronym for CIELAB; it is a perceptual color space. This means it is closer to how we think about color than it is to how we print color. Using lab requires a reconversion back to RGB for print or display that's where the program takes care of the color space and converts the lab values to something really nice RGB or RGB.
The Lab color space is most often used for quick colorful solutions to easy imaging problems. I tend to gravitate towards lab because it is so fast. I can adjust the lightness and the contrast of the image without doing anything to the color; it makes it very attractive choice to open up with. In my experience the vast majority of problems with images have more to do with exposure than do with color therefore opening in lab for me is a natural choice.
My choice of color space also depends on the goals I have for the image, if the image is going to be published or used in a book or even printed on my local printer; I'm going to make most of my color corrections in RGB. The reason for this is that RGB is a much more accurate color space. I know this from experience, intuition, and having corrected a lot of images. RGB is also the color space most printing and processors use today.
CMYK on the other hand is used by commercial printers and offset press printing. It is pretty rare that we would use CMYK to send a picture out to say Kodak or snap fish. In Curvemeister we use CMYK since it is a very subtle color correction space and it can solve problems that RGB is incapable of working on. For instance a cyan cast; a true cyan cast is most easily corrected in CMYK since cyan is one of the color channels. In RGB you have to mess around with both the red and green channels to fully correct the cyan cast.
HSB is more of a specialized color space. We use it in Curvemeister to tackle specific problems for instance if I have a portrait of a person and their eyes; only their eyes, appear to be to cyan or blue, I can select with a mask; just their eyes, and reduce the saturation to make them appear much more white without effecting the color of the iris. I am just going to work on the whites of the eyes. We show you HSB in Curvemeister to make you aware of the possibilities and to give you another tool for you to use in your color correction it is by no means a primary color space and most of us would only go in there to solve a specific problem. If we don't show you that it's their how would you ever know.
Your second question Art, “Where do I place anchor points on the curves, and how do I know which curve to use?”, Really comes down to image analysis this is kind of the “art of seeing what's wrong” with the picture. I'm also going to answer your question about hue clocks in combination with this because they are tied together.
I'll ask myself the following questions: what is the brightest highlight?, what is the darkest shadow?, are there any natural neutrals in the image? Is there significant detail in the shadow or highlight? Is that detail important to the image? Is there enough color in the image? Is there a single strong color that I need to make sure to enhance or create multiple tones of? The answers to these questions are what forms the basis of where I place my hue clocks.
I want my hue clocks to give me information to help me repair issues that I've identified by answering those questions. Once I place those hue clocks on the image they dictate where on the curves I am working. For instance my shadow points, the shadow hue clocks are going to show me a numeric range on the curve; where my adjustment is going to make the difference. The same is true for the highlight.
What the hue clocks are telling me is where to adjust the curve. i.e. where to set those anchor points and how far I need to move those anchor points, to change the numeric values so that I achieve the goals I have set for the image. Many times it can seem like the choices that we’re showing you are arbitrary but in fact they're based on answering the questions above.
Since every image is different, it's impossible for me to point out a specific place for you to put hue clocks. I can only tell you that placing them in the deepest shadows with significant detail, brightest highlights with significant detail, and natural neutral gives you the best average and helps you make the easiest corrections to your images. At other times I'll sit hue clocks in specific places that I want to verify the color for instance; I'll set hue clock in the sky to make sure that it's not an impossible color. I might set a hue clock on some green foliage to make sure that it's not too green. I'll also set a hue clock on a face or someplace on a person that is not generally covered by make up this helps me verify that the skin tone is correct and that the person is not too cold looking or to warm looking.
As you can see hue clocks are more than just shadow highlighted neutral they also give me valuable information about various colored parts of the images. They give me a microscopic look at what's going on since the sample is only 3 x 3 pixels or 4x4 pixels or whatever actually I choose to set up to be. The boat example you listed is a use for this. If I mess up the color of the entire image these hue clocks are my canaries in the coal mine. It will tell me quickly that I have a problem.
Adjusting hue clocks is more a matter of setting specific goals; in general I adjust to the hue clocks only to remove a color cast or enhance a known color. For instance, if I have an item in the picture that I know should be gray, I'm going to adjust each individual color channel to make all the numbers equal.
The exact number I pick can be based on multiple different items. Some of the things I consider when choosing that numeric value are: Do I want the image brighter or darker; if I adjust all of the values up to the lightest number then the image is going to get brighter if I adjust all the numbers down to the darkest value then the image is going to get darker.
One of the fastest ways for me to choose the value I want is to take all three numbers, add them together, and divide by three. I get an average of the numbers which I then set all three color channels to be that on the hue clock. This ensures that the overall image tone doesn't change and it allows me to keep the feel in the image that is already there.
