Black

[lovelife65]

I'm one of the outliers here as well. The pure black works IMO. It creates a beautiful canvas. Dans photo of the flower on black is wonderful. As is Andres.

[Rufus]

My non-technical take on this is a fudge.

The effect of white clipping demonstrated by Mandred's 2 images in post #19 can certainly be seen in Lightbox with the hair being better defined without the clipping. Once you know it is there you can look for it, but the casual viewer or average punter would not see it. So it becomes a matter of pride in your work and how large you are likely to print an image that influences one's view of the clipping. However, I do take the point that pure white may show as a different texture on a print due the lack of any ink whatsoever in that area.

As for black, in principle I don't see anything wrong with an image that has areas that are intentionally totally black, but care needs to be taken to ensure that the fall off in the density of black occurs in an aesthetically acceptable way.

[george013]

My non-technical take on this is a fudge.

The effect of white clipping demonstrated by Mandred's 2 images in post #19 can certainly be seen in Lightbox with the hair being better defined without the clipping. Once you know it is there you can look for it, but the casual viewer or average punter would not see it. So it becomes a matter of pride in your work and how large you are likely to print an image that influences one's view of the clipping. However, I do take the point that pure white may show as a different texture on a print due the lack of any ink whatsoever in that area.

As for black, in principle I don't see anything wrong with an image that has areas that are intentionally totally black, but care needs to be taken to ensure that the fall off in the density of black occurs in an aesthetically acceptable way.

A little bit of cheating. Not only the background/clipping has changed, but all the picture.

George

[DanK]

A little bit of cheating. Not only the background/clipping has changed, but all the picture.

George

Not really relevant here, George. And in the case of the image I posted, it's not correct anyway: the relevant edits were done in a selection that included only the black areas.

[Round Tuit]

Fair enough, but this is considered a technical flaw, just as much as blown out highlights are.

If I were to post and image like this:




People would either be all over me for having clipped the highlights badly or they would suggest no big deal because I should be able to do what I want. I find that people are more sensitive to clipped highlights than crushed shadow detail, probably because we were taught it is better to crush shadows than to clip highlights.

The problem with pure white and pure black backgrounds is that they contaminate edges of the image and detail is lost. I don't have a suitable image to demonstrate what happens with a black background, but I the high key image I've been showing reveals this issue fairly clearly. Both of the following images are shown at 100%, which doesn't happen very often when displaying on screen but is something that is common in prints. If you look at the edges of your black background image, you are going to see the same type of impact.

Open one of these images in Lightbox and toggle back and forth between them.



Image 1 - edge contamination can be seen in many places, but the fine details of the hair show it most clearly.





Image 2 - no edge contamination





Some people will say this is minor and no one will notice. Once one is aware of the issue, we see it right away. This is very much along the same lines as highlights around edges that come from sharpening or Chromatic Aberration issues. We don't see them until we are sensitized to them and then they poop right out at us.

Manfred - the differences in the edges of the two photos is quite evident to me and not just on the hair but also on Melanie's shoulder and the boy's head. It looks to me though that you clipped the white background by lightening the entire photo. If this is the case, then you would start to lose edge detail even before reaching the clipping point. Darkening your photo would enhance edge detail. This is so because edges in the real world are always somewhat "feathered'.
Dan and I selected the background so that the darkening only applied to it and made sure that the selection protected the main subject. If you look at our photos with the crushed black shown in blue, you will see that the crushed black does not reach our subject. The edges are protected by a narrow band of dark grey. The width of that protective band is easily controllable by adjusting the feathering of the selection.
Selecting the white background of your photo while protecting the fine light hair would of course be more difficult to achieve in practise but in theory should be feasible.

[xpatUSA]

Speaking of edges and not wishing to start a Foveon vs CFA war, this image might be of interest:



The problem with high-contrast edges for "most of us" (but not me) is the good old de-mosiacing a.k.a "smart guess-work". Of interest to members in this thread is the background which is not black. However, the only detail by and large is noise, see histogram.

Of particular interest in that shot is the approx 1 pixel-wide object at the top right of the lemon which is showing no sign of "contamination" at all, in spite the obvious over-sharpening of the lemon edge (halos):



The noise of course prevents a perfectly homogenous-level back ground and it is certainly not "black"; but still the subjects have high enough contrast to be similar to the thistle-down shot posted earlier. I'll hazard a guess that the small object would show no contamination at all against pure black. On the other hand, if the background had been somehow changed to pure white, it would likely have disappeared, perhaps along with the lemon.

A question is: would this image have passed scrutiny in the hypothetical competitions mentioned earlier by others? If so, why?

Not my image, by the way ...

