Leaves

[Gerry F]

DSC05882 by gerald fraser, on Flickr

[Manfred M]

Sorry Gerry, but this badly overexposed image does not work at all. If you look at all of the red in the screen shot from Adobe Camera Raw, the red areas show the places where the highlights have been clipped, i.e exceeded your camera sensor's ability to capture the data.

[xpatUSA]

Sorry Gerry, but this badly overexposed image does not work at all. If you look at all of the red in the screen shot from Adobe Camera Raw, the red areas show the places where the highlights have been clipped, i.e exceeded your camera sensor's ability to capture the data.

Manfred, I wonder if the "badly exposed" was artistically deliberate, i.e. not the horrible mistake that is implied?

There is also the possibility that the camera captured the scene well enough and that the background got clipped (deliberately or otherwise) during conversion and post-processing. After all, ACR does not warn of clipped raw data, only converted/rendered levels.

[Manfred M]

Manfred, I wonder if the "badly exposed" was artistically deliberate, i.e. not the horrible mistake that is implied?

Exposing to blow out the highlights is bad practice, regardless if it is deliberate or not.

[xpatUSA]

Exposing to blow out the highlights is bad practice, regardless if it is deliberate or not.

A bit blunt, I think. So, no highlights should be blown?!!

[xpatUSA]

Gerald, is the posted image processed deliberately to be high-key, soft and low contrast?

[Manfred M]

A bit blunt, I think. So, no highlights should be blown?!!

No, there are circumstances where this is generally considered to be an acceptable practice; specular highlights, shots directly into light sources (sun, artificial light). That being said, I do know photographers that would disagree with that statement and would concerned with any significant blown out areas.

You tend to be quite concerned when even one colour channel is clipped. Why would you find clipping all channels acceptable?

[xpatUSA]

No, there are circumstances where this is generally considered to be an acceptable practice; specular highlights, shots directly into light sources (sun, artificial light).

I knew that; hence my response to the all-encompassing statement in #4.

That being said, I do know photographers that would disagree with that statement and would concerned with any significant blown out areas.

You tend to be quite concerned when even one colour channel is clipped. Why would you find clipping all channels acceptable?

The question "why" is easily answered in 'Exposure 101' articles and also your "there are circumstances where this is generally considered to be an acceptable practice; specular highlights, shots directly into light sources (sun, artificial light)" - rather than your earlier "Exposing to blow out the highlights is bad practice, regardless if it is deliberate or not".

I still think the "blown" background in Gerald's posted image is neither a mistake nor is it bad practice.

Not wishing to speculate further or to bat basic exposure back and forth, I hope that Gerald will clarify his image for us both.

[Manfred M]

I knew that; hence my response to the all-encompassing statement in #4.



The question "why" is easily answered in 'Exposure 101' articles and also your "there are circumstances where this is generally considered to be an acceptable practice; specular highlights, shots directly into light sources (sun, artificial light)" - rather than your earlier "Exposing to blow out the highlights is bad practice, regardless if it is deliberate or not".

I still think the "blown" background in Gerald's posted image is neither a mistake nor is it bad practice.

Not wishing to speculate further or to bat basic exposure back and forth, I hope that Gerald will clarify his image for us both.

Ted - carefully said, I will stick with my original statement as a general "rule" that most experienced photographers will work with. Gerald's image does not contain areas of specular highlights, so far as I can tell.

The exception to that general rule has always been that sometimes, given a sensor's dynamic range limitations, there are some circumstances where the practice can be deemed to be acceptable. This is not a universally held opinion and many photographers will try different techniques to avoid over exposing even extreme highlights, just as they will work to avoid crushing the shadow detail.

It appears unlikely that this was the case with Gerald's flowers and I suspect he is experimenting to see what people will say. If he is trying for a high key look, I would encourage that, so long as the image is well exposed. I have run into quite a number of photographers who thought high key = overexposed, when that is definitely not the case.

[Wavelength]

Before reading the discussion above, when i looked at the image, i felt that it was a deliberate over exposure, to create this kind of effect, where most of the subjects is lost keeping some suggestive remainings only....then i felt its OK, but at the same time it may generate the tendency to withdraw the eyes as if looked at the glowing Sunshine that sores eyes

[billtils]

A quick look at Gerald's flickr page (link in the OP) indicates that the image as posted is meant to look like that.

