Uuglypher Loc: South Dakota (East River)

[quote=selmslie]I have told you that I understood it completely and have tried it but it did not make any visible difference with a better camera like my D7000 or D610.
xxxxxxxxxxxxxxxx
Scotty said: "I have told you that I understood it completely and have tried it but it did not make any visible difference with a better camera like my D7000 or D610."

Once again, Scotty, you claim to have tried EBTR, but again you seem to have forgotten that you admitted that you have never determined the ERADR of the cameras you have used. That's a problem. The latter is necessary to carry out the former. Thus, you have NOT tried EBTR. This gets us back to a problem my characterization of which you didn't like. Bad memory...or...? You have to get your stories straight.

I suggest you actually go through the process...actually try EBTR... admit your objections are theoretical, not experiential, and drop your charade of " just knowing" that EBTR won't work. I really don't want you to take my word for it! From the start I've wanted naught but that you give it a fair try yourself. You claimed to have done it, but just a few days ago let it slip that you had not taken the first necessary step of determining your camera's extra RAW-Accessible Dynamic Range (ERADR)!

You've seen the ETTR and EBTR images that clearly demonstrate far less noise in the latter than in the former, even at high ISOs. They were made with a Sony Alpha99 with ERADR of one and 2/3 Stops, but innumerable similar images have been made with a wide range of cameras. For most of the past decade I have successfully practiced EBTR with Sonys (A350, A55, A77, and A99, and a Canon sx50). Make all the demands you like, Scotty, but until you actually resolve to accomplish EBTR yourself, it's obvious that no further information or more images equally and still unequivocally proving the efficacy of EBTR will suffice to change your mind.

Theorize till the cows come home, Scotty, but remember.... honeybees still pollinate flowers and make honey, in spite of aeronautical engineers theoretical "proof" that they can't fly!

You have seen images proving that EBTR fulfills its purpose, and I understand that was probably aggravating to you, to the extent of having to deny the obvious! It has reached the point that denying the efficacy of EBTR is more an article of faith than of fact to you, Scotty, and I long ago resolved not to participate in religious arguments.

Until such time as you try it for yourself, which by your own admission you have definitely NOT done, further discussion on the matter is futile. The greatest mystery to me is why, having spent good money on a decent DSLR, you have failed to assure that you are taking full advantage of its full dynamic range of exposure and, indeed, have failed even to find out how much additional Dynamic range it has. Isn't that a bit embarassing to admit?
Ah well, I've given up on trying to understand your fervent denial of the obvious.

Dave

selmslie Loc: Fernandina Beach, FL, USA

Which purpose, noise reduction or improved shadow tonality?

Both of us have shown that noise is reduced by adding exposure (signal). It is also clearly pointed out in http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html


This article restates the theoretical tonality argument for ETTR with which I have no quarrel.


You can get less noise by exposing at high ISO to the right. I can get the same noise reduction by using a lower ISO. Since I normally shoot between ISO 100 and 400, I get even less noise than you do at 1600 with ETTR. Your method is esoteric and time consuming. Mine is straightforward and efficient, and when in doubt, I bracket - a piece of cake.

So we both have an approach to noise that works to our satisfaction.

Before I complicate my life, I still need to see evidence showing how much improvement, if any, you can produce in your shadows. Words are cheap and theory is worthless unless you can act on it. I can't hang theory on my wall.

I need a reason to use ETTR. You have not given us one. I will even settle for a procedure that I could follow to produce my own example since you are either not willing or able to produce this evidence for us.

Pardon the bluntness, but it is time for you to put up or shut up.

selmslie Loc: Fernandina Beach, FL, USA

Here is an approach you have not mentioned.

Instead of starting at a high ISO and increasing the exposure time as you describe, start from base ISO like 100 or 200 and pick an exposure that clearly separates the right side of the histogram from the edge. It may initially look dark, but you will be assured of minimal noise.

Then, without changing either the aperture or shutter speed, increase the ISO until the histogram reaches the point where you want it, either at the right edge of the JPEG range or as far beyond as you feel is safe.

The second result may look light, but you will end up with minimum noise since you will still have used the highest possible S/N ratio.

Uuglypher Loc: South Dakota (East River)

I have learned to employ two strategies for EBTR:

When shooting landscapes from tripod when long exposures are no problem and there is no wind to move the foliage of foreground trees I almost always use the lowest possible ISO and, as always, use the camera's full allotment of ERADR, thus deriving the maximal of both routes to minimizing or, for all practical purposes, eliminating noise. (recognizing that the tag end of the half of the patabolic curve never really reaches zero, albeit continually approaching thereunto!)

The same approach vis a vis low ISO and capture as far beyond the right as possible applies to any reasonably stable scene or subject under constant illumination, thus ruling out most sports ( except chess and full-body-contact World Tournament SUDOKU), active children at play, and much wildlife. This includes studio portraits and still-lifes, product illustration, floral and food photography, fashion, etc.

A happy exception out-of-doors is on days with heavy overcast when "sunny f/16" suggests f5.6 to f /4 and the light is diffuse and essentially the same from all directions. The higher ISOs permit much faster shutter for active, moving wildlife (including active dhildren), sports, and zoo animals - one becomes far less choosy re: smaller apertures and deep DOFs. It's then that one compromises for LESS noise (rather than no noise) using all the ERADR to mitigate the higher noise consequent upon use of higher ISOs.

I learned almost a decade ago that no ETTR could trump EBTR for image quality in terms of noise OR in terms of tonal spectrum throughout the range from darkest to brightest.

