How to test your camera's high-ISO capability

StewartR

StewartR

Messages
11,513
Name
Stewart
Edit My Images
Yes
At this time of year lots of our clients are off in search of the Northern Lights, and that can be a demanding challenge for both camera and lens. Not least because you need to shoot at high ISO - but how high can you go before your images become too noisy? It's not something you can just look up on the internet, for two reasons. Firstly different people have different standards when it comes to the acceptability of noise. And secondly - something which isn't usually mentioned - it depends on the size at which you're viewing your photos. But fortunately it's not too hard to test for yourself. I wrote this blog post primarily for our clients, but it occurred to me that it may be useful to the TP audience too.

Comments and suggestions welcome.

http://lensesforhire.blogspot.co.uk/2014/12/northern-lights-part-1-testing-high-iso.html
 
Good piece Stewart :thumbs:

One thing about noise though, is it varies with the colour temperature of the light. If the light is warm, as most forms of artificial light are (typical of many situations where you need to crank up the ISO) then the blue channel gets much less effective exposure than red and green. When that is corrected, either in-camera with white balance or in post processing, then the blue channel has to have much more gain applied. It is 'amplified' more, effectively as if it was running at a much higher ISO, so the noise is more noticeable.

Basically, if the ambient light is warmer than normal daylight, the 'acceptable' noise level will be lower, by maybe one or two stops lower.
 
HoppyUK said:
Good piece Stewart (y)

One thing about noise though, is it varies with the colour temperature of the light. If the light is warm, as most forms of artificial light are (typical of many situations where you need to crank up the ISO) then the blue channel gets much less effective exposure than red and green. When that is corrected, either in-camera with white balance or in post processing, then the blue channel has to have much more gain applied. It is 'amplified' more, effectively as if it was running at a much higher ISO, so the noise is more noticeable.

Basically, if the ambient light is warmer than normal daylight, the 'acceptable' noise level will be lower, by maybe one or two stops lower.
Click to expand...
EXAMPLE at 25600 (D7000)

1491762_10153656846985305_2066730739_n.jpg


1538859_10153656846980305_1101473327_n.jpg
 
Phil Young said:
Colour corrected and selective NR & Sharpening.

What would you be looking to see in a cooler temperature??
Click to expand...
Less noise according to @HoppyUK or have I misunderstood?

My interpretation is that when you colour correct the warm light, you introduce more noise into the blue channel. With your example I was hoping to see a comparison of noise when shooting with warm vs cool light. I might have misinterpreted what you are trying to show, or completely missed the point that Richard was getting at...
 
the idea is blue pixels will be more starved of light in warm light, so more likely to be noisey, but thats a physical thing, so correcting wb wont help i guess?

temp in C (i e snowing temps) can reduce noise, but dont know by how much
 
Again it's one of those things where it's can sometimes depend on the photo you have shot,some photos will have great value just as long as you can see what's in the photo :)
 
StewartR said:
You've really lost me there Phil. One image has had NR applied and one hasn't. So what are you trying to show?
Click to expand...
Ah yes, should have been more clear when I started talking about NR...

The face of the bear hasn't had any NR... just a colour correction in PP. The BG has had a lot of softness applied to it.

Also. Sorry, there hasn't been any sharpening so as can be seen, it doesn't appear to have *less* noise.
 
Last edited:
OK I was going to say the noise looks much worse in the second (corrected) image.

I think to illustrate this example (not that we are calling @HoppyUK a liar or anything, I'm happy to take his word for it and I'm beginning to understand it more now) would be to take two shots at the same high(ish) ISO, one shot at a very warm colour temp (tungsten lighting for example) and one at something much closer to daylight. You can then colour correct both in post (ignoring any other processing) and compare the results. Hopefully you should be able to see more noise in the one originally shot at the warmer temperature.

I was thinking how you could do this but keep the amount of light hitting the sensor the same. I guess the best way would be to use flash and a warming gel.

I suppose the opposite is also true. If you are shooting in a very blue light, you will introduce more noise into the red channel when colour correcting. I guess it all depends how far you deviate from a "daylight" temp. Is that right? Obviously for night shots (of the aurora as given in the original example) you will want to keep the images fairly cool anyway...
 
Last edited:
Ok - I'll go take a photo now outside work in tungsten light and then the same ISO in daylight tomorrow...

*EDIT* Just tried but only have my compact camera... not really enough resolution for it.
 
Last edited:
It's good to see that somebody is addressing the issue that is almost always ignored when deciding on useful ISO settings -

Printing an image 4 foot wide and checking it out with a magnifying glass is stupid.

And so is the digital equivalent.
 
Interesting blog Stewart, shame the thread has gone off on a bit of a tangent!

What lenses do you see people hiring for the northern lights? I looked into it but in the end purchased a samyang 14mm as once I'd sold it on I had lost about 30 quid!
 
Phil Young said:
Ok - I'll go take a photo now outside work in tungsten light and then the same ISO in daylight tomorrow...

*EDIT* Just tried but only have my compact camera... not really enough resolution for it.
Click to expand...
Not to worry, apparently we are taking the thread off on a tangent!
 
Last edited:
hollis_f said:
It's good to see that somebody is addressing the issue that is almost always ignored when deciding on useful ISO settings -

Printing an image 4 foot wide and checking it out with a magnifying glass is stupid.

And so is the digital equivalent.
Click to expand...

Exactly the point I was making in a recent post on the 7D2 owner's thread, we tend to get carried away with the technicalities of the cameras rather than if they are fit for the purpose we will use them for, if I can print a shot, cropped or not, at A2 or A3, I'm happy.
 
It's also worth mentioning that the two most common enemies of noise are under-exposure and excessive cropping. Both increase noise significantly.

Attempting to increase brightness in post processing just makes everything worse. It's much better to ramp up the ISO higher in the first place, as ISO gain is applied ahead of the analogue-to-digital converter and it's more efficient.
 
Very interesting blog , I will look forward to the next installment with interest . Having seen the Northern Lights from the deck of a moving ship I didn't get the chance to take any photos but it is a spectacle I intend returning to see & and takes some photos.
 
HoppyUK said:
It's also worth mentioning that the two most common enemies of noise are under-exposure and excessive cropping. Both increase noise significantly.

Attempting to increase brightness in post processing just makes everything worse. It's much better to ramp up the ISO higher in the first place, as ISO gain is applied ahead of the analogue-to-digital converter and it's more efficient.
Click to expand...

...and using the histogram display to set your exposure is the easiest I find...
 
You must log in or register to reply here.
Back
Top