Golden Pollen (shot at 1x1)

[DanK]

It took me less than a second after I typed in the question to pull up multiple articles on the advantages of ETTL

you can find anything on the web. In any case, the author of the post you linked to wrote this:

It’s so simple to state the combination of these two philosophies that renders both extremes silly: You should expose as bright an image as you can without clipping.

It seems that the author didn't understand, but what he is recommending is not "combining these two philosophies." What he is recommending is ETTR: expose as far to the right as you can without clipping.

The ETTR principle is straightforward math, not philosophy. It is true that it matters less than it used to because sensors are better, and it matters less with a low-noise camera than with a high-noise camera. I can get away with a lot more using my 5DIII than with the 50D it replaced. However, the principle is simple: ETTR gives you more signal, and that means the signal/noise ratio is higher. Brightening any image, whether in post or by amplifying the signal using a higher ISO, will amplify noise as well as signal. Darkening an image will decrease noise as well as signal.

So the bottom line is that ETTR will not always matter, but ETTL makes no sense. Someone correct me if I am wrong, but I have never seen a persuasive argument that exposing ETTL is in any way better than ETTR and then darkening in post. I do a fair amount of night photography, and I often expose to the right even then, even though the images then have to be darkened to look like nighttime. The result is smooth, low-noise images with rich colors that need no noise reduction, e.g.:

[Manfred M]

If it's all down to maths and physics then we can replace requests for C&C with some AI generated score that measures conformity to these and forget about creativity.

I prefer to follow the advice to shoot and look, specifically look at the image as shot and its histogram and use the latter to decide whether or not to tweak any settings. Or more accurately, I sometimes am sensible and follow the advice to ... .

It's situational and a slavish conformity to ETTR or ETTL (or any other "rule") is less than optimal. An example would be using exposure compensation to get the best image of a swan or other large white bird - on a dark pond with a dark background or in flight against a bright sky.

Bill - I would never suggest that everything in a creative field is purely down to math and physics (and chemistry if we include film photography). Photography has always been a blend of the technical side of things, which requires math and physics to describe them (sometimes referred to as the "craft" aspect of photography) and the creative side (how the elements are arranged in the image).

Even the technical elements have a creative aspect; while we can write about the technical considerations; focal length, aperture, shutter speed, sensitivity (ISO), these too are often not clear cut and purely technical. We can discus the appropriateness of the selected focus plane, shutter speed, etc. In some cases it is easy to get agreement that the settings are not appropriate but at other times, the differences can be quite subtle.

The one area where there is general agreement among knowledgeable photographers, especially when shooting in raw, is that getting as much usable data as possible is desirable. Having values of 0, where all shadow detail is crushed or values of 255, where highlights are clipped is generally not desirable and restricts what can be done with the image in post.

That is precisely the argument I am making. In general, it should the the photographer's goal to get a good capture and when there is enough "headroom", crushing shadow detail or clipping highlights is not good technique. This gives the photographer the maximum amount of data to work with. ETTR is nothing more than optimizing the quality of the data, assuming that there is room in the highlight end of the exposure.

How that data is handled in post processing is a totally different matter, as the photographer can make choices that give him or her the visual effect that they want to create.

How to handle situations where the dynamic range of the camera's sensor is exceeded by the scene is another completely different issue and there are some basic ground rules most experienced photographers tend to follow, but even here there are exceptions.

[xpatUSA]

Something odd going on here, Brian.

When I downloaded the new sRGB image it looked the same as the old one, more or less.

But when I use the right-click Mozilla EXIF viewer to see the histogram on-line, I get:



I'll need to look at this some more ...

I've got that one figured out, I think, and it is indeed a red herring. The histogram above is of the actual screen pixel values, not the RGB values in the posted image file. So all of Brian's original images remain over-saturated, even though the screen driver or the screen profile or indeed my screen settings say otherwise.

[DanK]

ETTR is nothing more than optimizing the quality of the data, assuming that there is room in the highlight end of the exposure.

