True Blacks and Whites

It seems like for as long as I have been studying photography, there are of course countless topics that get debated over time and time again. One that I just can't seem to wrap my head around, probably because I see so many different opinions on it, is the subject of blacks and whites in an image. So here are a few things I've come across and some questions I have about them:

1) The idea of setting your black and white points to a predefined luminance value, such as 10 and 244, respectively. So just to make sure I have the concept, this is remapping the darkest pixels in the image to 10, and the lightest to 244. I've read that for some this is a regular part of the workflow, and others say that it's outdated and not necessary anymore. What are your thoughts and why?

2) I'm a happy subscriber to Lynda.com, and I watched a video the other day by one of my favorite "instructors" Ben Long - it was on developing pictures for successful printing. In the course, he frequently referenced making sure that you are printing "true blacks and whites," going on to say that when you print true white, the value is 255 and no ink gets applied to the paper in those areas. Is this a desirable thing?

3) I've read that the darkest luminance value that still contains detail is 8, and the lightest is somewhere around 244 - 248. If this is in fact the case, I would assume you would want to have your important pixels where you want detail and are not intentionally clipping to have these values (assuming you are after dynamic range and contrast). If so, this would support some things I've mentioned above and oppose others.

So confused!

Confuse it more by the added discussion of theater enthusiasts who discuss screens which are able to display BTB (blacker than black) and WTW (whiter than white).

drvnbysound wrote in post #16201125
Confuse it more by the added discussion of theater enthusiasts who discuss screens which are able to display BTB (blacker than black) and WTW (whiter than white).

In video, there are such things. Black can fall below 0 IRE and white can extend beyond 100 IRE.

In printing, they don't exist.

RandMan wrote in post #16201028
1) The idea of setting your black and white points to a predefined luminance value, such as 10 and 244, respectively. So just to make sure I have the concept, this is remapping the darkest pixels in the image to 10, and the lightest to 244. I've read that for some this is a regular part of the workflow, and others say that it's outdated and not necessary anymore. What are your thoughts and why?

If you set your black point to a specific value, you are remapping that value, for example 10, to be zero, or black. Everything that is 10 (or below) will be clipped to black at this new setting. The way you are expressing it above is the opposite of what is happening. Often, the idea is to set your black point at, or near, the darkest luminance values to anchor your shadow tones. Same for highlights. If you have highlight information that still contains detail in the 246 range and you set your white point to 244, you will clip (i.e., set to pure white) everything at 244 and above.

Maybe I am misunderstanding what you mean by white and black point in this regard. What do you mean by setting your white and black points to predetermined values?

RandMan wrote in post #16201028
2) I'm a happy subscriber to Lynda.com, and I watched a video the other day by one of my favorite "instructors" Ben Long - it was on developing pictures for successful printing. In the course, he frequently referenced making sure that you are printing "true blacks and whites," going on to say that when you print true white, the value is 255 and no ink gets applied to the paper in those areas. Is this a desirable thing?

You cannot print white, so you are depending upon the "white" of the substrate upon which you are laying ink to give you your white. White varies across papers and the ability to lay down no ink is a function of your printer, printer+ink profile and your image data. Because paper+ink has a dynamic range that is lower than most output devices, optimizing your data for print to use that entire dynamic range (if that is the goal of that print) is something that involved understanding how best to condition your image data for the specific printer+ink+paper and the associated profile. Printing neutral blacks is also a function of your printer and profiles - for example, does your printer have a dedicated black mode, or is black (gray) composed of C,M,Y and K?

RandMan wrote in post #16201028
3) I've read that the darkest luminance value that still contains detail is 8, and the lightest is somewhere around 244 - 248. If this is in fact the case, I would assume you would want to have your important pixels where you want detail and are not intentionally clipping to have these values (assuming you are after dynamic range and contrast). If so, this would support some things I've mentioned above and oppose others.

So confused!

The above depends on many things, including the desired final output. Your display may be able to support detail at a luminance value of 8, but your printer may not and printer detail may suffer at 15 or 20 or whatever.

