Why ETTR Is Wrong - Do This Instead

[NorthernFocus]

I agree, I think some older cameras the gain from iso was applied after exposure but before the analog to digital conversion. So it would not help shot noise but could really help read noise to set the right iso in camera.

I did a good bit of testing with the D300. At the time there was a lot of on-line chatter on the topic so I did my own testing. ISO definitely had a huge impact on noise and there was a big difference with under exposed vs ETTR images(with ISO as the variable setting). Also the D300 applied in-camera NR to the RAW file above a certain ISO setting(640?). Low/poor light photography was a real balancing act. As an aside I've gone back and re-worked a few of those files with current NR software and it's amazing what can be done.

 

[Viathelens]

Engage humor mode for this one.

Today I learned that Steve knows nothing about photography. His video got posted shell, a link to it) over on Ugly Hedgehog and there was a lot of good discussion about getting more light and all that. Then the resident troll over there…a jerk named Paul with a handle of CHG CANON piped up and pronounced that Steve knows nothing. Paul is a smart guy but unfortunately spends most (like 60-70%) of his posts just being an arrogant jerk. He did get slapped down by everyone else…but his mind is made up and he doesn’t want to be confused by the facts. I rarely post over there due to the rudeness and unfriendliness…but the jokes and funny cartoon sections are worth it.

I've deleted what I originally wrote.

 

[Viathelens]

Exposure appears to be a complex issue and most photographers, I think, believe that ISO causes noise, not something I have ever believed. I've always believed that not enough light for that specific situation and subject is what results in noise but I don't usually say that because I don't want to get negative feedback on the statement and the subject can definitely be confusing at times. I'm glad to see Steve talking about this issue. I can shoot a shot in low ISO and get noise, it is all about the light in photography.

ETTR is based on how a linear sensor collects data. The majority of the ability to collect data is on the right hand side of the histogram (not sure I've stated this quite correctly). I'm assuming a mirrorless camera is the same in this regard as a dSLR. The shadow tones have little data in them. I've included a linear sensor chart to show this. Someone said, "so we shoot as if we were shooting film?" I don't know that the statement is correct? Film was always, "shoot for the shadows and process for the highlights" (based on the limitation of film to collect highlights) and digital was "shoot for the highlights and process for the shadows," which was how slides were shot. Things are moving quickly these days in photography gear so it can be hard to decipher at times as we toss old knowledge out and make room for the new. I shoot for the highlights, setting the focus point so that it is on the brightest part of the scene, and process for the shadows. I shoot manual mode and if the light is not steady I use auto-ISO, letting the ISO "fall where it may," as Ansel would say. I'm always trying to figure exposure out and how it works! This is an interesting discussion.

You can only see EXIF info for this image if you are logged in.

 

[Anjin San]

Ha! Ha! Good to hear that someone else thinks about this guy like I do, he is one of the reasons that I seldom go to that site (there are other reasons not to view it): he's been around for many years. He's got a real desire to be "king" of UHH.

Yep…and it’s too bad because he can occasionally make pleasant and obviously well thought posts…and he’s technically astute. But…bees and honey I guess…

 

[ButlerKid]

Please refrain from making negative comments about members of other forums. BCG strives to be a friendly place offering camaraderie and great information without judging others.

 

[viper699]

Very helpful Steve, thanks!

 

[SCoombs]

I never tested it in a systematic way, but I have definitely found over time that applying ETTR in the "traditional" way tends to give me results that are worse off. At best, I find it's good to keep it in mind to avoid drastically underexposing, but it's not something I think helps as a "technical trick" like it may have used to.

 

[SCoombs]

Related to this topic of getting exposure right, has anyone else found that the Z8/9 meter tends to underexpose in natural light? When shooting with indoor lighting I find the camera's meter to be pretty spot on most of the time, but anytime I am outdoors or shooting indoors with light from a window I find that whether I am doing animals or people I need to add around +.7 to the EC to get it right.

 

[ButlerKid]

Related to this topic of getting exposure right, has anyone else found that the Z8/9 meter tends to underexpose in natural light? When shooting with indoor lighting I find the camera's meter to be pretty spot on most of the time, but anytime I am outdoors or shooting indoors with light from a window I find that whether I am doing animals or people I need to add around +.7 to the EC to get it right.

For years, Nikon cameras have been known to underexpose by 1 to 2 stops. True for DSLRs going way back. I know of some photographers that routinely set their cameras at +1 or more for all exposures. I have never found that necessary but do appreciate being able to lift exposure a bit in post when needed - and not having blown whites in other situations.

 

[JAJohnson]

Related to this topic of getting exposure right, has anyone else found that the Z8/9 meter tends to underexpose in natural light? When shooting with indoor lighting I find the camera's meter to be pretty spot on most of the time, but anytime I am outdoors or shooting indoors with light from a window I find that whether I am doing animals or people I need to add around +.7 to the EC to get it right.

Nearly all of my photography is outdoors. I’ve not found the need to routinely add EC while I’m shooting. In my experience, Z cameras expose accurately when using matrix metering.

 

[Cameron T]

Nearly all of my photography is outdoors. I’ve not found the need to routinely add EC while I’m shooting. In my experience, Z cameras expose accurately when using matrix metering.

I largely agree. I'll dial exposure comp a lot of the time and pull back on shutter speed to gain shadow detail when I pull the final image back some, but that's more a personal choice and definitely not something I feel I have to do to get the correct exposure in most situations.