How that data is handled in post processing is a totally different matter, as the photographer can make choices that give him or her the visual effect that they want to create.

Exactly my point as well.

[billtils]

The one area where there is general agreement among knowledgeable photographers, especially when shooting in raw, is that getting as much usable data as possible is desirable. Having values of 0, where all shadow detail is crushed or values of 255, where highlights are clipped is generally not desirable and restricts what can be done with the image in post.

+1 to that Manfred.

[JBW]

I do.

Brian, it's as if I had spent no time at all here explaining that the problem is one of over-saturation, not lack of blue.

I took your explanation seriously. But there was no way reducing the saturation made the shot look better.

[JBW]

The second camera profile selection is indeed the more common Hewlett Packard version sRGB IEC61966-2.1. A pity it didn't fix everything but there's no real reason why it should.

The Mozilla histogram is the same on all your "golden" shots in this thread - it shows a decent amount of blue with little to no bottoming. For now, I propose to treat it as a red herring.

I opened your "x3" in the the GIMP and decomposed the image into R, G and B layers. Blue was still crammed up to the left. I expected no different.

I'll keep looking at the posted images but I'm running out of steam to be honest.

Do you have any other app that can open the raw file, or was it an OOC JPEG?

I've got the one that sony puts out. Ted, relax you've done enough.

[Manfred M]

I took your explanation seriously. But there was no way reducing the saturation made the shot look better.

That is a creative choice Brian. In part that is driven by both the equipment you are using to view the shot, the viewing conditions (look at your screen at night with the lights turned way down versus in the daylight) and how your visual system sees things (the eye / brain connection). The image that you are seeing is likely somewhat different than the one I am looking at for those reasons. Even if we were standing side by side, looking at your computer screen, our visual system will still have us seeing a slightly different image.

It is also driven by your own personal taste. That is what it is and there is no right or wrong way of looking at things with that overlay.

The only comment that I was trying to make in #3 is that if you are missing data due to the choices you made in exposure, you have reduced your ability to manipulate your capture in post and that ultimately restricts how much you can do with the image.

[Manfred M]

So the bottom line is that ETTR will not always matter, but ETTL makes no sense. Someone correct me if I am wrong, but I have never seen a persuasive argument that exposing ETTL is in any way better than ETTR and then darkening in post

Agree 100% Dan. In fact there is no serious discussion of ETTL in this context that I am aware of. Using a technique that deliberately reduces the amount of data that the camera captures seems quite counter-intuitive.

That being said, there is a photographic approach known as ISO-invariance that sort of moves in that direction. In this practice, the camera is set to the base ISO and rather than using in-camera amplification, gain is applied in post-processing. Only certain fairly recent camera models seem to fall into this category; all use sensors that have an extremely wide dynamic range. Shooting in low light would effectively move in the direction of expose to the left.

[xpatUSA]

I took your explanation seriously. But there was no way reducing the saturation made the shot look better.

Ok. You're right. Reducing saturation rarely makes an image "look better". The difference is a little more subtle than that. In this case, with virtually no blue channel, your image is just red+green - one step towards monochromatic. More colorful, yes.

[xpatUSA]

Do you have any other app that can open the raw file, or was it an OOC JPEG?

I've got the one that Sony puts out. Ted, relax you've done enough.

Ok. It would have been nice to know what actually forced those blues to zero, though.

[DanK]

That being said, there is a photographic approach known as ISO-invariance that sort of moves in that direction. In this practice, the camera is set to the base ISO and rather than using in-camera amplification, gain is applied in post-processing. Only certain fairly recent camera models seem to fall into this category; all use sensors that have an extremely wide dynamic range. Shooting in low light would effectively move in the direction of expose to the left.

