Canyon Conumdrum

[derekfountain]

I'm puzzled by one thing - I tried to copy your curves to see what was going on, and I noticed that I could not get the histograms to be anything like yours (see below), especially the red and green channels - so I guess you did at least 3 passes.  The first to lighten, then RGB, then Lab - is that right?

No, 2 passes, of which 90% of the image correction happened in the first RGB pass. I put a shadow point in the darkest rock at the very left of the image and a neutral on the large whiteish clump of rock in the foreground. I couldn't find a highlight. I then moved the neutral around a bit to find a sweet spot I was happy with - there's no 'right' answer here I don't think. That was the basic correction which got the blue cast out and the reds and yellows looking as I wanted them. On top of that I manually added the slight boost in the RGB curve highlights. From there I had another pass in Lab mode to boost the lightness and saturation.

I can honestly say I hardly ever look at the histogram. I find the direct feedback from the image, together with the stand alone Hue clock, give me all the information I ever use.

[mikemeister_admin]

I am absolutely lost with trying to understand the difference between the histogram representations.

If I follow your curves and then Lab adjust, I can get your rendition Derek, but what ever I tried I could not load an image that gave angthing near your channels.

So either you are opening the image in a colour profile I never use (I tried nought, RGB & sRGB), but surely profiles will make no difference to a histogram?
Or the different versions of CM show different results.

I know some people do not use the histogram at all - is this because it is just so unreliable I wonder?

Here is an image of CM and Reindeer's idea of the blue channel - PS itself was useless as the top of the peaks were out of range!

Mike, Greg - can you throw any light (!) on the problem?

Thanks
Chris

[ggroess]

I think you need a course in exposures!!

I think that is absolutely excellent Derek,

I see how this is going to play now....sure.....i get it....

I love to parse sentences.....LOL

Good job both of you.....
Greg
P.S. the histogram technology is Mikes bag...I'm gonna let him explain it to me as well...

[mikemeister_admin]

re histograms - I did a little more testing and they come out differently depending on the mode of the image before CM is opened!

But still the peaks do not agree with other representations.

I would have thought that it is simple matter to add up the values to get a histgoram, but something else must be going on!

What's the answer Mike?

Chris

[derekfountain]

What's the answer Mike?

Perhaps more to the point, what's the question? :) What are you trying to achieve with the histogram? The only reason I ever look at a histogram in Photoshop is to see whether there's an empty area at one or both ends, which tells me the image isn't using the entire tonal range. I always use the histogram on my DSLR to check exposure, and I use it in RAW conversion, again to check exposure. But once in Photoshop I rarely pay it any attention.

Under the CurveMeister settings, there's an option to turn the histogram on/off. What happens to your workflow if you uncheck that box? It's not a flippant question! What are you getting (or trying to get) from the histogram that I'm missing? (Apart from hassle and trouble?)

[mikemeister_admin]

Ah, if I knew the answer to the question, I'd be well away!

One thing that I look for when sampling points is whether it is in a large or small area and the effect it may have on the rest of the image when I make a change.  I'm still very much a beginner in attempting to correct images and I feel that the more information I can get the more hope I have at making a good adjustment.  I dont know why, but I feel sort of naked without a histogram representation - sort of gives me something to hold on to, rather than a blank canvas.

In Lab mode (my favorite), where the colours exactly start seems, to me, fairly critical as I increase the saturation - I would not want to clip/blow some colours, just as one could by badly adjusting the endpoints of a lightness channel.  In the image we are considering, there is a significant peak in the B channel, that just holds the sky - how would I now that without a histogram?  I (faithfully?) follow Greg's earlier suggestions, which is (always?) to cheat!  One solution to this Lab image challenge is to completely isolate the sky, play with the rest of the image, and then pop it back (modified a bit).  I am quite happy to modify parts of an image to bring out the detail/colour etc that I want to draw the viewer's attention to.  I am a great fan of the early painters, who effectively used unsharp masks and darkening masking to make a painting come alive.  Adjusting colour is only one part of preparing an image.

In the 'good' old days, I was happy to adjust photos before I took them, by changing the light with reflectors etc or using filters.  In this new digital age, I've completely changed how I take pictures - I now attempt to pre-visualise what I'll do with them in PS.  I frequently bracket both exposure and focus, so that I can combine the images to create the effect I want to portray - it is not photography (as I define it), but art-photo?  I dont manipulate the actual image, just the rendition, tones and colours.  The resultant 'image' is sometimes completely different from what was there and what one could have got with a film/print.  I'm really enjoying this art-playing and find that previously I would take an hour to set up and take a shot (I'm into nature at the moment) and that would be it.  Now I still can take a hour, but then spend up to 3+ hours playing with it! A rather nice wild flower took me all day from start to finish, just to get a single print, but it is kind of special and the result would be completely impossible to achieve with film.

