Histogram differences in different programs

[Randy Stout]

Happy Easter to all:

I have noticed different histogram results when looking at images in different programs. I use Nikon NX2 primarily. I just commented on a nice Northern Shoveller image from Mike Lentz, that the breast looked a bit hot on my monitor, and in NX2 there were def. a few blown pixels in the red channel. It showed up when you used the blown highlight feature. Kim Rollins just replied that he didn't see any in PS, at least on the histogram.

Have others noticed this discrepancy in how different programs read the histograms, or perhaps it is just the case that when only a few pixels are toasted, the spike on the histogram is usually too small to be seen, while it will show up on the blown highlight screen?

Thanks!

Randy

[Axel Hildebrandt]

Not with different programs but when I save a TIFF as JPG, it sometimes happens.

[Harold Davis]

randy, i have had a problem with that with the two nik programs. ViewNX and CNX2 will give you different histograms every now and then. dont know why.

[Dave Taylor]

I read somewhere that some cameras use the preview (or jpeg) produced by the camera as the source for the histogram on the camera lcd. Is it possible that some imaging programs do the same thing, producing a histogram off a rendered preview that is less accurate than the RAW file in order to save time? I don't know.
I guess if it is just a couple of pixels that were showing blown high lights, I wouldn't worry about it too much, they can be toned down or recovered.

[Roger Clark]

You get different histograms with different color spaces and in going from 16-bits to 8-bits/channel. Try changing from adobe RGB to sRGB and you will see a significant change, for example.

[Dave Taylor]

You get different histograms with different color spaces and in going from 16-bits to 8-bits/channel. Try changing from adobe RGB to sRGB and you will see a significant change, for example.

Of course this will also happen if you compare RAW files to jpegs.

[david cramer]

Randy isn't talking about comparing jpgs to RAW or one color profile to another. He's bringing up a comparison of the same image and profile (in this case a jpg in srgb) on different imaging processors, and getting different results. I brought the image he refers to into both CS4 and NX2 and sure enough, NX2 indicates a few blown highlights in the red channel, whereas the image in CS4 has no blown highlights. It's a minor descrepancy, but it is there. Any ideas as to why?

[Roger Clark]

Randy, David,
Do you see any change in shape of the histogram, or just at blown highlight indicator? If just blown highlights, it could be something as simple as the cutoff point for what is blown is different. For example, one program could say only 255 is blown (on an 8-bit scale), and another could say greater than or equal to 254.

[Charles Glatzer]

They will all vary slightly according to the manufactures default preferences. Higher contrast will often cause the highlights to clip prematurely..

Chas