That's not my understanding of ISO invariance, although I may misunderstand it because none of my cameras are ISO invariant. In the case of cameras that are not ISO invariant, the penalty from amplifying the signal in camera by boosting ISO is less than the penalty from amplifying it in postprocessing, up to several stops above base ISO. (Above that, it makes no difference.) In ISO-invariant cameras, the loss of DR is the same regardless. (I have no idea why, although I have observed that as you wrote, ISO-invariant cameras tend to have high DR.) As a result, with ISO-invariant cameras, there is no reason to raise ISO on the camera in order to ETTR. However, if there is enough light to ETTR at base ISO, the logic above still applies.

[Manfred M]

Ok. It would have been nice to know what actually forced those blues to zero, though.

My guess would be the -0.3 EC that Brian mentioned in #1.

[xpatUSA]

My guess would be the -0.3 EC that Brian mentioned in #1.

Sorry, Manfred, I can not agree with that.

-0.3 EC is a ratio of about 0.8. So, if there were tones equivalent to say 20 in the scene, they would come out as 16, not zero. (simplified estimate, gamma notwithstanding).

[JBW]

Agree 100% Dan. In fact there is no serious discussion of ETTL in this context that I am aware of. Using a technique that deliberately reduces the amount of data that the camera captures seems quite counter-intuitive.

That being said, there is a photographic approach known as ISO-invariance that sort of moves in that direction. In this practice, the camera is set to the base ISO and rather than using in-camera amplification, gain is applied in post-processing. Only certain fairly recent camera models seem to fall into this category; all use sensors that have an extremely wide dynamic range. Shooting in low light would effectively move in the direction of expose to the left.

Now we are getting somewhere!
If I know how I want my shot to turn out what's wrong with setting up the camera to be close to the result I am after? To me it is philosophically no different than working a shot in live tethered until I get to where I am happy.

[Manfred M]

Now we are getting somewhere!
If I know how I want my shot to turn out what's wrong with setting up the camera to be close to the result I am after? To me it is philosophically no different than working a shot in live tethered until I get to where I am happy.

There are two parts to this Brian.

1. Framing and composition - shooting tethered will let you confirm those items and especially focus much better than you can do with the screen on the back of the camera.

2. Good exposure - going directly into Capture One will give you a better overview as to how good the capture is from an exposure standpoint. As a raw shooter, your aim should be to get as much data as possible for you to work with in post, so avoiding clipping and getting the histogram where it needs to be.

In the film days, I handled film much the same way as I handle raw data; getting the best exposure possible and then working the print in the enlarger to get a moody dark look or a light airy look. The place to start was always a good negative. A thin, underexposed negative gave poor results just as much as a heavy, overexposed one did.

The only time I can see using the approach you are looking at is if you are in a rush and don't plan to do any PP work at all and you are planning to use a SOOC JPEG as your final result. I dpn't think you will find a lot of photographers at CiC advocating that approach. If you are looking for a moody selfie for Facebook or Instagram, why not?

[JBW]

There are two parts to this Brian.

1. Framing and composition - shooting tethered will let you confirm those items and especially focus much better than you can do with the screen on the back of the camera.

2. Good exposure - going directly into Capture One will give you a better overview as to how good the capture is from an exposure standpoint. As a raw shooter, your aim should be to get as much data as possible for you to work with in post, so avoiding clipping and getting the histogram where it needs to be.

In the film days, I handled film much the same way as I handle raw data; getting the best exposure possible and then working the print in the enlarger to get a moody dark look or a light airy look. The place to start was always a good negative. A thin, underexposed negative gave poor results just as much as a heavy, overexposed one did.

The only time I can see using the approach you are looking at is if you are in a rush and don't plan to do any PP work at all and you are planning to use a SOOC JPEG as your final result. I dpn't think you will find a lot of photographers at CiC advocating that approach. If you are looking for a moody selfie for Facebook or Instagram, why not?

Today i went out and sot at 0 compensation. Admittedly the darker 0. (yes my camera has 2 0's) I got what I expected a lot more blown out areas. I'll continue to shoot at -0.3

What was unexpected was that with everything i shot, from lizards to poppy plants only the vanishing blue flower shows no blue. (even when shot at 0 compensation)

Is it possible that the yellow and purple combo is simply too much for my camera?