What does 'hi' means?

[EchoingWhisper]

Why would you purposefully limit your dynamic range and throw away data (the highlights that will get clipped) when you can capture it all? I believe that if you buy that D7000 you are considering, that you will change your opinion on what you will ALWAYS do pretty darn quickly!

Megan, thanks for posting the results of your tests with the D5100: it's nice to hear it also works well for this type of technique.

One sentence response: learn to expose properly and you won't blow highlights. Imagine that.

No matter how you expose properly, when using higher ISO, you wouldn't have the dynamic range of lower ISOs, so a scene exceeding the dynamic range of your selected ISO will clip the highlight or make shadow details become hidden under noise.

 

[nickzou]

Define "properly exposed". The example shown in my link above shows a scene photographed with two settings:

1/30, f2.0, ISO 100
1/30, f2.0, ISO 6400

Most people would traditionally say that only the second one was "properly exposed". In either case, the exact same amount of light hit the sensor of course. And with cameras like the D7000 that have very flat read noise curves, one can set the camera at ISO 100, capture approximately 14 stops worth of dynamic range, instead of setting it at 6400 where you'll only get about 7 stops (the camera will "clip" highlights). The jpg displayed on your on camera lcd will be terribly dark - perhaps totally black. But when you import your raw file into, say, Lightroom software and push the exposure 6 stops (while protecting highlights), the photo will reach the brightness you would have had at ISO6400, won't have blown highlights, and will have suffered no noise penalty.

In other words, with these types of sensors, in low light situations it's possible to forget about ISO when considering "exposure": leave it at iso100, pick the aperture you want for the DOF you want, and set the shutter speed you want to control motion, and then as long as you are within 6 or 8 stops, you can develop the image correctly on the computer instead of asking the camera to process it first (which is what changing the ISO on camera does, but in a less sophisticated manner than the computer can do).

WHATT?? Okay, I'm definitely trying this right now!

Please try, I'd like to see the result too.

Okay, so this is at ISO3200

And this is at ISO100 recovered. There's a fair bit of noise, probably more comparable with the ISO6400. Did I do something wrong?

And yeah before you jump all on my sh%*, this was shot on a D7000.

 

[EchoingWhisper]

WHATT?? Okay, I'm definitely trying this right now!

Please try, I'd like to see the result too.

Okay, so this is at ISO3200

And this is at ISO100 recovered. There's a fair bit of noise, probably more comparable with the ISO6400. Did I do something wrong?

And yeah before you jump all on my sh%*, this was shot on a D7000.

Lightroom really isn't the right software for this.

The "ISOless" sensor - Page 8 - PentaxForums.com

One has to be careful in pushing ISO in PP. One trap in LR is black level. If it isn't set to zero prior to pushing then the result is ugly. Moreover, Pentax clips the black at higher ISO levels which results in darker (but less noisy) shadows. Eventually, when pushing a lot, one would need a per-color option to clip the blacks which LR does not offer. W/o this option, dark tones tend to become tinted.

If done properly, there shouldn't be a visible difference between a pushed and high ISO image with the K-5.

So, the perfect PP software for "ISOless" sensors has yet to appear.

 

[EchoingWhisper]

WHATT?? Okay, I'm definitely trying this right now!

Please try, I'd like to see the result too.

Okay, so this is at ISO3200

And this is at ISO100 recovered. There's a fair bit of noise, probably more comparable with the ISO6400. Did I do something wrong?

And yeah before you jump all on my sh%*, this was shot on a D7000.

Seeing it again. I don't see any visible difference though. :/ How did you boost 5 stops in post?

 

[nickzou]

Please try, I'd like to see the result too.

Okay, so this is at ISO3200

And this is at ISO100 recovered. There's a fair bit of noise, probably more comparable with the ISO6400. Did I do something wrong?

And yeah before you jump all on my sh%*, this was shot on a D7000.

Seeing it again. I don't see any visible difference though. :/ How did you boost 5 stops in post?

Upped the exposure to +4, and then a combination of fill light and brightness until I got approximately the same historgram.

