ISO 1600 vs correcting exposure in photoshop

[Daniel Browning]

So, higher ISO is a trade off between lower noise in the shadows, but higher noise in the highlights? Or, is it the other way around?

In the context of fixed exposure (i.e. in Manual mode), it's a trade off between noise and highlight clipping.


For example, take a low light scene with these settings:


1/250, f/8, ISO 100 = 7 stops of highlight headroom (very hard to clip highlights), but lots of noise


1/250, f/8, ISO 1600 = 3 stops of highlight headroom (4 stops are clipped compared to ISO 100), but less noise.

[Dallasphotog]

There is the ever-so-slight chance that some of you guys are too smart.


I was just going to say that when things get on the dark side I push ISO up enough to get shutter speeds I can hand hold and that allow me to freeze the subject motion. From there my motto is "In Photshop I trust."


This is especially true shooting football or baseball at night where the camera is trying to correctly expose areas of the background and I really only care about the subject. I set up manually andlive with underexposed histograms. Usually, a little workwith the shadows and highlights slider in Photoshop curesall.


For weddings, I use RAW files for theversatility provided in Digital Photo Professional. Dark churches never allow an optimum combination of shutter speed, aperture and ISO.


I don't own a 50D or a 5D, but I can say thatI've never had very good images off the 1DMKII above ISO-800 and that the new 5DMKIIseems to produce revenue generating product atISO-12,800; with verynicematerial at ISO-6400. It certainly has the best image quality of the bodies I own or use including the 1DsMKIIIwe use at work.

[Chuck Lee]

iND and Daniel,


We have an excellent post going here. There is alot ofvery interest info to glean here.


It's funny as crap. I thought for all these years that if you changed the exposure increments to 1/2 it would also change the ISO increments to 1/2.... Silly Me.

By the way, Chuck, you might be interested to see the chart of dynamic range over ISO settings, made by John Sheehy

Yeah, Monkey Face Guy.... His Avatar screams "TAKE ME SERIOUSLY" I like his graph...it's nice, but I much more the enjoyed theargument/debate he was having with Gabor. Gabor is very deep like you Daniel. I see over on LL where you guys hit this topic in Jan. Lot's of good stuff there. Thanks.


When photographing, my faculties don't quite reach that deep. I just know what I know from trial and error.


I'm not seeing the headroom in post that Sheehy shows in his 5D MkII graph for the 40D or 5D.


For right nowshooting for reduced noise is more important to me than knowing what my dynamic range headroom is. I find in practice that the 40D when shooting with the ETTR tweenies can post an extra 2ev in highlights whereas the 5D consistently gives an extra1ev across all ISOs. (except 3200). Apparently the 50D and 5D MkIIfollow the 30D and 40D trend.


The 40D tweenies are the same as the 30D so if the new 5D MII is configured the same way 160,320,640 etc. would be the tweenies I'd be shooting with.


On the 5D all the results and opinions I have seen say turn the tweenies off and don't use them. That is correct. I don't.


Question: Has anyone done a dark noise histogram plot of the 5D MkII like the one done here?: Canon EOS 5D Dark Noise Tests


I'm still working with the white balance for RAW "green" histograms and will post some question/results/conclusions when more time is available.


Awesome Guys!

[Jon Ruyle]

Canon changed the way tweener ISO are handled in the 5D2. Now they are 30D-style "digital" tweeners instead of analog tweeners like the 5D Mark 1 and 1D series. That's actually a good thing, because the analog tweeners on the 1D series and 5D1 are implemented by a separate gain amplifier that adds enough of its own noise to make it just as noisy as pushing a lower ISO.

Hmmm. I wonder if they could get even lower read noise by pusing that even further... maybe get rid of iso 200, 800, 3200 and leave only 100, 400, 1600.


As an aside, I believe that read noise always decreases less than linearly. That is, you'll always get lower noise (but lower dynamic range, too) by doing a single low iso exposure than several high iso ones (ie, a single 16 sec iso 100 exposure will have less noise than 16 1 sec iso 1600 exposures averaged out or added together). Does anyone (read: Daniel) know if this is correct?

[Oren]

Ok Daniel, I've just taken 2 pictures both with the same shutter speed and aperture but one with ISO 100 and the other with ISO 1600, but not surprisingly, they look very different - exposure wise, how do you expect me to compare such different pictures?

[Daniel Browning]

Gabor is very deep like you Daniel.

Thanks. That's one of the things I really like about him. (There are some personality quirks of his that aren't as enjoyable, but no amount of faults will make up for the great service he did in building Rawnalyze.)

