DPP doesn't have a genuine Exposure control?

Guillaume, after seeing your results I went back to DPP, this time looking at the numerical values at the bottom of the window, instead of the histogram which apparently was misleading me. Comparing the changes in RGB values with the same three patch card used above, at Brightness 0 and Brightness +1.00. The results are (numbers are approximate because of variations between pixels):
Highlight patch: 223 and 251, a change of 12.5%.
Midtone patch: 107 and 160, 50%.
Shadow patch: 30 and 57, 90%.
Comparing my screen captures from yesterday gave similar results.
Three samples isn't enough to determine the shape of the curve (there could be a straight line that passes through those three points or a curve), but it is clear that +1.00 does not double every value.

However, following Tony's advice I ticked the Linear box. The values I obtained became:
Highlight - 80 and 160.
Midtone - 16 and 32.
Shadow - 3 and 6.
Doubled in each case. My conclusion is that the Brightness (Exposure) adjustment is indeed linear EV adjustment, but it is applied before gamma correction, the Picture Styles profile, the contrast curve and conversion to color space - all of which skew and curve the RGB output. (BTW, the numbers above were taken with Standard P.S., Contrast = 0 and sRGB output [which, as you know, has a TRC that is not exactly gamma 2.2]). To test this I put Contrast down to -4 and obtained 21%, 35% and 64%. It would be interesting to try other P.S.s and other spaces, but beyond what I have time for. And really, although I respect your investigations and have read your site with great interest, this is just geeky silliness; we calibrate our monitors and edit by eye until we have an output that looks the way we want it to. What went on backstage is interesting but ultimately irrelevant.

I didn't read much of the thread. Exposure changes the right hand point of the histogram, and stretches the whole thing. Brightness changes the mid point. Blacks changes the left point. The real world raw book linked from the book thread in my sig (via my faq thread) explains it.

Ah, but Tim. we're talking about DPP!

tim wrote in post #12932131
I didn't read much of the thread.

It shows

tzalman wrote in post #12932038
Doubled in each case. My conclusion is that the Brightness (Exposure) adjustment is indeed linear EV adjustment, but it is applied before gamma correction, the Picture Styles profile, the contrast curve and conversion to color space - all of which skew and curve the RGB output.

I agree, it is not about gamma (gamma is conmutative regarding linearity), but the other settings (contrast, bright, saturation,...).

I have just done a test in ACR setting some arbitrary parameters, and changing only exposure by 1EV, and this is the result:

HTTP response: NOT FOUND | MIME changed to 'image/jpeg'

HTTP response: NOT FOUND | MIME changed to 'image/jpeg'

tzalman wrote in post #12932038
this is just geeky silliness; we calibrate our monitors and edit by eye until we have an output that looks the way we want it to. What went on backstage is interesting but ultimately irrelevant

I disagree here. Knowing how a particular setting is affecting your image is the best knowledge to use that control properly (or simply not using it), rather than trying random settings until we get something we like. I now know by certain (no feelings, perceptions or third party opinions here) that the right workflow in ACR is White Balance, next Exposure for white point (knowing it's a genuine exposure control that won't affect colours), then Brightness for overall luminance (knowing it will desaturate and de-contrast highlights), and I would never use Contrast (since it has a fixed turning point located at 128).

But this is the richness of humankind, not everyone having the same needs, thinkings and interpretation about our environment.

Regards

_GUI_ wrote in post #12927378
...is like the old Bright control in Photoshop CS2, a totally unrecommended adjustment since it destroys highlights information and washes the shadows when pushed, and clips the shadows when reduced

Slightly deviating, but I thought CS2 was a version that was past and resolved the destructive Brightness / Contrast control in PS.

_GUI_ wrote in post #12932205
I now know by certain (no feelings, perceptions or third party opinions here) that the right workflow in ACR is White Balance, next Exposure for white point (knowing it's a genuine exposure control that won't affect colours), then Brightness for overall luminance (knowing it will desaturate and de-contrast highlights), and I would never use Contrast (since it has a fixed turning point located at 128).

Is this a fixed order of operations? What happens if you have to make a large shift in exposure (one-third stop or greater) since it is influenced (clipping) by white balance?

Picture North Carolina wrote in post #12932572
Slightly deviating, but I thought CS2 was a version that was past and resolved the destructive Brightness / Contrast control in PS.

These are Bright and Contrast in CS2:

PixelMagic wrote in post #12932606
Is this a fixed order of operations? What happens if you have to make a large shift in exposure (one-third stop or greater) since it is influenced (clipping) by white balance?

It should be if you don't want to go back to repeat previous operations. You first make white balance (that changes exposure on individual channels), then set general Exposure (typically to avoid any clipping), then use Bright to set the desired level of exposure but with the guarantee you won't clip highlights.

If you for instance set Exposure before WB, and next you modify WB, you can get clipping or non maximum exposure. So you will have to adjust Exposure again. If you set WB first (without caring for clipping) and then set Exposure, you will never have to go back to WB.

Regards

One thing that can be said of DPP, it produces some wonderful conversions and its NR capability is very, very good (and automatic when the camera is set up properly). The price is also right...

I respect your knowledge and insights but please note that your statement slightly contradicts what is probably the most cited book on Adobe Camera Raw, "Real World Adobe Camera Raw with Adobe Photoshop CS5."


Page 208 states the following:

Then on page 209 it states:

So I suppose your workflow is predicated on always nailing the exposure precisely. But for people like me its more an iterative rather than fixed process.