So, yo need to know your output device and plan for it. Color management and appropriate target values for display calibration and profiling become really important. Viewing a print under appropriate light is the key, if printing is the final output, because this is the reference to which you are attempting to match your display. Some folks will make display profiles for each paper they print on, because the specific paper dictates the white point and contrast ratio (black point) needed to establish the display target values.

kirk

In my opinion, when printing, or when saving for screen viewing, every image needs at least a little black. i.e. something that is O, as well as something that is close to white i.e. 255. However, the significant whites need detail (93% in Lightroom or about 248 rgb.) And also blacks with details (14% in Lightroom or about 10 in rgb.)

That's also about how I did B&W prints in the darkroom with film, paper and chemicals.

And I like a fairly steep tone curve in between. I mostly printed on Contrast IV or above on paper.

Something like this:

FlyingPhotog wrote in post #16201153
In video, there are such things. Black can fall below 0 IRE and white can extend beyond 100 IRE.

In printing, they don't exist.

Sure... I was referring to viewing on screen as opposed to printing.

dmward wrote in post #16203109
In my opinion, when printing, or when saving for screen viewing, every image needs at least a little black. i.e. something that is O, as well as something that is close to white i.e. 255. However, the significant whites need detail (93% in Lightroom or about 248 rgb.) And also blacks with details (14% in Lightroom or about 10 in rgb.)

That's also about how I did B&W prints in the darkroom with film, paper and chemicals.

And I like a fairly steep tone curve in between. I mostly printed on Contrast IV or above on paper.

Something like this:

I'm going to voice dissent, David!

If you meter things with a one degree spotmeter, one might find that the range of inherent brightness levels might span only 3EV range...from brighter than middle grey to white, for example. Let us use the range of EV10-EV13, with exposure set to EV10. So in that type of scene there is no 'black', when you exposed middle grey to record as middle grey.

Similarly, with the black cat in the coal mine, the range might be EV2-EV5, with middle grey reading EV5, so there is no 'white' in the scene.

Now for artistic interpretation of the scenes, one might want to print either photo so that we increase the contrast range of the final print to span a wider range of tones, but that is not portraying anything to the range of inherent brightness levels.

kirkt wrote in post #16201168
If you set your black point to a specific value, you are remapping that value, for example 10, to be zero, or black. Everything that is 10 (or below) will be clipped to black at this new setting. The way you are expressing it above is the opposite of what is happening. Often, the idea is to set your black point at, or near, the darkest luminance values to anchor your shadow tones. Same for highlights. If you have highlight information that still contains detail in the 246 range and you set your white point to 244, you will clip (i.e., set to pure white) everything at 244 and above.

Maybe I am misunderstanding what you mean by white and black point in this regard. What do you mean by setting your white and black points to predetermined values?
kirk

No, no misunderstanding whatsoever - you described what I was describing more or less. I was just wondering about the concept of taking your darkest and lightest areas of importance and "forcing" them to be certain luminance levels; that's what I was referring to when I referenced "predetermined values." As in going into the levels dialogue and changing the bottom gradient slider (the outputs) to read something like 8 and 244 for black and white point outputs, then pulling the black and white points in the upper area (the inputs) until they just barely hit the clipping point. This would end up forcing your darkest detail areas to 8 and brightest detail areas to 244. I was wondering if this is a desirable practice most importantly.

kirkt wrote in post #16201168
You cannot print white, so you are depending upon the "white" of the substrate upon which you are laying ink to give you your white. White varies across papers and the ability to lay down no ink is a function of your printer, printer+ink profile and your image data. Because paper+ink has a dynamic range that is lower than most output devices, optimizing your data for print to use that entire dynamic range (if that is the goal of that print) is something that involved understanding how best to condition your image data for the specific printer+ink+paper and the associated profile. Printing neutral blacks is also a function of your printer and profiles - for example, does your printer have a dedicated black mode, or is black (gray) composed of C,M,Y and K?
kirk

This I understand, but my main curiosity was whether this is a good thing or not, which I'm still unclear on. By the way your describing things, it seems like it is completely normal and expected to have your paper serve as your whites. Can you just confirm? I do have a cheap "photo" printer that has BCMYK, but I send my images to mpix for printing. And on a side note, after reading your response to this, at what point does the average high quality professional printer stop being able to print very light grey and just lets the paper show it?

kirkt wrote in post #16201168
The above depends on many things, including the desired final output. Your display may be able to support detail at a luminance value of 8, but your printer may not and printer detail may suffer at 15 or 20 or whatever.