In order to understand values presented in the hue clocks you need to remember where those numbers come from the numbers in the hue clocks if you're in the RGB color space are from 0 to 255. These are the range of values possible for any given color in the RGB color space so for instance if you look only at the red channel the darkest black in an image can have a value of zero. What this really means is there is no red in that particular pixel or point on the image. On the other side of value of 255 means that all of the red is possible at that point.
In the RGB color space color is additive what that means is you add all of the color values together to get the hue of the pixels that you're measuring. To me RGB is more like stage lighting you can add colors together to create white; if you add red plus green plus blue and all of the values are 255 you get pure white. If one color is larger than the other two a color cast is created.
Let's look at the values you listed in your question to Steve; 135,190,175 it is a grayish teal. OK, how do I know that…The numbers are fairly close together, that means they are going to be a grayish color but the green and the blue are stronger than the red so it’s somewhere between Green and Blue.
What that specific shade is, is less important than the fact that the color has a lot of green and blue in it. I realize that's subjective, but you need to understand that's where the decision to correct the color comes from.
Do I want the image brighter or darker; if I adjust all of the values up to the lightest number then the image is going to get brighter if I adjust all the numbers down to the darkest value then the image is going to get darker. To take this idea a little bit further, if I know that leaves on a tree are supposed to be green and the red component of the RGB sample point is greater than the green component even by one; even by a difference of one, that I know I need to correct the red channel because there's too much red in the image.
Adjusting the control points is more a matter of personal choice. Some people use the mouse to adjust the curve line up or down, personally when it comes to the fine adjustments you're asking about, I set a control point on the numeric value I want to change, I use the numbers inside the curve window to show me the exact value, set a control point, and I then use the arrow keys up or down or left and right to make small changes so that I do not have a loss of curve control.
As you move the mouse over the picture that thick line that you're talking about in the curves window is what we call the “color worm”. What it's really showing you is how much of the picture is affected by changing the pixel underneath the mouse. This can help you identify where a small change can greatly affect the image.
Regarding multiple control points on the curve, many times I'll make an adjustment to a color and because the adjustment affects more than the area I wanted to affect, I will add additional control points to bring the curve back in line and reduce the area of the image my curve is correcting. How do I know where to place these points? I go back to the color worm and check to see where my adjustment is affecting the image; that usually helps me set the upper and lower boundaries (control points) and helps me pull the curve back so that I just create a little bump.
Art, I'm only going to mention masks in passing. Masks are a tool that we use to protect certain parts of the image and you don't have to use them. They're a very powerful tool that many people who use Photoshop can use to make short work of difficult imaging problems. For instance, you mentioned that you can select your way out of trouble; I'm betting, that if I need to mask something; it affects more than just a single object I'm working on; and if that's the case an overall mask can help me correct that problem across the entire image. Actual mileage may vary.
In general Art, I guess I would ask you to continue to monitor the class and I hope that you choose to stay long enough to participate in the screen share or to work with me to arrange a private session. I certainly understand now why you are reluctant. The program I use to screen share is called Mikogo and it is a small web based program that you run and delete when you are done...It does not install per se. I install it on my PC to share but you don't have to in order to see. The audio portion can be done via telephone but again I understand.
I think that the discussion that occurs and the examples that are used as your questions arise, will clarify a lot of these issues for you and make this a lot simpler process.
Greg
I’m going to try to respond to your questions one at a time and give as much explanation as possible so please bear with me if this gets lengthy.
Your first question, “How do you know when to use LAB, CMYK, HSB or RGB?”, this really is a matter of choice. Many times I'll open image and I'll start looking at it in LAB. The reason I choose LAB Is for me; currently it is the fastest and easiest workspace to do quick image analysis. LAB is an acronym for CIELAB; it is a perceptual color space. This means it is closer to how we think about color than it is to how we print color. Using lab requires a reconversion back to RGB for print or display that's where the program takes care of the color space and converts the lab values to something really nice RGB or RGB.
The Lab color space is most often used for quick colorful solutions to easy imaging problems. I tend to gravitate towards lab because it is so fast. I can adjust the lightness and the contrast of the image without doing anything to the color; it makes it very attractive choice to open up with. In my experience the vast majority of problems with images have more to do with exposure than do with color therefore opening in lab for me is a natural choice.
My choice of color space also depends on the goals I have for the image, if the image is going to be published or used in a book or even printed on my local printer; I'm going to make most of my color corrections in RGB. The reason for this is that RGB is a much more accurate color space. I know this from experience, intuition, and having corrected a lot of images. RGB is also the color space most printing and processors use today.
