expose to the right?

[jjj]

The proof is in the pudding - all this hypothectical wittering about various types of exposure can best be judged by looking at pictures produced by the photographers who advocate each method.

So do you think that Michael's photos are all flat, dull, poorly exposed, and uninteresting? He shoots with MFDB and DSLR, and was one of the first to write about ETTR as the optimal exposure strategy when shooting digital.
 [{POST_SNAPBACK}][/a]

And I read the articles with interest when he did so. But as I said in a post above, I was simply comparing two specific photographers involved in each side of the debate, not all photographers and not Michael. As if I had, I would have mentioned him in my post. You were commented on simply as you had offered up your images for inspection. So I was not generalizing from that, to all photographers who use ETTR. IMO, your pictures are unlikely to get people to use ETTR as I do not think they best demonstrate ETTR's virtues.
 Michael's images certainly are not flat like the ones you showed us in this link,
[a href=\"http://www.visual-vacations.com/ProfessionalServices/Portraits.htm]http://www.visual-vacations.com/Profession...s/Portraits.htm[/url]
 - your images look a bit like tone mapped images you get in weak HDR shots and are not to my taste. But I'd also say Andre's images are a nicer quality than Michael's too. Again as in most artistic judgements, that's just better to my taste and not to everyone else's.
It would save a lot of time if people read more carefully what was actually written. Heck I'm not even anti-ETTR as a concept per se, but I think it has it's place in certain situations and for certain photographers. But for many others it's a complete waste of time. Especially if you can easily produce images how you want them, without it.

[jjj]

Yes it's true, but this is adjusting in ARC. When you take the pictures you have in camera 1/3 stop. After the 4 exposure seconds I tried 1/3 more, then the highlights are clipped. ETTR is going to the right as long as you can without clipping isn't it. Well then that's it. The camera can't fine tune the exposure like ARC. Another aspect is that if you have something white in the image the the room is much smaller to expose to the right.
[a href=\"index.php?act=findpost&pid=152130\"][{POST_SNAPBACK}][/a]

Are you relying on the camera's histogram to tell you it's clipping? As that is not accurate enough for this. Keep increasing exposure by the smallest amount possible for shots up to +2 stops as Andrew says. And then examine and tweak in ACR.

[Jonathan Wienke]

Yes it's true, but this is adjusting in ARC. When you take the pictures you have in camera 1/3 stop. After the 4 exposure seconds I tried 1/3 more, then the highlights are clipped. ETTR is going to the right as long as you can without clipping isn't it. Well then that's it. The camera can't fine tune the exposure like ARC.

My experience using ACR with several different cameras is that even if there is a small amount of clipping with exposure at 0, the RAW is not necessarily clipped. Dialing in -.3 to -.5 stop exposure will often solve the clipping. When the actual RAW data is clipped, the histogram will look something like this with exposure set at -1 stop or so:



Note the little spikes on the right of the color channels, even though they are not touching the edge of the histogram. Does the histogram of the 6-second exposure look like this, even with exposure set to -1 stop?

Another aspect is that if you have something white in the image the the room is much smaller to expose to the right.

I agree.

[jing q]

I feel like I'm in dpreview forums.

I'm outta this thread now.

[samuel_js]

I thought I'll give it a try and explain my personal experience so, not being a technical person at all,  here we go.
When I was learning photography I early adopted Ansel Adam's Zone System to determine the right exposure. The ZS is a representation of the photographic scenes in 12 stops, being 0 total black and 12 pure white.
My opinion is that a normal histogram should be centered, with a little room at left and right, meaning not clipped shadows or highlights.  Like Ansel said, a good photograph should have something almost white and something almost black in the greyscale. This isn't an universal truth, is something for everyone to think about. My personal experience using a DB is the same as using B&W film when metering. That's one of the reasons for me to use a DB, the DR. The ZS method gives very accurate metering and very real representation of the scene "as visualized by the photographer. A normal contrast scene, metered to a grey card or with incident light usually gives a centered histogram. The 3seconds exposure sample I posted is an example of this. I metered both to the card and the incident light and the reading was 3 seconds at f8. If the histogram is slightly to the left could be the light changing, as It was getting dark, but still good exposure, no big deal.