So, yo need to know your output device and plan for it. Color management and appropriate target values for display calibration and profiling become really important. Viewing a print under appropriate light is the key, if printing is the final output, because this is the reference to which you are attempting to match your display. Some folks will make display profiles for each paper they print on, because the specific paper dictates the white point and contrast ratio (black point) needed to establish the display target values.

kirk

This makes tons of sense. If you send your pictures out to get printed professionally however or don't know exactly where they'll end up, how do you gauge this other than test prints? Do you just wing it or are there kind of commonly expected values. For example, "Oh yeah - whenever I send my files off to mpix or _____ or ______ or any place like that, I make sure my blacks are at least x and my whites don't go above/below x."
If I had my own high end Epson (like my mother does - lucky son of a gun) then this would be an easy process because I could just experiment whenever I want.

dmward wrote in post #16203109
In my opinion, when printing, or when saving for screen viewing, every image needs at least a little black. i.e. something that is O, as well as something that is close to white i.e. 255. However, the significant whites need detail (93% in Lightroom or about 248 rgb.) And also blacks with details (14% in Lightroom or about 10 in rgb.)

That's also about how I did B&W prints in the darkroom with film, paper and chemicals.

And I like a fairly steep tone curve in between. I mostly printed on Contrast IV or above on paper.

Something like this:

This is the stuff that dreams are made of. That's the silvery quality I'm always after in my black and whites. I know it has a LOT more to do with things other than number values, like lighting and contrast and other things, but that is awesome tone.

Wilt wrote in post #16203156
I'm going to voice dissent, David!

If you meter things with a one degree spotmeter, one might find that the range of inherent brightness levels might span only 3EV range...from brighter than middle grey to white, for example. Let us use the range of EV10-EV13, with exposure set to EV10. So in that type of scene there is no 'black', when you exposed middle grey to record as middle grey.

Similarly, with the black cat in the coal mine, the range might be EV2-EV5, with middle grey reading EV5, so there is no 'white' in the scene.

Now for artistic interpretation of the scenes, one might want to print either photo so that we increase the contrast range of the final print to span a wider range of tones, but that is not portraying anything to the range of inherent brightness levels.

I knew this would come up at some point, and I didn't think of specifying in my original post. I am, for at least this thread, referring to more full dynamic range photos with a full range of tones from deep shadows all the way through sparkly highlights.

Wilt wrote in post #16203156
I'm going to voice dissent, David!

If you meter things with a one degree spotmeter, one might find that the range of inherent brightness levels might span only 3EV range...from brighter than middle grey to white, for example. Let us use the range of EV10-EV13, with exposure set to EV10. So in that type of scene there is no 'black', when you exposed middle grey to record as middle grey.

Similarly, with the black cat in the coal mine, the range might be EV2-EV5, with middle grey reading EV5, so there is no 'white' in the scene.

Now for artistic interpretation of the scenes, one might want to print either photo so that we increase the contrast range of the final print to span a wider range of tones, but that is not portraying anything to the range of inherent brightness levels.

Wilt,
I understand what you're suggesting and in some situations I agree. But even when the tonal range of an image is less that true white to black, I have alway preferred to attempt to get something white and something black.
Personal preference is for a steep S curve as well.

dmward wrote in post #16212013
Wilt,
I understand what you're suggesting and in some situations I agree. But even when the tonal range of an image is less that true white to black, I have alway preferred to attempt to get something white and something black.
Personal preference is for a steep S curve as well.

We have a difference of approach. No one approach is less valid than the other...realism vs. artistic expression. There is a valid time for each. Sometimes the client demands exacting reproduction of reality from the commercial photographer, other times the art director want something that 'grabs' the viewer for an ad campaign.

from a Google search of high key photos...(not mine)

http://www.google.com …18,s:0,i:135&tx​=109&ty=74

A very effective use of dark tones in a predominantly high key shot (also from previous Google search)

http://www.google.com …ved=1t:429,r:95​,s:0,i:371