What is the max setting folks are using with the D800 and getting acceptable pics? This may be a stupid questuion but I am trying to learn this beast.

Visit my Nikonians gallery.

I've used 6400 without any issue. When you downsample a large picture to web display size, you're removing a great deal of noise in the process. But that's also the tricky part: what's your end result? If web, then you'll be fine with anything most likely (assuming exposure is set properly). Large 36x24 prints... maybe not so much.

Acceptable for me may not be acceptable for you, though.

Like almost anything else "it depends". I shoot sports at ISO 6400 and think they are OK if I've got most of the frame, but quite limiting if heavily cropped. 3200 is much better. I've shot 12800 and they are usable as "snapshots".

But if I were doing something for real print use (i.e. that was to get near a frame) or pictures of people that someone might look closely at, or for heavy cropping, I always try to get down in the 1600 range or under, where it is VERY clean.

It also depends (at least to the tune of 2-3 stops) how much dynamic range the subject has. If your lighting is even and not over a lot of range, then you can shoot higher. If you have deep shadows and bright highlights, and it is quite dark, I try hard to keep it down under the 2000 range or so as it's there you get the best dynamic range. IT also helps in shooting higher if there are no people in it, as then you don't get that obvious plastic look in post processing.


Linwood

Comments welcomed on pictures: Http://captivephotons.com

I was just testing my D800e today with tripod and 500mm with test target at 75 ft with your question in mind. I took jpegs and viewed side by side images at actual pixels. Of course you could improve the acquisition technique if you took raw files without sharpening. (In camera, sharpening set to none).

I reviewed unsharpened jpegs and also did noise correction and mild sharpening in LR 4.3, as processing in LR is my workflow. (I typically have a RAW not jpeg workflow but I didn't want to make this too complicated).

In any case, I thought noise became personally bothersome(acceptable is subjective) and not easily corrected at ISO 2000, without beginning to get a plastic appearing correction. That is to say, ISO 1600 had essentially no noise without any loss of sharpness.

Not sure if 800 is same as 800e but it was very helpful to correct the color noise as the ISO went up. In my tests there was some magenta showing up in areas of gray, easily corrected.

What you can get away with is different than what is ideal. I found ISO 3200 no problem at all for reproducing posters.

My ideal for the 800e is ISO 200 for best low light contrast. I found details best at ISO 400.
I agree with what is written above.
David

Thanks for the thoughtful replies. I use LR 4.3 with raw. Most pics are of birds so I need to get the speed up. I use a 70-200mm f2.8 with a 2.0x on a tripod. Most pics are displayed on a calibrated projector. I will start with iso 1600 and see how it goes.
Glenn

Visit my Nikonians gallery.

Most projectors are very low resolution - amazingly so, in fact. It is a very rare - and expensive - projector that shows 4mp (ie 2048x1536 - actually that's just 3mp). Most are more like 1.2mp. A surprising proportion are still 1024x768 or well under 1mp. Most of these resolutions are similar to what you get on the web!

What's even worse is that a fairly typical 1366x768 projector usually puts its million pixels on a pretty big screen area. It is very common to see that million pixels displayed on a 92" or 100" screen, meaning that the pixel density is a paltry 17 dpi.

_____
Brian... a bicoastal Nikonian and Team Member

My gallery is online. Comments and critique welcomed any time!

>
>In any case, I thought noise became personally
>bothersome(acceptable is subjective) and not easily corrected
>at ISO 2000, without beginning to get a plastic appearing
>correction. That is to say, ISO 1600 had essentially no noise
>without any loss of sharpness.
>

>David

Hi David,
I have the same opinion and set my absolute limit of nothing above 2000, I see it visually at 2200 ISO. I usually set 1600 as my upper end. This is for wedding photography.

I see a lot of folks posting images at 3200 and 6400 on the web, but like you and others point out, downsizing cleans up a lot noise. For print, and that's what I shoot for, I limit my ISO. Anything above 800 is going to lose color fidelity. I shot my formals mostly at 640 ISO or lower.
Regards,
Craig

While noise and freezing subject speed are important, keep in mind that at higher ISO levels you are sacrificing dynamic range and color. If you have a bird with white or even black feathers, be careful about high ISO as you'll have much less latitude for correction in post.

You might have situations where you should use Manual Mode with Auto ISO to let the camera select the lowest possible ISO for your chosen aperture and shutter speed.

Eric Bowles
Nikonians Team
My Gallery
Workshops

Nikonians membership — my most important photographic investment, after the camera

I've found that the noise is not bothersome (to me) on full resolution shots up to about ISO 2000 when shooting wildlife.

Previously on D300 the "bothersome threshold" was at about ISO500 to ISO640.