 

[Viathelens]

Related to this topic of getting exposure right, has anyone else found that the Z8/9 meter tends to underexpose in natural light? When shooting with indoor lighting I find the camera's meter to be pretty spot on most of the time, but anytime I am outdoors or shooting indoors with light from a window I find that whether I am doing animals or people I need to add around +.7 to the EC to get it right.

I shoot with two Z8 Nikons, also had the Z9 for awhile and various dSLRs over time. I do not find that my Z8 underexposes in natural light. I do find that I always wish there was more natural light! I do not routinely use EC as I shoot manual for shutter and aperture and generally, if the light is not stable, auto ISO. I'm most often shooting at f/6.3 or f/5.6, my widest apertures on the two lenses most often used. Usually the camera seems to be right on target. I expose for the highlights, metering on the brightest part of the scene if possible. I do not shoot very much indoors as I'm normally photo-ing wildlife.

 

[bleirer]

Related to this topic of getting exposure right, has anyone else found that the Z8/9 meter tends to underexpose in natural light? When shooting with indoor lighting I find the camera's meter to be pretty spot on most of the time, but anytime I am outdoors or shooting indoors with light from a window I find that whether I am doing animals or people I need to add around +.7 to the EC to get it right.

That's something highlight weighted metering would do, maybe check if it is turned on?

 

[SCoombs]

As an example of what I am talking about, this photo is a pretty typical representative.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

The leftmost is the straight out of camera photo the way it metered it (matrix metering). Now this shot DOES look like some out of doors scenes might, but it's definitely not what the actual scene looked like standing there in person. The middle is a +2/3 increase in the exposure slider, and this is pretty close to what it looked like in reality or maybe a touch brighter - maybe +1/2 would have been closer. The rightmost photo is an additional edit with tweaks to shadows, highlights, etc. to achieve what I would consider a reasonable final edit.

 

[DRwyoming]

As an example of what I am talking about, this photo is a pretty typical representative.

View attachment 106673

View attachment 106674

View attachment 106688

The leftmost is the straight out of camera photo the way it metered it (matrix metering). Now this shot DOES look like some out of doors scenes might, but it's definitely not what the actual scene looked like standing there in person. The middle is a +2/3 increase in the exposure slider, and this is pretty close to what it looked like in reality or maybe a touch brighter - maybe +1/2 would have been closer. The rightmost photo is an additional edit with tweaks to shadows, highlights, etc. to achieve what I would consider a reasonable final edit.

Given the bright jacket taking up a large portion of the frame that’s exactly what one would expect. IOW, the relatively bright areas are exactly where half a stop or so of positive exposure compensation would help. The image isn’t overall neutral toned and with any large area metering I’d expect some positive exposure compensation would be needed. If it was spot metered on the brighter areas it would be more like a stop or more of positive exposure compensation.

That doesn’t mean the camera underexposed the image it means substantial portions of the image are brighter than neutral.

 

[bleirer]

If your Nikon is similar to my Canon, I'd guess it is, then in evaluative metering (matrix) the meter is weighted around the confirmed focus point. I'd say the face is pretty well rendered in the first shot.

 

[Cameron T]

If your Nikon is similar to my Canon, I'd guess it is, then in evaluative metering (matrix) the meter is weighted around the confirmed focus point. I'd say the face is pretty well rendered in the first shot.

fwiw, this was basically the discussion point when he brought it up on DPR. Everyone else agreed skin tones were fine before raising exposure, and raising exposure made the skin tones look too bright.

 

[SCoombs]

fwiw, this was basically the discussion point when he brought it up on DPR. Everyone else agreed skin tones were fine before raising exposure, and raising exposure made the skin tones look too bright.

The face in the first shot is substantially darker (by as much as around 20%) than the measures that people on DPR agreed were appropriate for skin tones.

Another, maybe more objective way of discussing this is to say that when I use Lightroom's Auto button on this photo, it puts the exposure slider to +0.68, and this is very typical for outdoor photos I have taken on this camera. Now I am not a huge fan of the auto button in LR, mainly because I find it tends to oversaturate stuff and to overbrighten the shadows indiscriminately in a way that can ruin the character of a photo, BUT it is nevertheless aiming to produce a neutral gray in a way similar to how a camera metering system does and it fairly consistently dials in something similar to +2/3 exposure for natural light shots while putting stuff shot with indoor lighting a lot closer to what the camera metered.

Here's another comparing original and LR's auto (which has the skin tones right about where the folks on DPR said they like to get them) for a photo that seems to have both some bright and some very dark areas which probably averages out to neutral or maybe just a tad brighter than neutral.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

This one seems to me to absolutely average out to less bright than neutral. Here is the original vs upping the exposure until the skin tones hit that "about 215 red channel" standard that people on DPR cited, which in this case was +1.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

 

[Jadiehl]

In this video, I’m taking a deep dive into ETTR (Expose to the Right) and exposing the real truth behind this once-popular technique.

I know, many photographers still swear by it and recommend it, but does it hold up with today’s modern camera sensors? Spoiler alert: it doesn’t—at least not the way you’ve been told or the way most people use it! In fact, it’s probably doing you more harm than good!

This video is about to blow the lid off of all the misinformation about ETTR – and it may just change the way you shoot forever!

[MEDIyoutube]4toOWWb8dp04[/MEDIA]

Thanks, Steve. This was very interesting.

 

[Lance B]

Steve, you should have been a teacher. Fantastic video as usual!
I have not been concerned all that much about ETTR either as from years of photography I found that it didn't make that much difference just like your brilliant video shows. Where possible, I have always lowered the shutter speed and open the aperture to get the best results and not been too concerned about where ISO is as long as I got the shot.