Dave

selmslie Loc: Fernandina Beach, FL, USA

Dave,

I suspected as much - that you had picked up this approach when digital sensors were not as good as they are today.

The camera you seem to be using (SLT-A77V) looks like a nice model with a 24MP APSC sensor that should produce better tonality and less visible nose than your camera(s) from a decade back. I can't figure out if you get a 12- or 14-bit raw file from it since Sony is a bit reticent to reveal this. From all I have read, 12-bit may be enough.

What we have been at odds about throughout this discussion is the need for the extra effort involved in ETTR and EBTR, and you have to admit, it is not as straightforward as simple metering with the caveat to keep an eye on the histogram and blinkies for potential problems and bracket when necessary.

The thing is, in low light situations, there is really not much you can do other than to try for generous exposure or live with the noise and print small. If you get noise because you could not avoid it, there are software solutions available. In daylight conditions, noise is no longer a problem unless you are shooting in deep shade or in twilight.

As for tonalities, there are plenty of value steps at all levels within three zones of middle gray. Digital has many more steps than needed above that and the only time I would ever be concerned is if the tones more than three stops below the middle were important, which they seldom are.

So if you are comfortable doing it your way, by all means, carry on. I do not see that the extra hassle will pay off.

Scotty

Uuglypher Loc: South Dakota (East River)

Scotty said: "So if you are comfortable doing it your way, by all means, carry on. I do not see that the extra hassle will pay off."
XXXXXXXXXXXXXXXXXX

Bottom line? What you call "extra hassle" I see the process of Expose Beyond the Right (EBTR) as the most simple of solutions to the significant differences between the capture of the tightly structured 8-bit JPEG image files on the one hand, and the multipotent, 12-bit or 14-bit sac o'RAW image data on the other hand.

And so I'll return the favor and grant you the comfort of that process in which you are most confident and from which you feel your image quality benefits the most.

Let's see, Scotty; does this thread make this denoument the second or third time we've agreed to disagree on this very topic? Perhaps we ought each compose a 500-word " white paper" and submit them together whenever this topic comes up in the future, flipping a coin each time to determine whose goes first on the page?
And, through use of the retrospectoscope, I acknowledge the need to apologize for the occassional lapse in civil decorum in this discussion.

If we have occasion to face-off again on this topic perhaps we ought, rather than criticize the other's technique, focus exclusively on the merits of our own technique.
Perhaps the same title would be appropriate to both our papers: " The One-and-Only Appropriate Method of Digital Image Exposure" with an independntly chosen subtitle?

Just a thought..

'Till the next Time...

Dave

selmslie Loc: Fernandina Beach, FL, USA

I am the last person to suggest that there is a single best exposure technique for available light. Any way you measure it there is a good chance that you will overlook something, misjudge it or misread or be mislead by your histogram. This can happen even after you are very familiar with one specific camera and lighting scenario.

Probably the only way that you can achieve consistent, repeatable results is in a studio with artificial lights - and plenty of power.

So I would advocate both leaving some latitude in your exposure setting as well as bracketing your shots. After all, additional digital images are free and nobody is perfect.

anotherview Loc: California

You write: "When I decrease the brightness to a pleasing level, the histogram obviously shifts to the left. Here is the rub. Comparing this adjusted photo with the one in which blinkies first occurred, both pictures have the same shadow detail, noise and comparable mid-tones and highlights. Bear in mind that I am not a pixel-peeker."

But you did not say how or where you adjusted the brightness. I presume you decreased the brightness via a control in Adobe Camera Raw. But please explain how you adjusted the brightness.

I have used ETTR for years, although not for every shot. The theory behind ETTR has to do with half all the information in the image residing in the right-most one-fifth of the image represented in the Histogram. Half of the remaining information resides in the next one-fifth segment of the image, and so on.

If you do this reduction of information all the way to the left-most segment of the Histogram, you will find only 6.5 percent of all the image information resides there.

The analog noise that the sensor and processor generate relates to the image noise in a ratio called "the signal to noise ratio." Given the same amount of noise, a strong signal will cloak any noise by its relatively tiny effect on the signal, leaving a clean signal.

So you can see with same amount of noise but a with weak signal, the noise will standout more than it does in the presence of as strong signal, and will show itself to view.

See these explanations:

http://en.wikipedia.org/wiki/Exposing_to_the_right

http://www.luminous-landscape.com/tutorials/expose-right.shtml

abc1234 Loc: Elk Grove Village, Illinois

Good question. I should have added that I adjusted the brightness in LR with the exposure control.

Let's talk about this statement of yours, "The theory behind ETTR has to do with half all the information in the image residing in the right-most one-fifth of the image represented in the Histogram. Half of the remaining information resides in the next one-fifth segment of the image, and so on."

I have a different notion about ETTR and that is when you expose so that the histogram is as far to the right as possible, then you have captured the maximum detail in all the intermediate values. How the data are distributed is unique to each shot. In practical terms, each shot has its own amount of blacks, shadows, whites and highlights. Therefore, if I understand you correctly, I disagree with your statements that half the information is in the right-most fifth of the histogram (lowercase, not a proper noun) or 6.5 percent in the left-most region.

As for noise, I accept the notion that it is constant throughout the histogram and its appearance depends upon the signal strength or the amount of light at a given bit depth. Fortunately, today's cameras produce so little noise. My analogy is that for my Canon 60D, ISO 6400 produces less noise and has more contrast than Tri-X at 400 produced grain and softness.

I hope this answers your questions.