Bothersome meaning that below this threshold, no special noise reduction measures are needed in post production and the pics are good for A3 prints with a little bit of levels, saturation and selective sharpening.

Visit my Nikonians gallery.

Ditto above.
I took some black birds this week. Even at iso 600 you will notice noise in the black feathers that may need correction. This assumes that you have pushed your histogram to right. If underexposed even worse.

FF... I try not to go below iso 400 if i can help it.

FF...I am a believer in tripods!

Enjoy your camera!
David

Glenn,

>What is the max setting folks are using with the D800 and getting acceptable pics?<

ISO 50 (Lo 1), but I am prepared to settle for ISO 100.

Rob
www.robbuckle.co.uk

>Glenn,
>
>>What is the max setting folks are using with the D800 and
>getting acceptable pics?<
>
>ISO 50 (Lo 1), but I am prepared to settle for ISO 100.

I never go below Base (100 ISO) due to compromised highlights.

Craig,

>I never go below Base (100 ISO) due to compromised highlights.<

I never go below base ISO either if the scene dynamic range will not fit within the reduced dynamic range of ISO Lo 1. However, when the scene dynamic range does fit, I obtain the lowest noise, sharpest photograph possible from the camera. The difference between base ISO and ISO Lo 1 is obvious.

Rob
www.robbuckle.co.uk

>Craig,
>
>>I never go below Base (100 ISO) due to compromised
>highlights.<
>
>I never go below base ISO either if the scene dynamic range
>will not fit within the reduced dynamic range of ISO Lo 1.

Could someone elaborate a bit? I realize that dynamic range is increased as you go lower in ISO, and I also realize that below the base the curve flattens out.

I always interpreted that to mean it didn't get any better, but you seem to be implying it gets worse.

I'm not sure who is more authoritative. Here's one chart:

http://home.comcast.net/~NikonD70/Charts/PDR.htm

(you probably have to select D800). The DXO charts are here:

http://www.dxomark.com/index.php/Publications/DxOMark-Reviews/Nikon-D800-Review/Sensor-performance

Those confuse me as they don't have the dots on the lines, so I'm not quite sure if they measured below 100 or not.

Anyway, my question is that if the curve is flat (not falling off toward the left), how is LO-1 actually WORSE, assuming you have adequate light for the deepest shadow you are trying to capture?

Or put another way, if I shot the same scene at 12,800 and continually slowed the shutter and lowered the ISO proportionally until I got to LO-1, I would have expected to see a growing ability to bring up the shadows and down the highlights in post processing without loosing detail, and obviously less noise as ISO went down, but that LO-1 and 100 would be about the same in regard to the dynamic range.

But you seem to be saying that's not true. That somehow on the LO-1 shot I'm going to lose some highlight detail I won't loose at 100 (allowing for appropriate shutter speed change so it is the same net exposure, and assuming there is no reciprocity failure in this range) Can you elaborate a bit?

Linwood

Comments welcomed on pictures: Http://captivephotons.com

I find the side by side comparison on the DxOMark site provides the best point of reference for base ISO and above.
http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/792%7C0/(brand)/Nikon/(appareil2)/814%7C0/(brand2)/Nikon

They only tested ISO Lo -1.0 for their ISO Sensitivity test - and found both ISO 100 and Lo -1.0 had a measured ISO of 74 - but that is probably an error. They did not test Lo -1.0 in any other test, and we all know that the actual exposure is different at Lo -1.0 than ISO 100. So I would throw out the Lo-1.0 result.

For virtually all tests, there is better or equal performance at lower ISO levels. Bill Claff has done some testing below base ISO, and found that with most cameras, base ISO and Lo-1.0 had approximately the same performance.
http://home.comcast.net/~NikonD70/Charts/PDR.htm

I have not seen any tests where there is a significant disadvantage to using Lo -1.0, but there is no advantage either.

There have been a few exceptions if you really want to split hairs. On a few cameras, the results are not completely linear at the low end and ISO 200 had lower dynamic range than ISO 170 and 230, but that is unusual and was not enough to change the way you shoot.

From my perspective, if I need a slow shutter speed I readily use Lo-1.0, but it still amounts to 2-3% of my images annually.

Eric Bowles
Nikonians Team
My Gallery
Workshops

Nikonians membership — my most important photographic investment, after the camera

Linwood,

>Could someone elaborate a bit?<

A photo shot at ISO 100 that completely spans one or more of the histograms, but is not over- or under-exposed, cannot be made to fit when shot at ISO Lo 1. In other words, the latter effectively has less dynamic range than the former. Several times I have had to change to ISO 100 from ISO Lo 1 in order to just capture the scene dynamic range without over- and under-exposing an image.

