Are you Shooting HAMSTTR? - ETTR - Expose to the right

HOSTED PHOTO
please log in to view hosted photos in full size.

...Is a more accurate term to describe what we really mean when we say we "Expose To The Right" or ETTR. (from this point on I will use only the term HAMSTTR. Suffice it to say, that the term is new, and prior to August 2009 the universally used term to describe HAMSTTR was "ETTR" )

A brief History of "HAMSTTR"©.

• In 2003 Michael Reichmann of Luminous Landscape fame, and early Digital Photography Pioneer, was conducting a workshop in Iceland attended by none other than Thomas Knoll ( Creator of Photoshop! )
  
  Thomas advised Michael to maximize the signal to noise ratio in Digital photography by adding what we call + EC (plus exposure compensation) while shooting RAW, and then if need be, using - EC during the RAW conversion to bring brightness levels back down to match the lighting at the time.
  Complete Article here:
  Expose to the Right - Maximizing S/N Ratio in Digital Photography
  
  
  
• Others had been seeing similar results, and very quickly word spread around that this was in fact a very effective technique.
  Further Reading:
  Roger Cavanagh, and early POTN Contributor on Expose to the right

• Total image information is increased as more of the histogram falls under that right most section.
  A 12 bit image is capable of recording 4,096 tonal values.
  The right most "stop" of the histogram (brightest) data contains 2048 of these steps — fully half of those available. Now we use 14 bit in modern DSLRs, but the effect of "HAMSTTR"© is the same.
  
  
• Noise is reduced as more of the image falls to the right side of the histogram.
  HAMSTTR will minimize the noise that potentially occurs in the darker regions of the image.

So far so good, now fast forward to August 2009 and a simple question regarding ISO settings posted on POTN results in a 20 page long debate re: use of the word "exposure"

In a nutshell, some old school purists would have it that we can not include the sensitivity of our cameras digital sensors in the "exposure" equation, because the current dictionary definition refers only to three components, ( luminance, Aperture, and shutter duration) ...not four. ISO is not part of the "triangle".
By this definition, even if your cameras' meter and histogram say otherwise, a shot taken @
1/100, f/4, ISO 100 is in fact the exact same "exposure" as a shot taken
1/100, f/4, ISO 1600

From this Epic thread;
https://photography-on-the.net/forum/showthre​ad.php?t=730218

By this definition the added sensitivity of the boosted ISO to 1600 does not alter the "exposure" but rather increases the brightness after the fact.

The trouble with this? Two fold:

1. It's true!
2. We simply do not use the word "exposure" that way any more! Call it "incorrect" but with all the terms we use in modern digital photography, EV, EC ETTR etc.. where "E" stands for Exposure in our camera metering system, all of these are effected by ISO. When in fact, by the "old school" standard, the correct definition, ISO has no impact on the "true exposure" as defined by "triangle" made up of Luminance, Shutter duration and Aperture.

In practice however, with digital (and even in film) we always include the sensitivity of the media or (ISO setting) in our calculations to make the image,. just as our equipment does this. In fact, the two tools we rely on most, absolutely include the ISO setting. More, they can not be forced to ignore it.

The Histogram and Meters in our cameras are effected by the ISO setting.
We can't escape that.

As much as all of this may seem self evident, the use of the word "exposure" in these terms seems to create a semantic paradigm.

Use vs. Definition

It was then suggested that a new term need be applied for the "other half" of our ETTR adjustments, the ISO settings. This term proffered is ITTR for "Iso To The Right"

This is fine to a point, however it draws a line that in practice the vast majority of us do not differentiate, or even recognize when shooting. It forces the use of Two Terms to do the same action, as that one action uses both techniques together, the ETTR and ITTR to push the Histogram and Meter to the right.

"HAMSTTR"© is a solution to remedy this cumbersome dichotomy.

HAMSTTR refers directly to the real world tools and setting we use to push "ETTR" and "ITTR"
We use the Cameras built in Meter and Histogram Settings to give us the "To The Right" adjustments.

For the vast majority, of ETTR shooters, HAMSTTR is what we've been doing all along. Combining both ETTR and the ISO push seamlessly in one set of adjustments to achieve one goal. An image pushed + EC to give us the most from our digital cameras and RAW files.

Further Reading on "HAMSTTR"©**

*In the following articles, please substitute "ETTR" and "Exposure" with "HAMSTTR©" and "Histogram and Meter" as all these guys are wrong.