Ah, apologies, I seem to have wandered off the subject a bit! - I was a scientist and am not very artistic, so I like working from facts rather than flare, hence my interest in why such a simple thing as a histogram can be so misleading.

[ggroess]

Well, since you opened the topic up a bit....

For me it has always been about the light...My first photography professor taugh me many things..most of all he taught me to "see" his term for it.  I never really got it until I was watching a baseball game on TV and I thought to myself..."jeeze what crappy lighting....they should move the main to the left a bit and then drop the shadows down...."

I had to put down my beer at that point and sit in amazement as I finally got it...

"seeing" to me is the art of understanding the light and the subject to the point that you want to change things like the lighting on a TV baseball game becasue to you it is just not right.  I think in this digital age many photographers are opting out of lighting control since you can make so much happen in the post processing software.  I almost hate to say it but I kind of miss the days of shooting correct exposure so that I get the shot right.  I still work hard on my exposure control, do not get me wrong here...but I know now that I have a safety net if I miss it by more than a little....

I have found that by not forgetting my basic photography, sticking to the lighting, watching the subject, and above all not setteling for average in the camera, I spend less time on the post processing software part of this and more time behind the lens shooting...which is where I really like to be.

As to the histogram...you knew it was out here somewhere....I think it is up to the display process of the software that is showing it off.  One thing I noticed is that the default in PS is to not show the histogram in the curve window.  At first I was lost...then I noticed something I was much more in tune with....the floating sample point that is always in the curve window when I use CM.  I can live without the histogram because I really make no judgement based on it...I cannot do a very good job in PS without the floating sample point because I have little reference as to where the curve needs adjustment. 

All the histogram really is, is representation of the distribution of the pixels by brightness across your image.  You mentally do a histogram when you previsualize the image...do I want this light or dark...do I have enough mid-tones.  The histogram is a tool that we never had until the digital age came along...it is a quick check or for me even a crutch.  As to why they are different??

I'm going with Math as the problem...since the histogram is a representation the math that creates it determines the values displayed...you said it yourself...the Blue was off the scale...whose scale and at what values....PHEW...glad to have that out there...

Let the flames begin....

Greg

[mikemeister_admin]

Ah you a lucky to be able to see how much of a colour is in an image - I just cant do that.  I can manage it with lightness, but not colour. 

RGB colour space is something I just can not visualise, despite printing out a RGB colour wheel, which I keep on my desk. 

Lab colour for some reason is so much more simpler and more in tune with what I can visualise - but I still really have no idea about the amount of green/magenta/blue/yellow there is in an image - so for me I have to push curves and watch the clock to see what is going on.  I can almost predict what will happen with the 2 lab channels, but have no hope with 3 - stupid really, when I worked I could easily hold 5+2 things in my mind at once.

What I am hoping is that by practice and watching what people like you, Greg and Derek, can do something will eventually click - or it could be that my little grey cells are dying too fast!

I entirely agree that light/shadow and contrast can completely make/break an image and if one is lucky enough to find your subject with the correct lighting it is just great - but often one must adjust (flash and/or PS) for what we actually find on the day.  I'm lucky in that I have moved into the closeup/macro world and therefore have more control in the wild (yes there are wild places in the UK!) than in non-macro shots.  Preparation is everything - my most cherished shot was researched 3 months in advance, the weather was perfect on the day and the film could not be developed until 8 months later!! - gone are those days and new horizons are opening.  Using CM and taking the course has really helped me progress, as has these terrible challenges.  Keep them rolling Greg.

[ggroess]

I have to say zog that I too prefer the LAB space..to me it is more like Color Printing in RA4.  You made simple changes that allowed you to verify that you were going the right way and you kept it very simple.  RGB is getting simpler for me as I teach it more and use it for my images.  When I first started in the class I could not use anything else.  Then I discovered LAB...waaa hooooo...It was an eye opener for me.  It made sense...It was responsive and it was pretty simple when you got to it....

Problem is... just like the image that started this string... LAB is not without limits.  That is where you have to be willing to go elsewhere and not beat your head too hard.  It is one of the best things about the CM tool is the fact that you can switch color spaces so easily.

Greg

[derekfountain]

Ah you a lucky to be able to see how much of a colour is in an image - I just cant do that.  I can manage it with lightness, but not colour. 

Are you still using the demo version? The thing you're missing appears to be the standalone hue clock. It's astonishingly useful for seeing how much colour is in an area of an image, and for spotting casts. If my understand is right, and you're trying to use the histogram to 'see' colours, try using the hue clock instead.