Eric,

>I have not seen any tests where there is a significant disadvantage to using Lo -1.0, but there is no advantage either.<

Sorry, but this is not correct. Shoot an image optimally at ISO 100 and again at ISO Lo 1 and the latter will clearly be sharper than the former. It's marginal, but its obvious. There is no question in my mind that my photos require less High Pass/USM sharpening when shot at ISO Lo 1 compared to ISO 100, hence the reason I shoot at the former setting whenever possible.

Rob
www.robbuckle.co.uk

>>Could someone elaborate a bit?<
>
>A photo shot at ISO 100 that completely spans one or more of
>the histograms, but is not over- or under-exposed, cannot be
>made to fit when shot at ISO Lo 1. In other words, the latter
>effectively has less dynamic range than the former. Several
>times I have had to change to ISO 100 from ISO Lo 1 in order
>to just capture the scene dynamic range without over-
>and under-exposing an image.

I get the definition you are using, but how do you know that to be true? Is it observation, or should I be able to see that on the charts being referenced?

Linwood

EDIT: PS: I am not trying to argue with you or say you are wrong; you sound very certain. I'm trying to understand how it comes to be and reconcile with the charts at that site.

Comments welcomed on pictures: Http://captivephotons.com

>Linwood,
>
>>Could someone elaborate a bit?<
>
>A photo shot at ISO 100 that completely spans one or more of
>the histograms, but is not over- or under-exposed, cannot be
>made to fit when shot at ISO Lo 1. In other words, the latter
>effectively has less dynamic range than the former. Several
>times I have had to change to ISO 100 from ISO Lo 1 in order
>to just capture the scene dynamic range without over-
>and under-exposing an image.
>
>Eric,
>
>>I have not seen any tests where there is a significant
>disadvantage to using Lo -1.0, but there is no advantage
>either.<
>
>Sorry, but this is not correct. Shoot an image optimally at
>ISO 100 and again at ISO Lo 1 and the latter will clearly be
>sharper than the former. It's marginal, but its obvious. There
>is no question in my mind that my photos require less High
>Pass/USM sharpening when shot at ISO Lo 1 compared to ISO 100,
>hence the reason I shoot at the former setting whenever
>possible.

I agree. It is sharper at ISO Lo-1 and as you point out, it is the highlights that get compromised. It is always a tradeoff and depends on what you are shooting and what you want to achieve.

The charts sighted here, tend to show no DR loss when lower than base, but I disagree. The highlights suffer.

You made a good point on a practical test, which I might do. I have a target that is black, middle grey and white. I bet it will show exactly what you point out between Base ISO and ISO lo-1 when all other things remain the same.

Regards,
Craig

>For virtually all tests, there is better or equal performance
>at lower ISO levels. Bill Claff has done some testing below
>base ISO, and found that with most cameras, base ISO and
>Lo-1.0 had approximately the same performance.
>http://home.comcast.net/~NikonD70/Charts/PDR.htm
>
>I have not seen any tests where there is a significant
>disadvantage to using Lo -1.0, but there is no advantage
>either.

I had assumed (that nasty word) that while dynamic range will not increase dropping to Lo-1.0, that noise would drop. But I don't know how to tell, frankly. At this site he has various charts, such as "Read noise in electrons vs ISO" which I would think for short exposures corresponds to noise-in-image, is that correct?

That has a flat area for the D800 below 100, but also above 1600 -- how can that be?

In fact the flat area is labeled with close circles indicating they are from digital, not analog gain control. On the D4 the closed circles are above 12,800, but for the D800 they are above about IS 1200. Not sure at all what that means, but seems to imply the D800 cuts off analog gain a lot more than one stop earlier than the D4?

So simple question - how does noise-in-image change going from 100 to Lo-1.0? Same, less noise, more noise?

Not dynamic range, but noise? Or are they in some fashion related? (because the Photographic Dynamic Range shown on one of those graphs also flattens on the low side, but not on the high side).


Linwood

Comments welcomed on pictures: Http://captivephotons.com

I've gotten good results at ISO 6400 for certain subjects. I wouldn't photograph a landscape at that ISO, but it's worked fine for wildlife and photos of people in environmental settings. Going higher than that definitely lowers quality, but you can play tricks like going to black and white, etc.