[Manfred M]

Manfred - the differences in the edges of the two photos is quite evident to me and not just on the hair but also on Melanie's shoulder and the boy's head. It looks to me though that you clipped the white background by lightening the entire photo. If this is the case, then you would start to lose edge detail even before reaching the clipping point. Darkening your photo would enhance edge detail. This is so because edges in the real world are always somewhat "feathered'.
Dan and I selected the background so that the darkening only applied to it and made sure that the selection protected the main subject. If you look at our photos with the crushed black shown in blue, you will see that the crushed black does not reach our subject. The edges are protected by a narrow band of dark grey. The width of that protective band is easily controllable by adjusting the feathering of the selection.
Selecting the white background of your photo while protecting the fine light hair would of course be more difficult to achieve in practise but in theory should be feasible.

André - your masking ability has to be much better than mine if you manage to separate the fine details from the background.

I found an image that will demonstrate with a dark background. I was very careful in how I built my masks and handled the PP work so the only difference should be a background that is almost black versus one that is totally black (crushed detail). Here are the ACR shots


1. Shadow detail blocked






2, Shadow detail not blocked (mostly)






3. Here is a closeup of the image with blocked shadow detail. There is contamination in the detail (hair).






4. Closeup of the image with minimal crushing of shadow detail.




Once again, the contamination and detail loss is visible. Short of hand masking every detail, the contamination cannot be avoided.

I also wonder how many people can see the difference between the black background in the #3 versus the very, very dark gray in #4.

One final thought; Adobe (in Lightroom and ACR), Phase One, with Capture One and DxO Labs with PhotoLab all have functionality to show detail loss in both the highlights and shadows. If these three well regarded software developers feel it is important to let us know that loss of detail problems have occurred, why we feel it is okay to push really hard to crush the shadow areas.

[DanK]

This thread is now conflating two entirely different issues:

1. The original issue: is it necessarily a technical error to print or display pure black?

2. The second issue: the risk of contamination when doing ANY significant edit to a selected area with a complex border:

if you manage to separate the fine details from the background.

These are not necessarily related at all. The question of printing or displaying at a value of 0 can arise with no selection at all, and the problem of contamination can arise (indeed, it has arisen in quite a number of my images) even when one is not displaying anywhere close to 0 or 255.

In actual practice, I have much more often had visible contamination when not bringing the selected area down to 0 because I generally only do that when the base tone of the selection is quite dark, and as Andre wrote, I keep the border away from the details.

If these three well regarded software developers feel it is important to let us know that loss of detail problems have occurred, why we feel it is okay to push really hard to crush the shadow areas.

Because very often--indeed, in my case, most of the time--we do want to avoid crushing shadow areas. Sometimes, we don't. My car alerts me when I cross a lane divider because often it's accidental, and then I don't want to do it. That doesn't mean that it is always an error to change lanes.

[george013]

I don't know the subject anymore. If it's about the background not much has changed. From 0,0,0 to average 1,0,0. Looks more that the red channel has been upgrade a little.

George

[Manfred M]

This thread is now conflating two entirely different issues:

1. The original issue: is it necessarily a technical error to print or display pure black?

2. The second issue: the risk of contamination when doing ANY significant edit to a selected area with a complex border:



These are not necessarily related at all. The question of printing or displaying at a value of 0 can arise with no selection at all, and the problem of contamination can arise (indeed, it has arisen in quite a number of my images) even when one is not displaying anywhere close to 0 or 255.

In actual practice, I have much more often had visible contamination when not bringing the selected area down to 0 because I generally only do that when the base tone of the selection is quite dark, and as Andre wrote, I keep the border away from the details.



Because very often--indeed, in my case, most of the time--we do want to avoid crushing shadow areas. Sometimes, we don't. My car alerts me when I cross a lane divider because often it's accidental, and then I don't want to do it. That doesn't mean that it is always an error to change lanes.

Sort of Dan. The point I had originally made was that significant crushing blacks is considered to be poor photographic practice. A little bit tends to be acceptable. The same issue is out there regarding blowing out the highlights, same type of issues except at the other end of the the scale. Loss of detail, whether that be in tonality or in texture is something that should be questioned.

I've tried to demonstrate why this is (as opposed to being something purely philosophical) and there is that aspect as well. Destroying data in a way that is not recoverable is something that some photographers believe in quite strongly. Others will say that there is no pure black or pure white in nature (and they are correct, but some colours can be close enough for the point to be academic.

[Manfred M]

I don't know the subject anymore. If it's about the background not much has changed. From 0,0,0 to average 1,0,0. Looks more that the red channel has been upgrade a little.

George

Averages don't really matter at this level. All of the blue areas in images 1 and 2 of #27 are black or have one of the channels crushed. The other areas that are not showing blue have a minimum value of at least 1,1,1. The average is higher, but as you are looking at a screen print, the colour data accuracy cannot be relied upon.