_GUI_ wrote in post #12932854
It should be if you don't want to go back to repeat previous operations. You first make white balance (that changes exposure on individual channels), then set general Exposure (typically to avoid any clipping), then use Bright to set the desired level of exposure but with the guarantee you won't clip highlights.

If you for instance set Exposure before WB, and next you modify WB, you can get clipping or non maximum exposure. So you will have to adjust Exposure again. If you set WB first (without caring for clipping) and then set Exposure, you will never have to go back to WB.

Regards

Just one Question.
You've established that DPP doesn't have a genuine exposure adjustment, so then what exactly does the brightness adjustment behave as?

IMHO, you're putting more thought than is necessary into it. Most people are not computer programmers and don't care how their software functions under the hood. They only care that it does work and that it does what they want it to, and that is what our eyes are used for.
Perhaps Canon has a reason the adjustment is labeled Brightness rather than Exposure.

PixelMagic wrote in post #12932945
Page 208 states the following:
Quote:
"Make any necessary big exposure moves. It’s important to do this before setting the white balance, because changes in the Exposure value can have a big effect on the white balance. Other than that, it doesn’t really matter when you set the white balance."

It is possible that due to matrix operations being done in ACR's WB, its WB produces a different result according to exposure, but I never experienced any visible change in WB because of changing exposure, no matter how much I changed exposure. If you can post an example that confirms what the most cited book on ACR says I'd be happy to look at it.

Moreover I'd say that if ACR's WB changes when changing exposure, this is a conceptual flaw in ACR's Exposure correction design. Think about it, why should it change? does a scene's colour temperature change if you shoot it at f/2.8 with respect to shooting it at f/11?.

On the other hand, since WB is itself a change in exposure on individual channels, a particular WB setting may clip some information that we'll bring back by modifying overall exposure afterwards, so it's a waste of time to do Exposure -> WB -> Exposure, when WB -> Exposure is enough.

See here, warming the WB means R channel getting clipped:

tkerr wrote in post #12933142
You've established that DPP doesn't have a genuine exposure adjustment, so then what exactly does the brightness adjustment behave as?

I guess you mean the Brightness adjustment in the RAW tab, with EV spacings. This is what I'm trying to find out, but without an original image where no other parameter is affecting the output (contrast, saturation,...) it's impossible to know. I want to believe it is a proper exposure correction.

tkerr wrote in post #12933142
IMHO, you're putting more thought than is necessary into it. Most people are not computer programmers and don't care how their software functions under the hood. They only care that it does work and that it does what they want it to, and that is what our eyes are used for.
Perhaps Canon has a reason the adjustment is labeled Brightness rather than Exposure.

Fine, now imagine you are shooting a bracketing with different exposures, and want to use DPP to match their exposures and fuse them in layers for HDR. Can you confirm if the resulting images will share exactly the same exposure (as expected from a genuine exposure control), or I'll find some other 'differences' among them?

I don't mean all this is necessary to everyone, I don't even mean it's useful to everyone. I just think it can be useful for some people.

Regards

I'm not afraid to admit that the technical stuff here is a little beyond me. I'm just curious as to the conclusion and what that means to me when I use the brightness adjustment. A less technical explanation would be great.

Thanks on advance.

I don't have the technical expertise either.

In general, I'd say a Linear adjustment moves all tonal values "forward" (or backwards) evenly, until either highlights clip at the RGB values of 255, 255, 255 or blacks clip at 0, 0, 0. In DPP, you get good RGB values only if you use the Linear option, although the Linear option does away with the pre-processing that the software does (which applies a non-linear curve the best I can understand) which makes most of the image, well, very dark.

In the non-linear mode, the curve has already been applied, so the Brightness slider goes along with the non-linear mode. It will be strongest when not dealing with the brightest areas (although you can still get clipping if you push it). I haven't really done any serious testing with shadow areas.

Know, though, that the Raw Brightness control only goes to a +/-2 EV adjustment, so you are limited there to how far you can "push" things -- a highlight would have to already be at +1 EV to clip if pushed by two stops. However, if your image really needs it, then you can go the two stops in Raw and then go to the RGB tab and push things even more. So there you can go to what, at least a four stop upping in exposure. Interestingly enough in the non-linear option, the boost in the RGB histogram does show, but it does not in the Raw histogram. But, for many scenes if you've set Brightness to the max in both tabs, much if not most of your images will likely have blown highlights. And, if you slide your pointers over the bright areas, the RGB reading at the bottom of the window will show the bright readings of the preview from the non-linear setting. But, if you turn on the Linear option. you will likely get a whole different picture -- you will see the linear Raw data after the linear brightness adjustment of four stops or so. Some highlights will likely be clipped, but much of what "lost" in the non-linear conversion is still visible!

And so, what practical value is all this? Hmm, don't know...although it is a good exercise in understanding some of the underlying "stuff", especially with Raw processing. But, I'll tell you, I am still very much in the "whatever works for you camp". There are undoubtedly some understandings that will help me work more efficiently -- for example, the DPP Brightness slider seems to act a lot like the Lightroom/Camera Raw Brightness, in that it (acting in a non-linear way) will "try" to slow down highlight clipping -- well that's fine for me -- Brighten does a pretty good job for what it advertises, brightening a scene, and if it can do that but not push too hard on highlights, I'm happy for that (and I'm happy with my Lightroom Brightness slider for the same reason)!