Now, moving an exposure to left or right depends on a lot of factors and thats why I'm not supporter of the ETTR method. It depends on the object, the lightning, the contrast, the colors and the most important, it depends on what you want to achieve.  So what's the point on moving the histogram to the right just because is the way or a rule to do it? Are you afraid of the shadows?  

Let me put an example:
Sunny day, say about 18:00 , you are going to photograph a stone bridge from the river level: I point my sekonic  to the blue sky, just before it star to get really dark blue and place it on Zone 6,5 or 7. The shadows under the bridge will be dark but with detail. Why should I move to the histogram to the right?  The light part of the sky would be almost white, the browns of the stone desaturated and the shadows under the bridge would no longer paint the arc's curvature. Remember that photography means "paint with light"?

Another example:
A portrait. Expose to the right? My experience is flat faces, not going in more detail...

If you absolutely need to avoid noise (what I don't necessarily think is a bad thing at least in my DB), the best way is to make good exposure. Registering the contrast of the scene as it is. And having a tripod if you need.
An underexposed shoot gives you unnecessarily too much noise because it's simply a bad exposure, but in a correct exposure, there's no need, in my opinion, to move the exp, to the edge, loosing color and the drawing in the shadows.

This applies Medium Format Digital which I experience more or less like film.   A 35mm has less DR and not as good color rendition, so I don't see the point doing ETTR either.
The rule should be ER (expose right).

I see making this practice (ETTR) a routine is like to stop looking at the pictures you take. It's simply not possible. A scene containing almost black and white (like the one I posted) can't be moved too much without clipping channels. By the way I can say that a clipped photograph is not necessarily a bad one. The thing is that looking at pure white or black you see no detail or texture, you se nothing, but it can be expressively very strong.

At last I'd like to add that ETTR like HDR are just creative techniques that should be used after your photographic vision, and not as a routine. HDR is a great thing but it's sad that it's so easy discoverable. Most people abuse it instead of using it to complete their creativity.
I don't think there's rules or truth about ETTR (or HDR), is like someone telling me that I couldn't create good images if I were forced to use another brand.

It's funny, so many years of grainy photographic history and now every picture must be completely clean...

/Samuel

[AndreNapier]

BACK ON THE SUBJECT!
Since I caused some unnecessary commotion to this thread I will try to repay you guys I with more composed approach to exposure for fashion and glamour photography. I originally  joined the thread with my initial respond to the question posted.

To me the amount of over or under exposure depends on the subject of the shoot.
With female faces the major issue is the quality of skin when choosing the exposure. Perfect skin
( does it really exist anymore ) I tend to expose right on or slightly under as it will work very well in the post production. Worst the skin - lighter I go and than move it to the left. It is all personal taste but it works well for me and my clients.
Andre
[{POST_SNAPBACK}][/a]

I stand by what I said with the quote and here is detailed explanation. First of all I need to state that I do not use Adobe as a Raw converter but rather my trusty Leaf11. I would love to use ARC as it is bullet fast and time is the commodity that I have the least, but after intense scrutiny with dozens of sample files I concluded that Leaf converter delivers much nicer and smoother results. As far as intentional over exposure or ETTR with fashion shots, the problem that one will encounter is as follow. With standard zone system the average Caucasian skin puts itself around two steps above average gray. That is for properly exposed flat lid image. Fashion photographers very seldom use flat diffuse light as a key light but rather a fill. To make image more attractive and dramatic we tend to give face an extra punch by either gridded spot lite or gridded soft box.
Generally that brings a face to a 3 full levels above the medium gray. Due to curvature of human face and to the nature of light it is unavoidable that even with most meticulous make-up artist application some parts of model face will reflect more lights than the other. This extra one half to full step brings us to 3and 1/2 to 4 steps over the gray. Exposing to the right by 1 to 2 steps will bring us to a level from which human skin would never recover to normal, attractive look.
I have tried it numerous times and always ended up with color fringing and ugly discoloration that were for all practical reasons impossible to retouch even with the most intense healing process.
I hope it puts thing in prospective for many of you who are anonymously reading LL and trying to learn from it. It is not that we are against ETTR as a principle. It is a good tool for some application but if you are a people photographer you need to thing twice before using it as a rule rather than exception. When you submit your work for publication now days, most editors want to see perfectly photoshoped skin with real texture left in it. Over exposure tend to produce plasticky look that is not in fashion unless you shooting for Hair Now magazine.  Intentional overexposure is a very complex issue and we are using it to accentuate certain parts of wardrobe and to put attention to the main subject of an image. When post processing an image I almost always work with several layers of differently process Raws and over impose them at the specific lightness I envision for each detail. That is on the top of employing an average of six to eight light sources individually controlled to one image. I hope my respond will help some of you.
[a href=\"http://Http://AndreNapier.com]Http://AndreNapier.com[/url]