[Cantab]

I’ve gone back to Fernando’s original post.

Um mundo lavado pelo Sol, não sofre de enigmas. Tudo nos parece transparente e lúcido, inclusive a ideia de Deus o Qual, assumindo formas diversas, pode aparecer ao entardecer para dois dedos de conversa. Por sua vez, o “Outro Lado” é a dominação que a tradição judaica dá ao que é demoníaco. Quando entras na sombra nunca sabes como vais sair. Esta imagem perdeu-se nas sombras, expurgou-se da relação anacrónica com a arte ou a beleza.

https://kaskaisphotos.wordpress.com/2019/02/16/black/

Google Translate gave me what appears to be a reasonable translation of Fernando’s Portuguese caption:

“A world washed by the sun, does not suffer from enigmas. Everything seems to us transparent and lucid, including the idea of God Who, assuming various forms, can appear at dusk for two fingers of conversation. In turn, the "Other Side" is the domination that the Jewish tradition gives to what is demonic. When you enter the shade, you never know how you're going to get out. This image was lost in the shadows, expunged from the anachronistic relationship with art or beauty.”

This caption is important, as is the title of the thread, “Black”.

Fernando used a camera to create graphic art which hopefully communicates something. As an aside, if he’d used black paint instead of a digital camera we’d not be having a discussion about whether the level of blackness of the black paint works in this piece of graphic art.

My personal view is that we need to take Fernando’s image and its caption on its own terms. Does it work as a piece of multimedia art?

[DanK]

As an aside, if he’d used black paint instead of a digital camera we’d not be having a discussion about whether the level of blackness of the black paint works in this piece of graphic art.

I agree completely.

My personal view is that we need to take Fernando’s image and its caption on its own terms. Does it work as a piece of multimedia art?

I find the caption quite a reach, but I agree that we should take the image on its own terms.


Sent from my iPad using Tapatalk

[John 2]

In a way, I agree with Dan and Bruce. Sometimes (not always)it's enough to look at an image and like it without analysing whether it necessarily conforms to a given set of criteria. I quite like this and understand it in the context of Fernando's final few words - "When you enter in the shade you never know how you're going to do. This image was lost in the shadows, expurgated-if anachronistic relationship with art or beauty." (imperfect Google translation but just about understandable)

[Round Tuit]

André - your masking ability has to be much better than mine if you manage to separate the fine details from the background.

I found an image that will demonstrate with a dark background. I was very careful in how I built my masks and handled the PP work so the only difference should be a background that is almost black versus one that is totally black (crushed detail). Here are the ACR shots


1. Shadow detail blocked

2, Shadow detail not blocked (mostly)

3. Here is a closeup of the image with blocked shadow detail. There is contamination in the detail (hair).

4. Closeup of the image with minimal crushing of shadow detail.

Once again, the contamination and detail loss is visible. Short of hand masking every detail, the contamination cannot be avoided.

I also wonder how many people can see the difference between the black background in the #3 versus the very, very dark gray in #4.

One final thought; Adobe (in Lightroom and ACR), Phase One, with Capture One and DxO Labs with PhotoLab all have functionality to show detail loss in both the highlights and shadows. If these three well regarded software developers feel it is important to let us know that loss of detail problems have occurred, why we feel it is okay to push really hard to crush the shadow areas.

Thank you Manfred.
This latest example of yours shows that with a better selection the loss of fine details can be alleviated to some extent but would be difficult to eliminate completely. By the way, my masking skills poor at best. This is why I chose a picture with fairly smooth edges as an example of crushed black that does not affect the image quality negatively.
This discussion has been beneficial to me. I will now keep a closer eye on the fine details along the edge when I darken my backgrounds as I definitively consider loss of fine details to be a technical flaw. And, like you, I take a certain amount of pride in the technical quality of my photos. I still hold that crushed blacks are sometime appropriate.
Finally, I would like to apologize to Fernando and his fans for hijacking his thread. In the future, I will endeavor to move such discussions to a different thread of my own.

[MrB]

Ansel Adams: "There are no rules for good photographs, there are only good photographs."

The image from the OP works for me.

Philip

[mugge]

Ansel Adams was exceptional. He has been dead for a number of years now. Let us try something new this time and if we fail - good.

Cheers Ole

[xpatUSA]

Been a lot of talk in this thread about RGB values and several color images have been posted to illustrate this or that.

However, it looks like the original posted image is a grayscale JPEG, not Y'CbCr (color).

So, some of the discussion might be moot:

https://en.wikipedia.org/wiki/JPEG_F...at#Color_space

That is to say that the chroma CbCr was not stored in the JPEG, only Y'.

I say this because, out of curiosity, I downloaded it and tried to decompose it in the GIMP but the GIMP would not do it because "decompose" in the color menu was grayed out.