EDIT: SH*% SH*$ SH*%! I screwed this up. Hold on.

THIS is ISO3200:

THIS is ISO100 recovered:

 

[EchoingWhisper]

Okay, so this is at ISO3200

And this is at ISO100 recovered. There's a fair bit of noise, probably more comparable with the ISO6400. Did I do something wrong?

And yeah before you jump all on my sh%*, this was shot on a D7000.

Seeing it again. I don't see any visible difference though. :/ How did you boost 5 stops in post?

Upped the exposure to +4, and then a combination of fill light and brightness until I got approximately the same historgram.

EDIT: SH*% SH*$ SH*%! I screwed this up. Hold on.

THIS is ISO3200:

THIS is ISO100 recovered:

You really shouldn't use fill light, brightness etc - they are different compared to exposure.

 

[fjrabon]

I was really skeptical but I have to agree. I took 2 shots - 1@ 1/13 f/1.4 ISO 3200 and the 2nd one @ 1/13 f/1.4 ISO 100.

The shot taken at ISO 100 (after raising exposure in post) looks pretty much identical to the shot taken at ISO 3200. I don't have a d7000 but the d5100 does have the same sensor performance.

even though I'm using a different camera, with a much different sensor, that this technique has never been claimed to work on, I'm going to tell you that YOU ARE WRONG!

 

[fjrabon]

But clipping highlights is extremely different from 'clipping' shadows. You don't lose information in the same way. That's why we have the saying "the left side of the histogram holds a heckuva lot more information than does the right side." The left side only gets blocked out when the amount of photons hitting the sensor falls to an EXTREMELY low number. You can bring back a shadowed out area most of the time, if you're willing to deal with the noise. And with today's digital noise removal improvements and sensor improvements, that's going away faster and faster. Whereas a highlight, when it's gone, it's gone.

The left side of the histogram doesn't hold a heckuva lot more information than does the right side.

ETTR - The only reason to overexpose (make it look gray) a black sock is because there is more data on the right side of the histogram then the left and of course less visible noise.

Yep - you need to maximise the amount of information (bit) you get, then you ETTR. The people who suggest using base ISO did not encourage you to ETTL, because ISO is not a part of exposure. Using amplification destroys information. Please read the links I've given you.

I guess I should have worded it more carefully. I meant the EXTREME ends of the histogram. You can always pull some information out of the extreme left. You can't pull any information out of the extreme right, because it's pure white.

 

[Destin]

Someone still needs to explain how you ETTR when shooting that produces a near black frame. How you you even calculate and exposure value when you don't have an iso number to use in the calculation. You can't do it. There is a reason EVERY light meter on the market requires you to tell it what iso your camera is set to...

 

[EchoingWhisper]

If you increase exposure - remember to turn down all fill light, curves, contrast, brightness, recovery, blacks etc - they will mess your exposure.

 

[KmH]

D7000 has very low read noise - so digitally amplifying it wouldn't increase much noise, as long as there is sufficient bits/dithering. Underexposing using lower ISO doesn't decrease dynamic range - you can only clip highlights, you can't clip shadows - it can only be covered by noise. When you increase ISO, you are essentially throwing away the highlight information, you can't amplify detail that is covered by noise because there is no detail.

Did you copy that from somewhere?

 

[EchoingWhisper]

D7000 has very low read noise - so digitally amplifying it wouldn't increase much noise, as long as there is sufficient bits/dithering. Underexposing using lower ISO doesn't decrease dynamic range - you can only clip highlights, you can't clip shadows - it can only be covered by noise. When you increase ISO, you are essentially throwing away the highlight information, you can't amplify detail that is covered by noise because there is no detail.

Did you copy that from somewhere?

No I didn't. Google it.

 

[rsbones]

Someone still needs to explain how you ETTR when shooting that produces a near black frame. How you you even calculate and exposure value when you don't have an iso number to use in the calculation. You can't do it. There is a reason EVERY light meter on the market requires you to tell it what iso your camera is set to...