[bjanes]

When I was learning photography I early adopted Ansel Adam's Zone System to determine the right exposure. The ZS is a representation of the photographic scenes in 12 stops, being 0 total black and 12 pure white.
My opinion is that a normal histogram should be centered, with a little room at left and right, meaning not clipped shadows or highlights.  Like Ansel said, a good photograph should have something almost white and something almost black in the greyscale.
[{POST_SNAPBACK}][/a]

Adams' description of the zone system in The Negative, 1981, describes eleven zones, 0-X, but he regarded the useful dynamic range to be Zones I - IX, or 9 f/stops.

If you want to use the Zone system for your digital camera, you should determine its useful dynamic range. [a href=\"http://www.clarkvision.com/imagedetail/dynamicrange2/index.html]Roger Clark[/url] determined that the Canon 1D Mark II had a DR of 10 stops. Kodak rates some of their sensors used in MF backs (e.g the KAF 39000) at around 72 db or 12 f/stops; DR is limited by the noise floor and depends on how much noise is tolerable in the shadows, and you may choose a more conservative value.

You should be aware that digital has no shoulder and clips without warning and there is also no knee. Film has a logarithmic response, whereas digital is linear.

Middle gray, Zone 5, is at about a pixel value of 117 with a gamma 2.2 space (aRGB, sRGB) and 98 with a 1.8 space such as ProPhotoRGB, so you might want to place your Zone 5 luminances at that pixel value. If you use ETTR, you would place the highlights just short of clipping. I suggest the latter is more consistent with Ansel's approach. When he discusses a short scale subject (a low contrast subject that uses less than the full scale of zones), he suggests that Zone III placement for the shadow areas even though a higher placement could be used without clipping of the highlights (page 67 of The Negative), since there will be less grain and higher acutance. The proper density in the print can be achieved in printing. He concludes, "optimal image quality will be obtained for all values using the minimum exposure consistent with securing desired shadow detail"

With digital, it is just the opposite. The highlights should be placed just short of clipping. In this way, you will have less noise in the shadows. With a full scale subject, you have no room at either end of the scale and Zone placement of the midtones could be used. However, in this case it might be better to look at the histogram (once you have determined how it relates to clipping of shadows or highlights) and expose just short of clipping. Alternatively, you could take a highlight reading with your spot meter and place the highlight just short of clipping as determined by previous tests. This is what Bruce Fraser recommended. To paraphrase Adams, "optimal image quality will be obtained for all values using the maximum exposure consistent with securing desired highlight detail"

[Dustbak]

Andre, you are a true gentleman


Now, not to take this thread of-topic too much I was wondering. I agree with you that LC11 does a better job on .mos files than ACR but have you tried Iridient RD? It appears to deliver the same quality as LC11 but with the speed of ACR. Have you tried it and discarded it because you did not like the results? (I am still having some difficulties believing how fast RD is and how good the files turn out).

[samuel_js]

Adams' description of the zone system in The Negative, 1981, describes eleven zones, 0-X, but he regarded the useful dynamic range to be Zones I - IX, or 9 f/stops.
[a href=\"index.php?act=findpost&pid=152258\"][{POST_SNAPBACK}][/a]

Yes, of course, the original ZS is ten stops. I was referring to the 12 stops of a DB. I didn's explain myself so clear.