[xpatUSA]

... many photographers will try different techniques to avoid over exposing even extreme highlights ...

It appears unlikely that this was the case with Gerald's flowers and I suspect he is experimenting to see what people will say. If he is trying for a high key look, I would encourage that, so long as the image is well exposed. I have run into quite a number of photographers who thought high key = overexposed, when that is definitely not the case.

Manfred, so that others here will be clear about our meanings of "exposed", I believe that you are equating image brightness with exposure, whereas I use the technical meaning, namely sensor illuminance H x time, i.e. lux-seconds. So "over-exposed" to me means saturated photocells whereas to you it means pixel values clipped at 255 (8 bpc).

I have actually hacked into my Sigma raw converter and changed the label for the "exposure" slider to "brightness" ...

[DanK]

I know that I am repeating myself, but I don't much care about rules, and I don't buy the notion of a general good practice that applies to all photography. My concern is with the results--whether I find the resulting image appealing and interesting.

This image doesn't work for me. It simply looks overexposed. The fact that most of the leaves are out of focus just exacerbates the problem, in my view. That's just my taste.

[Gerry F]

Firstly thanks for the lively discussion. I have learnt a lot from it already. The photo was taken in the early hours of Saturday morning when the sun was quite low and bright. I certainly intended to attain the soft effect with the strongly highlighted background. I used the look rather than basing my settings on the in camera technical parameters.
I was using a Sony 85 mm GM prime lens at f1.4, 1/160 since I didn't have a tripod for long exposure. ISO 100

To Manfred - just a comment on your opening sentence: "Sorry Gerry, but this badly overexposed image does not work at all."
This slams the door shut on any discussion that I could initiate. Neither does it suggest any way of obtaining a similar outcome using different methodology. Luckily for me some other members do not entirely agree with you and a great discussion was initiated.

[xpatUSA]

I know that I am repeating myself, but I don't much care about rules, and I don't buy the notion of a general good practice that applies to all photography. My concern is with the results--whether I find the resulting image appealing and interesting.

This image doesn't work for me. It simply looks overexposed. The fact that most of the leaves are out of focus just exacerbates the problem, in my view. That's just my taste.

Whereas many or most people aim for realism, there is a broad genre namely "Artistic Photography" an example of which is "Pictorialism" where focus is less important so long as an impression of leaves is given. Indeed, in the works of Turner, parts of some of his artworks are unrecognizable! See "Form Through Gradation" here.

My point being that there should be room for Gerald's image style here without such strong criticism as has been posted so far.

[xpatUSA]

Firstly thanks for the lively discussion. I have learnt a lot from it already. The photo was taken in the early hours of Saturday morning when the sun was quite low and bright. I certainly intended to attain the soft effect with the strongly highlighted background. I used the look rather than basing my settings on the in camera technical parameters.
I was using a Sony 85 mm GM prime lens at f1.4, 1/160 since I didn't have a tripod for long exposure. ISO 100

Thanks for the clarification, Gerald. It is as I thought.

[Manfred M]

Manfred, so that others here will be clear about our meanings of "exposed", I believe that you are equating image brightness with exposure, whereas I use the technical meaning, namely sensor illuminance H x time, i.e. lux-seconds. So "over-exposed" to me means saturated photocells whereas to you it means pixel values clipped at 255 (8 bpc).

I have actually hacked into my Sigma raw converter and changed the label for the "exposure" slider to "brightness" ...

Not really Ted - I am suggesting that we try to take images where at capture we have NO values that show up as either 0 or 255 when we examine the histogram. Ideally that goes for each individual colour channel. I am ignoring the fact that our cameras tend to use 14-bit data and that we commonly discuss this using 8-bit values. We also know that there are situations where this is not possible or potentially desirable and there have been many discussions here and elsewhere what should be done in those cases.

The second part of the discussion is how to process images and that is something we are actually looking at here. The final image is only as good as the captured data, so assuming that this was done properly, we can at least start off with good material.Let's also admit that pure white and pure black are theoretical constructs that real world objects only get close to matching. Let's also remember that in photography, those "rules" vary slightly when it comes to B&W photography versus colour work.