I'm not seeing the headroom in post that Sheehy shows in his 5D MkII graph for the 40D or 5D.

His chart is just for the green channel. As you know, most light is unbalanced (tungsten, sunset, etc.) so that can make a big difference. It could also be just the way your raw converter is showing things to you. You can eliminate that factor by opening the file in Rawnalyze. That will show you *exactly* where all the values are. (For example, you can drag a square around a card that is supposed to be middle gray, and it will tell you how many stops from *true* clipping it is, such as "-3.5".)

The 40D tweenies are the same as the 30D so if the new 5D MII is configured the same way 160,320,640 etc. would be the tweenies I'd be shooting with.

Yep. They are the same way.

Question: Has anyone done a dark noise histogram plot of the 5D MkII like the one done here?

I did for ISO 1600 and 3200 when I first got it in December, but I can't find them now. I'm sure there are some on the web somewhere. Remind me next week and I'll spend a little time shooting some blackframes. (I was trying to decide if the read noise improvements in ISO 3200 would be enough to sacrifice 1 stop of highlight headroom -- they weren't.)

[Daniel Browning]

Ok Daniel, I've just taken 2 pictures both with the same shutter speed and aperture but one with ISO 100 and the other with ISO 1600, but not surprisingly, they look very different - exposure wise, how do you expect me to compare such different pictures?

With nerdy image analysis tools, of course! Just kidding. Although the noise can be measured directly without adjusting the images, human eyeballs are better at seeing things when they are the same brightness. Just increase the brightness of the ISO 100 shot by +4 EV to match ISO 1600.

(By the way, they have the same "exposure" but different "brightness". The idea that ISO is a part of exposure is a common misconception.)

Hmmm. I wonder if they could get even lower read noise by pusing that even further... maybe get rid of iso 200, 800, 3200 and leave only 100, 400, 1600.

The thing is that all the 1-stop ISO are already performed by only one single amplifier. (Just as all the tweener ISO are performed by one amplifier, making a total of two amplifiers.)

Emil has a chart of the read noises (including tweeners) in his Noise, Dynamic Range and Bit Depth in Digital SLRs essay:

As an aside, I believe that read noise always decreases less than linearly.

Correct.

That is, you'll always get lower noise (but lower dynamic range, too) by doing a single low iso exposure than several high iso ones (ie, a single 16 sec iso 100 exposure will have less noise than 16 1 sec iso 1600 exposures averaged out or added together).

That's correct. Read noise adds in quadrature, but it still adds.

[Oren]

Yes, I meant exposure as how bright it was of course.


Anyhow, I'm not sure that boosting the brightness is a fair thing to do.

[bob williams]

Wow, This is the conversation I have been waiting for and the explanations "I thought" I needed to better understandexposure and and its realtionship with ISO and noise. Unfortunately,I nowconsider myself at least 3 stops dumber than I was when I started reading this string. Let me see if I have the basics right:


1. High ISO (up to 1600) doesn't cause noise in and of itself--


2. Use of high ISO in the wrong situation can result in noise---


3. If correctly exposed, High ISO results in less noise butcan result in blown highlights--


4. I have erroneously compared high ISO digital to high ASA film -- what was "grainy" in film does not neccesarilymean noise in digital.


Please correct or confirm these assumptions.


Now---the big question----


I normally shoot wildlife, landscape and nature with a 50d and a 100-400L or a 24-105 L (70-200 f2.8L IS is comming). I do prefer shooting within 30 minutes of sunrise or sunset---but many times I find myself in broad daylight and hopefully adjust accordingly. I normally base my settings on desired effect; i.e. Faster shutter for motion control, narrow aperture for DOF etc ---and I set ISO on lowest possible for the shooting situation----What is the error with my method of set-up.


Note--As you have probably already summised, I have discovered that sometimes I get noise when it is unexpected.


Your response is greatly appreciated,





Bob

[Daniel Browning]

You've got it right, Bob!

2. Use of high ISO in the wrong situation can result in noise---

Yes, if the high ISO causes you to reduce exposure, it will result in more noise.

Faster shutter for motion control, narrow aperture for DOF etc ---and I set ISO on lowest possible for the shooting situation----What is the error with my method of set-up.

That's fine, as long as you aren't "wasting" headroom. Low ISO gives you highlight headroom. If you don't need the highlight headroom, the best thing is to increase exposure. If you can't increase exposure, increase ISO. That is the essense of "ETTR then ITTR".