The thing is...you don't have to "properly" expose the image in the traditional sense because you will be deciding later (on the computer) how bright you want your image to be. Since we are talking about low light photography here, our goal is to capture as much light as possible. So...set the aperture at the absolute smallest you can to maintain the depth of field you would like. Then set the slowest shutter speed you can handle. This is what you would do normally, right? All we are skipping with this new way of shooting is the step where the iso is boosted. But as we remember, changing the iso setting is not going to change the amount of light we ultimately capture or are able to display in the finished photo.

If you want some idea of whether the aperture and shutter speed you've chosen will capture enough light to produce a decent picture, then go ahead and set your camera to iso 3200 or so and see what the camera tells you. Or use an external meter. Or take a sample shot and look on the jpg or histogram or whatever. If you're satisfied, then keep the aperture and shutter speed but just dial the iso back to 100 or 200 (some sensor engineers recommend iso200 on the D7000 for reasons that are a bit over my head to understand).

 

[Kerbouchard]

Someone still needs to explain how you ETTR when shooting that produces a near black frame. How you you even calculate and exposure value when you don't have an iso number to use in the calculation. You can't do it. There is a reason EVERY light meter on the market requires you to tell it what iso your camera is set to...

The thing is...you don't have to "properly" expose the image in the traditional sense because you will be deciding later (on the computer) how bright you want your image to be. Since we are talking about low light photography here, our goal is to capture as much light as possible. So...set the aperture at the absolute smallest you can to maintain the depth of field you would like. Then set the slowest shutter speed you can handle. This is what you would do normally, right? All we are skipping with this new way of shooting is the step where the iso is boosted. But as we remember, changing the iso setting is not going to change the amount of light we ultimately capture or are able to display in the finished photo.

If you want some idea of whether the aperture and shutter speed you've chosen will capture enough light to produce a decent picture, then go ahead and set your camera to iso 3200 or so and see what the camera tells you. Or use an external meter. Or take a sample shot and look on the jpg or histogram or whatever. If you're satisfied, then keep the aperture and shutter speed but just dial the iso back to 100 or 200 (some sensor engineers recommend iso200 on the D7000 for reasons that are a bit over my head to understand).

Except it results in less noise to amplify analog data instead of digital data. ISO amplifies analog data. Photoshop amplifies digital data. This may change in the future, but we aren't there, yet.

 

[EchoingWhisper]

Someone still needs to explain how you ETTR when shooting that produces a near black frame. How you you even calculate and exposure value when you don't have an iso number to use in the calculation. You can't do it. There is a reason EVERY light meter on the market requires you to tell it what iso your camera is set to...

The thing is...you don't have to "properly" expose the image in the traditional sense because you will be deciding later (on the computer) how bright you want your image to be. Since we are talking about low light photography here, our goal is to capture as much light as possible. So...set the aperture at the absolute smallest you can to maintain the depth of field you would like. Then set the slowest shutter speed you can handle. This is what you would do normally, right? All we are skipping with this new way of shooting is the step where the iso is boosted. But as we remember, changing the iso setting is not going to change the amount of light we ultimately capture or are able to display in the finished photo.

If you want some idea of whether the aperture and shutter speed you've chosen will capture enough light to produce a decent picture, then go ahead and set your camera to iso 3200 or so and see what the camera tells you. Or use an external meter. Or take a sample shot and look on the jpg or histogram or whatever. If you're satisfied, then keep the aperture and shutter speed but just dial the iso back to 100 or 200 (some sensor engineers recommend iso200 on the D7000 for reasons that are a bit over my head to understand).

Except it results in less noise to amplify analog data instead of digital data. ISO amplifies analog data. Photoshop amplifies digital data. This may change in the future, but we aren't there, yet.

The reason ISO have less noise is because they amplify the data before read noise is added. Photoshop amplifies read data - aka analog data + read noise - so there is more noise. But if you have a sensor with low read noise, there is no need to amplify the data (analog) before reading because there is barely any extra noise.