Andre, you are a true gentleman
Now, not to take this thread of-topic too much I was wondering. I agree with you that LC11 does a better job on .mos files than ACR but have you tried Iridient RD? It appears to deliver the same quality as LC11 but with the speed of ACR. Have you tried it and discarded it because you did not like the results? (I am still having some difficulties believing how fast RD is and how good the files turn out).
[a href=\"index.php?act=findpost&pid=152331\"][{POST_SNAPBACK}][/a]

I can add that I normally use Capture One and I get better files than with ARC or Lightroom too.

[Jonathan Wienke]

I stand by what I said with the quote and here is detailed explanation. First of all I need to state that I do not use Adobe as a Raw converter but rather my trusty Leaf11. I would love to use ARC as it is bullet fast and time is the commodity that I have the least, but after intense scrutiny with dozens of sample files I concluded that Leaf converter delivers much nicer and smoother results. As far as intentional over exposure or ETTR with fashion shots, the problem that one will encounter is as follow. With standard zone system the average Caucasian skin puts itself around two steps above average gray. That is for properly exposed flat lid image. Fashion photographers very seldom use flat diffuse light as a key light but rather a fill. To make image more attractive and dramatic we tend to give face an extra punch by either gridded spot lite or gridded soft box.
Generally that brings a face to a 3 full levels above the medium gray. Due to curvature of human face and to the nature of light it is unavoidable that even with most meticulous make-up artist application some parts of model face will reflect more lights than the other. This extra one half to full step brings us to 3and 1/2 to 4 steps over the gray. Exposing to the right by 1 to 2 steps will bring us to a level from which human skin would never recover to normal, attractive look.
I have tried it numerous times and always ended up with color fringing and ugly discoloration that were for all practical reasons impossible to retouch even with the most intense healing process.

Your concept of "expose to the right" is a bit exaggerated if you're thinking in terms of 1-2 stops. For most subjects, 1/3-2/3 of a stop is about all you can do before you start clipping one or more channels. The color fringing / discoloration you're referring to is probably caused by a blown red channel, and the whole point of ETTR is to push exposure as far as you can without blowing any color channels.

[jjj]

Wasn't it 1.2 stops that was the recommended amount for Andrew's setup?

[digitaldog]

Wasn't it 1.2 stops that was the recommended amount for Andrew's setup?
[a href=\"index.php?act=findpost&pid=152350\"][{POST_SNAPBACK}][/a]

In the case of THIS specific image and lighting, yes. That is, I was able to over expose (which isn't over exposure) 1 ½ stops over the incident reading while not blowing out the specular highlights.

[Tim Gray]

At what point (on the scale of dit depth) would ETTR be of no incremental value?  Obviously it's a no brainer at 8 bit - but at that level, probably equally an argument for RAW as ETTR.  At 12 bit there was still clear (to me) value in ETTR and less value, but still worth it for 14 bit.  I assume the value would be less again at 16 bit.  As a pure hypothetical, surely at 256 bits (assuming a constant dynamic range) there's little motivation to practice ETTR?  All I'm saying is that at sompe point between 8 bit and some greater bit depth, the practice becomes moot.  Based strictly on reading this extensive thread, I'd guess that 16 is at least close to the margin.

[jonstewart]

At what point (on the scale of dit depth) would ETTR be of no incremental value?  Obviously it's a no brainer at 8 bit - but at that level, probably equally an argument for RAW as ETTR.  At 12 bit there was still clear (to me) value in ETTR and less value, but still worth it for 14 bit.  I assume the value would be less again at 16 bit.  As a pure hypothetical, surely at 256 bits (assuming a constant dynamic range) there's little motivation to practice ETTR?  All I'm saying is that at sompe point between 8 bit and some greater bit depth, the practice becomes moot.  Based strictly on reading this extensive thread, I'd guess that 16 is at least close to the margin.
[a href=\"index.php?act=findpost&pid=152421\"][{POST_SNAPBACK}][/a]

This is why I suggest that you need to have some experience of a (relatively current) MFDB before you try and suggest *how* important ETTR is in practice at 16 bit.