Seriously though, these articles are MUST READ for new digital Photogs that have not been practicing HAMSTTR©, or even accomplished ones that have been but aren't sure why.

The following articles were written by some of the true pioneers of Digital imaging, hugely influential, even of you are familiar with all of there names, you've likely been following there advice for years, having been passed down now on forums by word of mouth.

Michael Reichmann on E (HAMS)* TTR July 2003
http://www.luminous-landscape.com/tutorial​s/expose-right.shtml

Roger Cavanagh on E (HAMS)* TTR Circa 2003
http://www.rogercavana​gh.com …e_right/expose_​right.html

And it's early impact on POTN with some threads I found circa 2004;

Scottes On E (HAMS)* TTR:
https://photography-on-the.net …23&highlight=ex​pose+right

A Question re: E (HAMS)* TTR:
https://photography-on-the.net …d.php?p=148310#​post148310

* In the above links I substitute "HAMS" for E as apparently Michael and Roger etc. are like me and did not realize that they can't say "E" since the meter is showing the "ISO" as well.

** The term HAMSTTR© is copyright © CyberDyneSystems 2009

Good article to raise here, and more people should be doing it. I have habitually ETTR on both the 1D and subsequent 1Ds Classic that I bought. Both of these are also traditionally rubbished as being 'very noisy' cameras. I know also that underexposure on both the 1D and 1Ds does result in a lot of noise when images are 'push-processed'.

When recommending ETTR to others I try to keep it simple: dial in positive exposure compensation even for an 'average' scene, as long as you don't blow too many highlights. Blow the highlights, and they're very difficult to retrieve on traditional sensors. The only camera that bucks this trend is the Fuji S5 (which I now also use), where it's easy to recover highlights to quite an accurate degree...this camera allows a user to push the ETTR technique to the max.

When shooting some low contrast scenes on a tripod, I would overexpose by as much as +1 1/3 stops for an average scene and still not blow any highlights on an old Canon sensor. 'Pulling' the image back in post processing then results in a very, very clean image with no noise issues even at 100%.

Excellent information for us newbs. Thanks.

Steve

Well I'm not gonna call it hamster. The mathematics involved in the whole signal/noise ratio thing are complicated enough without injecting a physics lesson.

But I will continue to practice ETTR as a matter of habit, in spite of the risks.

The risks? Well there's always a chance that there will be small areas of the image, too small to show up on the histogram or blink on your LCD, that will be bright enough to inadvertently clip some channels if you try to crank up the exposure.

There. You've been warned.

speak of the particular thread, C.D.S., it was interesting to follow in real time.

hamsttr ... mmmm.

isn't technology cool.... changing our vocabulary right in front of us.

Wonderful information. Thanks for posting this.

great article - thanks

I see this advantage,
"Total image information is increased as more of the histogram falls under that right most section."

But in low light situations where you want to achieve a sufficiant shutter speed, surely, if you can afford to overexpose slightly, or expose to the right, wouldn't you be better off bringing the iso down instead - to reduce noise?

I'm going to give it a proper read through tomorrow, as I'm just off to bed, it looks like an interesting article though, thanks for posting.

Regards, Ashley

I figure you're joking here.

Naturalworldphotograph​er wrote in post #8551444
But in low light situations where you want to achieve a sufficiant shutter speed, surely, if you can afford to overexpose slightly, or expose to the right, wouldn't you be better off bringing the iso down instead - to reduce noise?

Not in my experience. People with knowledge of the RAW/read noise levels will have to chime in, but in my own use (for example) 20D pulled ISO1600 shots were obviously better than those at a flat ISO800. If nothing else, the horizontal banding of that camera pretty much disappeared through ETTR.

When recommending ETTR to others I try to keep it simple: dial in positive exposure compensation even for an 'average' scene, as long as you don't blow too many highlights. Blow the highlights, and they're very difficult to retrieve on traditional sensors. The only camera that bucks this trend is the Fuji S5 (which I now also use), where it's easy to recover highlights to quite an accurate degree...this camera allows a user to push the ETTR technique to the max.

A better--more precise, more repeatable--way to do this is to observe the highlights that must retain detail (your artistic choice of which those are) directly and touch the spike of those tones to the right side of the histogram.