Rick Walker

My photos:
GeoVista Photography

Download from our library of Image Doctor podcasts here

Very nice posts!
Eric, I was also to believe no advantage to going lower than base. I love learning new ideas.
The ability to get more dynamic range is intriguing, such as shooting mid day. Will have to try it out!
One scenario that I use low iso frequently is with strobes that can't be made dimmer (my eilenchromes). Limited with shutter speed of 1/250 resulting in black curtain on bottom of image at 1/320. Wonder if this would be fixed with different curtain setting on flash?
Cheers!
David

Rob said,

"Sorry, but this is not correct. Shoot an image optimally at ISO 100 and again at ISO Lo 1 and the latter will clearly be sharper than the former. It's marginal, but its obvious. There is no question in my mind that my photos require less High Pass/USM sharpening when shot at ISO Lo 1 compared to ISO 100, hence the reason I shoot at the former setting whenever possible."

And less dynamic range?

Wow! Some serious voodoo magic. Will have to try it out.

Thanks for tip.
David

The DxoMark and Bill Claff tests all indicate lower ISO produces higher dynamic range, lower noise, better color, etc. The tests are backed by a methodology that is relatively established, and I generally agree with the findings.

When you get to ISO levels below base, the difference in performance is very small - if there is a benefit at all. You may also find a slight disadvantage but for those situations where you need a slower shutter speed, it's worth it.

We have one report in this thread of someone who sees something different than the data of quantitative results. There are a lot of variables involved - including camera settings and any processing of the RAW file, so I tend to place more weight on the quantitative findings of DxOMark and Bill Claff.

Noise, Dynamic Range, and Color are all different ways to analyze sensor performance. At low ISO levels, the noise is so low that it really is not a factor. You would be very hard pressed to visually see any difference between Low 1.0, ISO 100, and ISO 200. DxO Mark would say it is not significant.

It's very easy to introduce pixelation or artifacts with any kind of post processing. In particular, sharpening, contrast, shadow recovery, D-Lighting, ADL, auto levels, and saturation can all contribute more noise than the measured difference between Lo -1.0 and ISO 200.


Eric Bowles
Nikonians Team
My Gallery
Workshops

Nikonians membership — my most important photographic investment, after the camera

Wow. Who would have thought my tongue-in-cheek (but factually correct) original post would have caused such a stir?

Linwood,

>I get the definition you are using, but how do you know that to be true? Is it observation, or should I be able to see that on the charts being referenced?<

Observation. I am familiar with Bill's results (and have contributed to them), albeit the last time I looked at them was when the D800/E was released.

I first noticed the difference in ISO 100/Lo 1 quite by accident and raised the issue here, on Nikonians.

I cannot explain the cause of the results. My guess is that it simply is a reduction in noise from ISO 100 to ISO Lo 1, which increases image contrast and hence the perception of sharpness, but who knows?

As for the reduction in dynamic range, again, nothing more than observation on several occasions by me. I stand by what I have written. If others choose to not believe or accept it then that is their prerogative.

>PS: I am not trying to argue with you or say you are wrong...<

That's okay: I am not trying to 'convert' anyone. I simply invite you to run your own tests and decide for yourself. If you find me to be in error then you will have wasted no more than approximately 10 minutes of your life and I apologise in advance.

Eric,

>The DxoMark and Bill Claff tests all indicate lower ISO produces higher dynamic range, lower noise, better color, etc. The tests are backed by a methodology that is relatively established, and I generally agree with the findings.<

For the avoidance of doubt, I do not dispute Bill's findings.

>When you get to ISO levels below base, the difference in performance is very small - if there is a benefit at all.<

I have not claimed the difference is large. However, when you are a perfectionist, as I am, then a small difference is significant.

>We have one report in this thread of someone who sees something different than the data of quantitative results. There are a lot of variables involved - including camera settings and any processing of the RAW file, so I tend to place more weight on the quantitative findings of DxOMark and Bill Claff.<

>You would be very hard pressed to visually see any difference between Low 1.0, ISO 100, and ISO 200.<

Are you not able to see a difference in the photos in the link I provide above?

Honestly Eric, if you are correct then you can so easily prove me incorrect simply by shooting the two ISO images for yourself.

Rob
www.robbuckle.co.uk

I've posted a response in your other thread. I don't want to sidetrack this one.

I have no reluctance to shoot at ISO 50 (Low -1.0), but in general I use ISO 100 for landscapes and similar situations.

The D300 had an advantage to using Low -0.3 to Base ISO, but even in that case the difference was small.

The thing to remember is that higher ISO levels have lower dynamic range and color range than lower ISO levels. In every Nikon camera that is the case across almost every ISO level (there are a few ISO levels where it is not the case on some cameras). That means it is generally better to use lower ISO's when possible, and especially important for high contrast images that stretch dynamic range. Noise may be a factor, but we can manage noise so that's not the critical issue under ISO 1600 or even ISO 3200.


Eric Bowles
Nikonians Team
My Gallery
Workshops

Nikonians membership — my most important photographic investment, after the camera