CMYK on the other hand is used by commercial printers and offset press printing. It is pretty rare that we would use CMYK to send a picture out to say Kodak or snap fish. In Curvemeister we use CMYK since it is a very subtle color correction space and it can solve problems that RGB is incapable of working on. For instance a cyan cast; a true cyan cast is most easily corrected in CMYK since cyan is one of the color channels. In RGB you have to mess around with both the red and green channels to fully correct the cyan cast.
HSB is more of a specialized color space. We use it in Curvemeister to tackle specific problems for instance if I have a portrait of a person and their eyes; only their eyes, appear to be to cyan or blue, I can select with a mask; just their eyes, and reduce the saturation to make them appear much more white without effecting the color of the iris. I am just going to work on the whites of the eyes. We show you HSB in Curvemeister to make you aware of the possibilities and to give you another tool for you to use in your color correction it is by no means a primary color space and most of us would only go in there to solve a specific problem. If we don't show you that it's their how would you ever know.
Your second question Art, “Where do I place anchor points on the curves, and how do I know which curve to use?”, Really comes down to image analysis this is kind of the “art of seeing what's wrong” with the picture. I'm also going to answer your question about hue clocks in combination with this because they are tied together.
I'll ask myself the following questions: what is the brightest highlight?, what is the darkest shadow?, are there any natural neutrals in the image? Is there significant detail in the shadow or highlight? Is that detail important to the image? Is there enough color in the image? Is there a single strong color that I need to make sure to enhance or create multiple tones of? The answers to these questions are what forms the basis of where I place my hue clocks.
I want my hue clocks to give me information to help me repair issues that I've identified by answering those questions. Once I place those hue clocks on the image they dictate where on the curves I am working. For instance my shadow points, the shadow hue clocks are going to show me a numeric range on the curve; where my adjustment is going to make the difference. The same is true for the highlight.
What the hue clocks are telling me is where to adjust the curve. i.e. where to set those anchor points and how far I need to move those anchor points, to change the numeric values so that I achieve the goals I have set for the image. Many times it can seem like the choices that we’re showing you are arbitrary but in fact they're based on answering the questions above.
Since every image is different, it's impossible for me to point out a specific place for you to put hue clocks. I can only tell you that placing them in the deepest shadows with significant detail, brightest highlights with significant detail, and natural neutral gives you the best average and helps you make the easiest corrections to your images. At other times I'll sit hue clocks in specific places that I want to verify the color for instance; I'll set hue clock in the sky to make sure that it's not an impossible color. I might set a hue clock on some green foliage to make sure that it's not too green. I'll also set a hue clock on a face or someplace on a person that is not generally covered by make up this helps me verify that the skin tone is correct and that the person is not too cold looking or to warm looking.
As you can see hue clocks are more than just shadow highlighted neutral they also give me valuable information about various colored parts of the images. They give me a microscopic look at what's going on since the sample is only 3 x 3 pixels or 4x4 pixels or whatever actually I choose to set up to be. The boat example you listed is a use for this. If I mess up the color of the entire image these hue clocks are my canaries in the coal mine. It will tell me quickly that I have a problem.
Adjusting hue clocks is more a matter of setting specific goals; in general I adjust to the hue clocks only to remove a color cast or enhance a known color. For instance, if I have an item in the picture that I know should be gray, I'm going to adjust each individual color channel to make all the numbers equal.
The exact number I pick can be based on multiple different items. Some of the things I consider when choosing that numeric value are: Do I want the image brighter or darker; if I adjust all of the values up to the lightest number then the image is going to get brighter if I adjust all the numbers down to the darkest value then the image is going to get darker.
One of the fastest ways for me to choose the value I want is to take all three numbers, add them together, and divide by three. I get an average of the numbers which I then set all three color channels to be that on the hue clock. This ensures that the overall image tone doesn't change and it allows me to keep the feel in the image that is already there.
In order to understand values presented in the hue clocks you need to remember where those numbers come from the numbers in the hue clocks if you're in the RGB color space are from 0 to 255. These are the range of values possible for any given color in the RGB color space so for instance if you look only at the red channel the darkest black in an image can have a value of zero. What this really means is there is no red in that particular pixel or point on the image. On the other side of value of 255 means that all of the red is possible at that point.
In the RGB color space color is additive what that means is you add all of the color values together to get the hue of the pixels that you're measuring. To me RGB is more like stage lighting you can add colors together to create white; if you add red plus green plus blue and all of the values are 255 you get pure white. If one color is larger than the other two a color cast is created.