Say, you have the bride in the white dress. Zoom in to the highlighted side of her dress and take a shot. Note that spike and adjust exposure so that spike is kissing the right side of the histogram. That's it.

Here is what happens: You have deliberately placed the brightest highlight that must retain detail at the top of the sensor's range, thus being sure to preserve that highlight. At the same time, you've made sure to give the scene the fullest exposure possible (without losing that highlight), pulling as much shadow detail as possible above the base noise of the sensor.

That is the best exposure you can give the digital sensor, period.

Now, if you compare that exposure to an incident exposure, it will appear that you have overexposed by 1/2 to perhaps a full stop (hey, that's ETTR!). The JPEG image on the LCD will appear overexposed, too. But when you push it back down in processing, you will see that you have greater shadow detail and less noise than the incident meter setting would have given you.

If you need to work quicker, it's not difficult to calibrate a spot meter appropriately. In my case, I set my Sekonic L-558 to ISO 12, lay the spot on the brightest highlight that must retain detail, and get the setting I need (which works just as well with both ambient and flash lighting with that meter).

Hello. My name is Daniel, and welcome to HAMSTTR shooters anonymous.

CyberDyneSystems wrote in post #8534003
Expose to the Right - Maximizing S/N Ratio in Digital Photography

Michael is a great photographer, and his 2003 article correctly states that ETTR is good. However, I think it should be mentioned that the explanation he gives for the superiority of ETTR is incorrect. He says that it is due to the number of raw levels, but the correct reason is because it captures more light. This is proven by taking an image and removing many bits, like this example:

http://forums.dpreview​.com …rum=1018&messag​e=31239793

The highlights are barely affected by removing 5 bits. It is further demonstrated and explained here:

http://theory.uchicago​.edu/~ejm/pix/20d/test​s/noise/noise-p3.html#bitdepth

HAMSTTR combines two different methods for improving the SNR (Signal to Noise Ratio) and dynamic range of image. The most important part of HAMSTTR is light. Increasing the total amount of light (through f-number, shutter, etc.) makes a big difference.

The second part of HAMSTTR is ISO. Moving the histogram to the right with ISO (without increasing light) makes the raw levels ("ADU") higher, which makes the image brighter with a typical conversion. If it's too bright, that can be easily corrected in post processing by reducing brightness. The benefit of increasing ISO is that it has less read noise, which means the SNR is improved in parts of the image that are affected by read noise.

CyberDyneSystems wrote in post #8534003
The advantages are several.

• Total image information is increased as more of the histogram falls under that right most section.

It is useful to point out a few of the exceptions to HAMSTTR. For example, on the Canon 30D, ISO 3200 is simply a 1-stop digital "push" of ISO 1600, the same as if it had been done in post, except it is recorded to the raw file itself. In that case, the 3200 histogram is more "to the right" than 1600, but it does not have more total image information.

CyberDyneSystems wrote in post #8534003
Call it "incorrect" but the all the terms we use, EV, EC ETTR etc.. where E stands for Exposure, in our camera metering system, all of these are effected by ISO, when in fact, but this standard, ISO has no impact on the "true exposure" as defined by Luminance, Shutter duration and aperture.

I think it should be clarified that exposure can be used correctly with all of those terms (EV, EC, ETTR). They are often *affected* by ISO, but that does not mean increasing ISO is the same as increasing exposure. For example:

• To get the desired brightness, ISO 200 requires less exposure than ISO 100.
  
• At ISO 400, an exposure value (EV) of 1 results in the same brightness as EV2+ISO 800.
  
• ISO 100 with -2 EC will meter for the same exposure as ISO 200 with -1 EC, but it will not have the same brightness.

CyberDyneSystems wrote in post #8534003
In practice however, with digital (and even in film) we always include the sensitivity of the media or (ISO setting) in our calculations to make the image,. just as our equipment does this.

Excluding ISO from the definition of exposure does not mean we cannot use ISO in our calculations to make the image.

There are three examples below. The first sentence uses the colloquial definition of exposure. The second sentence uses the dictionary definition.

1. My image is too dark. I am already at the maximum f-number, shutter, and amount of light, so I am going to increase the exposure [colloq.] by changing ISO 100 to ISO 200.
   
2. My image is too dark. I am already at the maximum f-number, shutter, and amount of light, so I am going to increase the brightness by changing ISO 100 to ISO 200.

