Apaflo Loc: Anchorage, Alaska

Which is to say it is the most accurate and versatile light meter, when coupled with the blinking highlight display, that exists.

How that is "not an exposure meter" isn't worth quibbling about.

dar_clicks Loc: Utah

Understanding the use each of these tools to its best advantage has nothing to do with "quibbling."

sodapop Loc: Bel Air, MD

It has great value in the Photoshop curves application

Apaflo Loc: Anchorage, Alaska

But saying it is not an exposure meter does.

fantom Loc: Colorado

The poster you are responding to clearly said he understands histos. He is not the one who posed the original question.

pecohen Loc: Central Maine

I should probably have stated this more precisely:

... I've adjusted exposure so that there is no clipping on the right and then, without making any other changes, in another histogram for exactly that very same unaltered image, I find that the image is indeed clipped on the right - and I'm not sure why.

GENorkus Loc: Washington Twp, Michigan

*(read the 1st page responces)

Robeng Loc: California

Steve,

To me the histo is real important, it tells me where my highlights are and if they are blown out or not. I trust the hiso more than my view finder.

Rob

GENorkus Loc: Washington Twp, Michigan

Let's see if we can figure this out.

Base:
Is the origional image in jpeg or RAW or what?
What post edit program are you talking about?
What camera?

Your message:
1st part.
Origionally you adjusted either in camera or in a post edit program to rid the right side clipping.

2nd part.
I'm assuming your looking at the histogram in a post edit program and it showed right clipping.
,

pecohen Loc: Central Maine

Now you are asking something that challenges my memory a bit. At least the most common instance of this problem that I've encountered has been with using RAW Therapee on a RAW file.

In that program, at least as I have it configured, in the Adjust tab there is a histogram at the top of the window and generally when I edit a RAW file, I first apply the camera/lens distortion adjustments and then adjust the exposure so that the histogram shows the exposure right-adjusted. If I then open the tone-curve tool another histogram opens and it does not show the same pattern.

So as not to be operating entirely from memory, I just tried this on a RAW image that I picked at random - but without that first step of applying the camera/lens corrections first. In this instance, the histogram in the tone-curve tool showed an under-exposed image - one not adjusted far enough to the right. I believe I've seen this same kind inconsistency before when switching from one program to another - for example when saving an image as a tif file in Raw Therapee and then opening it in PSP.

One difference I notice now is that in this example I just tried, the first histogram (not the one in the tone-curve tool) is RGB whereas the tone-curve tool is monochrome.

These tools both work with floating-point data and convert to an 8-bit or 16-bit format when a file is saved. That makes me suspect that they do not bother converting to an 8-bit format just to display a histogram but rather use some ad-hoc algorithm to compute the x-axis. This would account for the different histograms but perhaps there is some other explanation. However, the accurate position of the upper limit of the x-axis is important if the histogram is to be used for exposure adjustment.

GENorkus Loc: Washington Twp, Michigan

You could be correct with the 8 bit 16 bit but I feel it's something else.

I'm guessing but it sounds like a programmer for them didn't update the histogram values to the other b&w section of the program.

Do you follow my train of thought?

Again it may be that they only updated the values for the color section and the b&w is starting over from the origional values based from the origional RAW image.

If that is so, then they should offer a free upgrade because of their oversight.

Just a guess.

Apaflo Loc: Anchorage, Alaska

It doesn't work like that. The displayed histogram is generated on the fly for the displayed image. Change the displayed image, and the histogram automatically shows the effect.

Of course comparing different kinds of histograms, even if the image is exactly the same, does not make sense. A composite histogram, or a luminance histogram, and an RGB histogram will all be different.

But likewise the histogram produced by one variation of demosiacing a RAW file (for example a histogram generated using the embedded JPEG file) and another variation (for example one the user creates by converting the RAW file with any external converter) is necessarily different.

And when an image gets passed from one program to another there are possible changes, for example resulting from bit-depth conversions, that will affect the histogram, but... the most likely difference is merely the way the exact same histogram data is presented! Some programs normalize the vertical scale, others don't. Some histograms scale the horizontal data differently. But regardless of all of that, each different editor program uses a different internal RGB format (for example with a different default gamma). Not to mention that if the image was compressed using lossy compression that will also change it.

In the end the point is still that if you look at histograms for two different images they will be different!

GENorkus Loc: Washington Twp, Michigan

pecohen Loc: Central Maine

Which raises the question of which of the two histograms to use when adjusting exposure to the right?

Notorious T.O.D. Loc: Harrisburg, North Carolina

Since there is no perfect exposure doesn't this ultimately become subjective? You use the histogram as a reference and then make changes to the exposure or other components of the image to get the result that you desire. Not necessarily what someone else would do with the same image at all. To me the histogram is a rather course guide that can give me some guidance.

Best,
Todd Ferguson
Harrisburg, NC