Let's look at the values you listed in your question to Steve; 135,190,175 it is a grayish teal. OK, how do I know that…The numbers are fairly close together, that means they are going to be a grayish color but the green and the blue are stronger than the red so it’s somewhere between Green and Blue.
What that specific shade is, is less important than the fact that the color has a lot of green and blue in it. I realize that's subjective, but you need to understand that's where the decision to correct the color comes from.
Do I want the image brighter or darker; if I adjust all of the values up to the lightest number then the image is going to get brighter if I adjust all the numbers down to the darkest value then the image is going to get darker. To take this idea a little bit further, if I know that leaves on a tree are supposed to be green and the red component of the RGB sample point is greater than the green component even by one; even by a difference of one, that I know I need to correct the red channel because there's too much red in the image.
Adjusting the control points is more a matter of personal choice. Some people use the mouse to adjust the curve line up or down, personally when it comes to the fine adjustments you're asking about, I set a control point on the numeric value I want to change, I use the numbers inside the curve window to show me the exact value, set a control point, and I then use the arrow keys up or down or left and right to make small changes so that I do not have a loss of curve control.
As you move the mouse over the picture that thick line that you're talking about in the curves window is what we call the “color worm”. What it's really showing you is how much of the picture is affected by changing the pixel underneath the mouse. This can help you identify where a small change can greatly affect the image.
Regarding multiple control points on the curve, many times I'll make an adjustment to a color and because the adjustment affects more than the area I wanted to affect, I will add additional control points to bring the curve back in line and reduce the area of the image my curve is correcting. How do I know where to place these points? I go back to the color worm and check to see where my adjustment is affecting the image; that usually helps me set the upper and lower boundaries (control points) and helps me pull the curve back so that I just create a little bump.
Art, I'm only going to mention masks in passing. Masks are a tool that we use to protect certain parts of the image and you don't have to use them. They're a very powerful tool that many people who use Photoshop can use to make short work of difficult imaging problems. For instance, you mentioned that you can select your way out of trouble; I'm betting, that if I need to mask something; it affects more than just a single object I'm working on; and if that's the case an overall mask can help me correct that problem across the entire image. Actual mileage may vary.
In general Art, I guess I would ask you to continue to monitor the class and I hope that you choose to stay long enough to participate in the screen share or to work with me to arrange a private session. I certainly understand now why you are reluctant. The program I use to screen share is called Mikogo and it is a small web based program that you run and delete when you are done...It does not install per se. I install it on my PC to share but you don't have to in order to see. The audio portion can be done via telephone but again I understand.
I think that the discussion that occurs and the examples that are used as your questions arise, will clarify a lot of these issues for you and make this a lot simpler process.
Greg
-
mikemeister_admin
- Posts: 4927
- Joined: Fri Sep 20, 2013 8:29 pm
Thanks for the detailed explanations Greg. I didn't want you to waste your Saturday typing out a lengthy explanation for me. I'm sure you have more important things to do. :)
I've read your post several times, and I've tried playing with CM while reading it, and I'm still stuck on several points.
Here is a screen shot of the fruit picture. I have placed three hue clocks on the picture, one on the whitest spot I could find, one on the darkest spot I could find, and one on what looks to be a gray area.
Looking at the numbers on the hue clocks, I don't understand what they mean. In the white hue clock, the values are L = 86.....a = -6....and b = 13. The pointer on the hue clock is towards yellow. What does this information mean? What does an "L" of 86 mean? Does the 86 mean that in the "L" curve, the point that I sampled from has a brightness of 86?
Looking at the L curve while holding the cursor on top of the white sample point, you can see the "worm" is located between 75 and 100. (I assume the worm is there because the "L" is 86, which is between 75 and 100?) Okay, so what do I do now? Do I set a control point somewhere between 75-100....maybe at the 86? If so, then after I set a control point there, what do I do with it?
If that is correct, then does that mean that the "a" value of -6 means that in the "a" curve, the sample point is located on the negative side of zero.....towards the green?
If I have that correct, then what do I do with that information? Do I set a control point at -6 in the "a" curve? If so, what do I do with it?
I've read your post several times, and I've tried playing with CM while reading it, and I'm still stuck on several points.
Here is a screen shot of the fruit picture. I have placed three hue clocks on the picture, one on the whitest spot I could find, one on the darkest spot I could find, and one on what looks to be a gray area.
Looking at the numbers on the hue clocks, I don't understand what they mean. In the white hue clock, the values are L = 86.....a = -6....and b = 13. The pointer on the hue clock is towards yellow. What does this information mean? What does an "L" of 86 mean? Does the 86 mean that in the "L" curve, the point that I sampled from has a brightness of 86?