1. My shutter speed is too slow. If I increase it, the image will get too dark, I can't compensate because I'm already at the maximum f-number and amount of light, so I'm going to increase the exposure [colloq.] by changing ISO 200 to ISO 400.
   
2. My shutter speed is too slow. If I increase it, the image will get too dark, I can't compensate because I'm already at the maximum exposure, so I'm going to increase the brightness by changing ISO 200 to ISO 400.

1. My image is too noisy. I am using ISO 1600, 24mm f/2.8 lens wide open, and the slowest shutter speed I can manage. I am going to spend an extra $1,400 on a 24mm f/1.4 lens to use wide open. I will keep exposure [colloq.] the same by reducing ISO from 1600 to 400.
   
2. My image is too noisy. I am using ISO 1600, 24mm f/2.8 lens wide open, and the slowest shutter speed I can manage. I am going to spend an extra $1,400 on a 24mm f/1.4 lens to use wide open. I will keep brightness the same by reducing ISO from 1600 to 400, but I will increase the exposure by two stops thanks to f/1.4.

CyberDyneSystems wrote in post #8534003
In fact, the two tools we rely on most, absolutely include the ISO setting , in fact can not be asked to ignore it. The Histogram and Meter are effected by the ISO setting.

As photographers, the brightness of the image is very important to us. The histogram and meter are two tools that help us get the brightness we desire. In order to get the correct brightness, they must factor in the choice of ISO. That allows them to indicate what exposure to use to get the desired brightness.

But that does not mean ISO is a part of exposure.

CyberDyneSystems wrote in post #8534003
This term proffered is ITTR for "Iso To The Right"

This is fine to a point, however it draws a line that in practice the vast majority of us do not differentiate, or even recognize when shooting.

There are some ways that we do differentiate. For example, f/2.8, ISO 100, 1/125 compared with f/11, ISO 1600, 1/125. They both have the exact same histogram. The meter, too, is the same. But we know that one has much more noise. We know to first increase light (f-number and shutter) before we increase ISO.

Light is an important part of photography. The word "photography" comes from the Greek and means "drawing with light".

Many photographers will never care about how much light was used in a photograph. All they care about is how bright it is. For them, ISO and f-number are equally important, because they both affect the brightness of the image by the same amount. That is fine, and there are plenty of words for them to use to talk about that idea, such as "brightness", "darkness", and "apparent exposure".

Other photographers do care about the amount of light. They spend lots of money trying to get more light, and they spend lots of time getting the exact right amount of light. One of the biggest reasons for this is that the amount of light is the most important factor in the amount of noise.

For a raw photographer, brightness is something that can easily be changed in post. The amount of light cannot. That's why it is useful to have a word for it.

bacchanal wrote in post #8534116
I'm going to stick with the "tell the people using the old school wacky (luminance) definition of exposure to stuff it" school of thought.

I would say it the other way around. Thinking that ISO is a part of exposure is "old school". It got started back in the days of film, when the only convenient way to increase brightness was to increase exposure. (Sure, there were some inconvenient ways, such as changing the film, "pushing" chemically during processing, etc., but most people just used more exposure instead.)

This lead to many film shooters to think that exposure was the same as brightness. It wasn't really, of course, but with film the difference wasn't important. If the only method you have is exposure, then the word "exposure" might as well mean the same thing "brightness".

But digital is different. Now it is very easy to "push" (in camera or in post). Unfortunately, some people are still saddled with film-era vocabulary.

The amount of light and the amount of brightness are two separate things. I think it's better to leave the film baggage behind.

Naturalworldphotograph​er wrote in post #8551444
I see this advantage,
But in low light situations where you want to achieve a sufficiant shutter speed, surely, if you can afford to overexpose slightly, or expose to the right, wouldn't you be better off bringing the iso down instead - to reduce noise?

No. If you keep f-number+shutter the same, and increase ISO, it will improve the SNR. Here is an example using the 5D2. These are very raw. Not even white balance or demosaic.

• ISO 100 f/4 1/500
  
• vs ISO 1600 f/4 1/500

The danger of exposing to the right of the histogram is that if your white balance isn't perfect, you run the risk of blowing out one color channel and not being able to recover it, even if you shoot in RAW mode. I do shoot almost everything in RAW mode (except sports), but I still try to get the white balance as close as possible by using either a pre-set or a custom white balance. This insures that I don't blow out any one color channel and it speeds up post-processing.