As Gerald is dealing with a colour image, let's stick with that for now. The way he has pushed his images, I suspect he is trying to create images that are heavy in the light tones. There is a correct way to do that and an incorrect on. Unfortunately, he has chosen the wrong way to approach this, as can be seen by the ACR screen capture I posted in #2.

By the "wrong way", what I am saying is achieving a brighter image by pushing the data to the lighter side, without regard to how much of it exceeds a value of 255, i.e. pure white. The "correct way" is to push the data to the right without pushing it to be pure white (and through that process preserving highlight details and textures), which make for a much stronger image.

Let me demonstrate what Gerald has done and the problem with that method and what I would deem to be a more correct way of doing the same thing.

Here is a very light coloured image, an arctic fox in the winter. It is already dominant in the light tones and falls within my definition of being "correctly exposed". For the most part I will be showing screen shots of Adobe Camera Raw screen captures.




When I take Geralds method of turning up the exposure, I turn much of the image brighter.






When I turn on the clipping indicator, all of the red shows the areas that have been clipped.






One of the "correct" ways to do this is to leave the black points and white points alone, but change the mid point. I've used Photoshop's Levels control to do this, although in practice, I use Curves.






When I bring the image back to ACR, it shows a tiny bit of clipping on the snout, but for the most part, textures have been protected and not turned into pure white. This would be considered a stronger approach.

[DanK]

Whereas many or most people aim for realism, there is a broad genre namely "Artistic Photography" an example of which is "Pictorialism" where focus is less important so long as an impression of leaves is given. Indeed, in the works of Turner, parts of some of his artworks are unrecognizable! See "Form Through Gradation" here.

My point being that there should be room for Gerald's image style here without such strong criticism as has been posted so far.

I agree. Hence my caveat that what I described was only my taste. Likewise, my comment that I am not very sympathetic to rules applied generally.

I don’t care for the pictorialist genre, FWIW. However, I find some non-realist photography both interesting and attractive, e.g., some of Olivia Parker’s work. And as I have pointed out in an earlier thread, a good bit of her work violates the “rules.”


Sent from my iPad using Tapatalk

[xpatUSA]

Not really Ted - I am suggesting that we try to take images where at capture we have NO values that show up as either 0 or 255 when we examine the histogram. Ideally that goes for each individual colour channel.

OK Manfred, I understand your POV, now that you have said "at capture". So, ideally, the LCD histogram should range at most from 1 to 254 levels at the time of the shot. With the caveats a) that if one shoots raw and later changes one's mind about WB, JPEG settings, color space, etc., there could still be clipping or bottoming - and b) a 'between 1 and 254' LCD luminosity histogram could still have clipped channel values and c) the aforesaid exceptions regarding specular highlights and such.

Do pardon the huge snip. I did read your entire post and I agree as regards the use of curves in the production of high-key images. The fox examples were really illustrative of your main point, BTW.

[Manfred M]

OK Manfred, I understand your POV, now that you have said "at capture". So, ideally, the LCD histogram should range at most from 1 to 254 levels at the time of the shot. With the caveats a) that if one shoots raw and later changes one's mind about WB, JPEG settings, color space, etc., there could still be clipping or bottoming - and b) a 'between 1 and 254' LCD luminosity histogram could still have clipped channel values and c) the aforesaid exceptions regarding specular highlights and such.

Do pardon the huge snip. I did read your entire post and I agree as regards the use of curves in the production of high-key images. The fox examples were really illustrative of your main point, BTW.

Ted - carefully said, I did not say just at capture, although that is critical. I also strongly advocate protecting detail in the final image. While this example is all about highlights, the same can be said about shadow details.

None of the examples I posted dealt with capture, although it is a prerequisite for a strong image. Anyone I have studied under with formal photographic training (college / university programs in photography), especially the people that used densitometry as a working tool in the film days, are absolutely adamant about this.

I find an image like Gerald's example with lots of blown out highlights hard to look at. There is a way to get a very bright image without clipping, so what's wrong with doing that?

P.S. I am getting towards the end stages of doing online judging of a national competition; 425 images. One of the most common errors I am seeing is exactly this. The images that are done without blowing out the highlights (or crushing the shadow details) simply look a lot better and tend to get higher scores. There is one image with significantly crushed blacks that works well and it is probably the only one I did not penalize in the scoring.