Looking at the L curve while holding the cursor on top of the white sample point, you can see the "worm" is located between 75 and 100. (I assume the worm is there because the "L" is 86, which is between 75 and 100?) Okay, so what do I do now? Do I set a control point somewhere between 75-100....maybe at the 86? If so, then after I set a control point there, what do I do with it?
If that is correct, then does that mean that the "a" value of -6 means that in the "a" curve, the sample point is located on the negative side of zero.....towards the green?
If I have that correct, then what do I do with that information? Do I set a control point at -6 in the "a" curve? If so, what do I do with it?
- Attachments
-
- fruit01a-jpg (135.31 KiB) Viewed 8957 times
Art,
I use "Dragon Speak" to type for me on the really long answers like yours...It took me about 15 minutes to dictate and correct the posting...
Start by lifting the bottom Left corner of the A channel While you lift it watch the A values of the hue clocks. When they hit Zero you have made the A channel neutral. Then Drop the Top Right corner of the B Channel until the B values of the hue clocks are zero or near zero. you then have made the B channel neutral.
The L values are a brightness scale that goes from 0 to 100. 0 =black 100 = white. everything else is a shade of gray.
DO NOT WORRY THAT YOU DO NOT GET THIS PERFECT...I never do...
This is a great start. You are asking the right questions...the other colors in the image are going to go all over the place but for now just try to make the A and B channels neutral in 2 out of the 3 hue clocks that you have set.
Greg
I use "Dragon Speak" to type for me on the really long answers like yours...It took me about 15 minutes to dictate and correct the posting...
Start by lifting the bottom Left corner of the A channel While you lift it watch the A values of the hue clocks. When they hit Zero you have made the A channel neutral. Then Drop the Top Right corner of the B Channel until the B values of the hue clocks are zero or near zero. you then have made the B channel neutral.
The L values are a brightness scale that goes from 0 to 100. 0 =black 100 = white. everything else is a shade of gray.
DO NOT WORRY THAT YOU DO NOT GET THIS PERFECT...I never do...
This is a great start. You are asking the right questions...the other colors in the image are going to go all over the place but for now just try to make the A and B channels neutral in 2 out of the 3 hue clocks that you have set.
Greg
-
mikemeister_admin
- Posts: 4927
- Joined: Fri Sep 20, 2013 8:29 pm
I can't get two of the hue clocks to read zero. When I have one clock reading zero, the other clock is at -8. Plus, why am I moving the corners of the "a" and "b" channels? Do you always move the corners of those two channels, or do you make adjustments to the centers of them? I'm not clear on why you chose to adjust the corners.
Art,
You are starting at the end of the curve line because the hue clock is not at zero, zero, The "real" neutral is not on the center of the grid so we have to move it. This is where the histogram lies to you...ignore it for now. The hue clock is telling you that as it is now you have to move +8 to get back to zero. For this type of correction you need to start at the ends and then go to the center.
Start with one of the hue clocks then. Make it as close to zero as you can by moving the end of the curve as before. Then go to the center of the grid. Set a point as close to the center as possible and move that point until you have both Clocks neutral for A channel then repeat for the B Channel.
Shot 1 is the B channel and Shot 2 is the A channel. The moves are very small and you might be better using the arrow keys to get the hue clock close.
Shot 3 is the same settings with some saturation slider added and a third hue clock just to check the overall cast.
You are getting there.
Good stuff.
Greg
You are starting at the end of the curve line because the hue clock is not at zero, zero, The "real" neutral is not on the center of the grid so we have to move it. This is where the histogram lies to you...ignore it for now. The hue clock is telling you that as it is now you have to move +8 to get back to zero. For this type of correction you need to start at the ends and then go to the center.
Start with one of the hue clocks then. Make it as close to zero as you can by moving the end of the curve as before. Then go to the center of the grid. Set a point as close to the center as possible and move that point until you have both Clocks neutral for A channel then repeat for the B Channel.
Shot 1 is the B channel and Shot 2 is the A channel. The moves are very small and you might be better using the arrow keys to get the hue clock close.
Shot 3 is the same settings with some saturation slider added and a third hue clock just to check the overall cast.
You are getting there.
Good stuff.
Greg
- Attachments
-
- screenshot001-jpg-53 (175.95 KiB) Viewed 8957 times
-
- screenshot002-jpg-39 (165.32 KiB) Viewed 8957 times
-
- screenshot003-jpg-21 (270.21 KiB) Viewed 8957 times
Return to “Curvemeister 101 September 2010”
Who is online
Users browsing this forum: No registered users and 13 guests