It seems that most of those extolling it's virtues have most or all of their experience in 35mm / 12bit where (I know from experience) it is important. Please don't misinterpret this as demeaning: it's certainly not. I've only had mine for a month or so, but am still in awe of the differences between it and 35mm.

[Jonathan Wienke]

There's more to it than just the number of bits in the camera's ADC. ISO is a huge factor. When shooting at the camera's base ISO you can underexpose a stop or two below ETTR and still get an acceptable, even excellent image. But at ISO 1600, ETTR is pretty much mandatory or the noise will overwhelm the image.

[Dustbak]

There's more to it than just the number of bits in the camera's ADC. ISO is a huge factor. When shooting at the camera's base ISO you can underexpose a stop or two below ETTR and still get an acceptable, even excellent image. But at ISO 1600, ETTR is pretty much mandatory or the noise will overwhelm the image.
[a href=\"index.php?act=findpost&pid=152437\"][{POST_SNAPBACK}][/a]

But here I can say that that is something that I have noticed with all my MFDB's. When using higher ISO you better expose right (in both senses) or you will be rewarded with really ugly noise.

All other things I can only say I expose what I want to have and adjust later. I don't go for ETTR perse, only when I know that most zones are in my image otherwise I expose for the things I find important. Not sure why maybe because I am lazy and it works.

[digitaldog]

This is why I suggest that you need to have some experience of a (relatively current) MFDB before you try and suggest *how* important ETTR is in practice at 16 bit.

It seems that most of those extolling it's virtues have most or all of their experience in 35mm / 12bit where (I know from experience) it is important. Please don't misinterpret this as demeaning: it's certainly not. I've only had mine for a month or so, but am still in awe of the differences between it and 35mm.
[a href=\"index.php?act=findpost&pid=152436\"][{POST_SNAPBACK}][/a]

This is getting to be a pointless set of posts because some here seem to believe (based on what I have no idea) that somehow a larger chip or a different bit depth somehow spreads the data differently in a linear encoding. That's simply not the case. If you want to dismiss the math for religious reasons, well tell us now, so that the rest of us who refuse to enter into religious debates can use our valuable time posting elsewhere.

ETTR is only important IF you care to understand how linear encoded data is handled in all situations. No one here is, as far as I've seen, is demanding anyone expose this way 'or else', we're simply representing the facts of data capture of Raw, scene referred linear data. If you wish to under expose your capture, by all means do so. Again and for the last time, ETTR isn't over exposure, its correct exposure based on digital, linear encoding. Do you want to debate the math? Or Thomas Knoll? If so, you're on your own here. Expose the data any way you want. But PLEASE view the original question, top of the post (the one where the OP states correctly, this may be a stupid question). I usually don't believe there's such a thing as a stupid question when it is presented honestly as a question. This discussion if you can call it that is now way past fact based, one of personal opinion like what's better, a Mac or a Windows box, a Canon or a Nikon (or for this audience, a Leaf back or a Phase back).

If those of you who like to argue over and over again about religious imaging theory, not simple math, may want to continue, over on the Yahoo Color Theory list. At least you can argue about flat earth theories san's any scientific data till the cows come home.

[TorbenEskerod]

xx

[digitaldog]

Mr. Dog I think you are the religious one - not wanting to accept real life photographers experiences with MFD - you just keep hanging on to your theory based on Canon files, and simply cant get into your head that MFD is another ball game.
Torben
[a href=\"index.php?act=findpost&pid=152483\"][{POST_SNAPBACK}][/a]

It has nothing to do with Canon files. Linear encoded data is linear encoded data.

The math is undeniable. End of story.

[digitaldog]

My personal experience with the Canon 5D versus Phase P45 shooting architecture....
Just like I don’t read MTF carts when choosing a lens, I simply try different models and look for the sharpest lens using my eyes.

My Audi doesn't have an altimeter therefor its not necessary in an transportation device?

With your above statement, about the Canon/P45 and how you prefer to only use your eyes, I'm certainly happy you're not an airline pilot.