Apparent noise v aperture.

Hello.

This may be common knowledge but here goes.

I was taking some snaps of the autumn leaves and stuff yesterday afternoon.
During processing I noticed something new (to me).

The camera I was using is a 7Dii with an 85 1.2 and an iso of 800. With the lens at f1.4 there was considerably more edge/shadow noise than at f5.6 or f8.0.
I was shooting in raw with all in camera enhancements switched off.

Anyway, is it a physics thing that makes noise look worse as the aperture size increases?
Moreover, is it normal?

My thanks in advance.

Noise, or out of focus? F/1.4 and 85mm lends itself nicely to a lot of out of focus area, depending on composition and depth. A 7D MkII at ISO 800 shouldn't have much, if any, noticeable noise. Post a shot if you can.

I'll see if I can upload a comparison.

Maybe instead of noise, you could be refering to color fringing. All lenses exhibit a certain amount of chromatic aberration in areas that have strong contrast of backlighting (which can be fixed in post processing). If it's actual "noise", then it's exposure and post processing.

When I get home from work I will try and post some examples although I'm not sure what the image posting rules are.
In an ideal world I would do a series at all apertures.

I correct CA's using canons DPP v4. I also use vignette control which could possibly cause what I'm seeing.

I'm definitely (visually) seeing more noise in both the shadows and the part of the image that is in focus just by making the aperture bigger.
Asides the shutter speed changing to maintain the correct exposure the iso and all other parameters remain constant.

When you say you see noise in shadows, is that while processing in DPP? If you raise shadows, there's a possibility for seeing noise. So I don't think it's specifically "aperture"...think it might be exposure and post processing.

You can post a jpeg of the noise you're seeing. There are tabs for embedding resized jpeg files on the site.

Here are a couple of examples.
In truth, since converting them to jpeg and uploading, what I'm seeing is not as pronounced as in the raw files.
I have not applied noise reduction to the pictures.

I'd also like to point out that I don't think the iso noise is bad but it does look different across the aperture range.

http://i628.photobucke​t.com …nr%20f1.2_zpsc8​tyhezi.jpg

http://i628.photobucke​t.com …nr%20f5.6_zpsyu​ovsc4u.jpg

I think it's all about the different amount of blurred, detailless areas, where noise is more evident.

I downloaded the f/11 version. My first impression was that this is a very clean file. I ran it through LR and Noiseware, independently. There is a little bit of noise, but it is easy to work with. I'd be happy with files like those at ISO 800 from a crop camera. I have no problem using my 70D at ISO 6400. I'll use ISO 12,800 but I realize they are not going to be stellar IQ. Your 7DII seems to be about equivalent. Maybe you just need to develop the ability to pixel peep while processing, and only judge the final results. The file I looked at is more than acceptable.

With your permission, I'll post before and after screen shots of my LR edit, full frame and 100% crop.

Bassat wrote in post #17750166
I downloaded the f/11 version. My first impression was that this is a very clean file. I ran it through LR and Noiseware, independently. There is a little bit of noise, but it is easy to work with. I'd be happy with files like those at ISO 800 from a crop camera. I have no problem using my 70D at ISO 6400. I'll use ISO 12,800 but I realize they are not going to be stellar IQ. Your 7DII seems to be about equivalent. Maybe you just need to develop the ability to pixel peep while processing, and only judge the final results. The file I looked at is more than acceptable.

With your permission, I'll post before and after screen shots of my LR edit, full frame and 100% crop.

Please do.

I'd like to ask if you can see a difference in noise between 1.2 and 11 though?

Noise is controlled by exposure and iso levels. CA is noticeably different by aperture though. Noise can be accentuated, ie more noticeable, by white balance and looking closely at OOF areas. Your aperture is not directly causing any differences in noise, however slight exposure differences would explain that.

I looked at all of them online. I only downloaded the f/11, since it would have the most DOF. I am not sure if that means the noise would be more visible or less. I'd shoot that camera at 6400 without hesitation.
B/A, then 100% B/A

Canon's CMOS sensors cannot use all of the light coming in at very large apertures because of how the sensor wells work, so Canon implements a 'silent' ISO boost when you select apertures larger than f/2.8.

The effect is very small from f/2.8 to f/2.0, and quite a bit bigger from f/2 to f/1.4.

The smaller the physical size of the pixels, the more of this kind of background 'cheating' is employed. I read a study on this (I think from DxO) some time ago, and for the 7D (state of the art 1.6X Canon at the time of the test) the camera was boosting by nearly a full stop ISO at f/1.4 as compared to f/2.8.

This is a little confusing for a lot of people, so let me give an example. Suppose you take a picture at f/2.8, 1/100 and ISO 1600 and it is correctly exposed.

Now you change to f/2.0, 1/200 and ISO 1600. But because the camera cannot gather the full extra stop from f/2.8 to f/2.0, it will tell you it is shooting at ISO 1600 while silently boosting the ISO 1/3 stop and actually giving you the noise level of ISO 2000.

And imagine you change to f/1.4, 1/400 and ISO 1600. But now the camera is unable to gather much more of the light from the very large aperture, so while it continues to tell you it is shooting at ISO 1600, it will be applying a full 1 stop boost internally and giving you the noise of ISO 3200.

Here is a link to an article at Luminous Landscape about the phenomenon:
https://luminous-landscape.com …jor-camera-manufacturers/

JeffreyG wrote in post #17750354
Canon's CMOS sensors cannot use all of the light coming in at very large apertures because of how the sensor wells work, so Canon implements a 'silent' ISO boost when you select apertures larger than f/2.8.

The effect is very small from f/2.8 to f/2.0, and quite a bit bigger from f/2 to f/1.4.

The smaller the physical size of the pixels, the more of this kind of background 'cheating' is employed. I read a study on this (I think from DxO) some time ago, and for the 7D (state of the art 1.6X Canon at the time of the test) the camera was boosting by nearly a full stop ISO at f/1.4 as compared to f/2.8.

This is a little confusing for a lot of people, so let me give an example. Suppose you take a picture at f/2.8, 1/100 and ISO 1600 and it is correctly exposed.

Now you change to f/2.0, 1/200 and ISO 1600. But because the camera cannot gather the full extra stop from f/2.8 to f/2.0, it will tell you it is shooting at ISO 1600 while silently boosting the ISO 1/3 stop and actually giving you the noise level of ISO 2000.

And imagine you change to f/1.4, 1/400 and ISO 1600. But now the camera is unable to gather much more of the light from the very large aperture, so while it continues to tell you it is shooting at ISO 1600, it will be applying a full 1 stop boost internally and giving you the noise of ISO 3200.

Here is a link to an article at Luminous Landscape about the phenomenon:
https://luminous-landscape.com …jor-camera-manufacturers/

Wow, learn something new every day. Thanks Jeffrey.

rrblint wrote in post #17750405
Wow, learn something new every day. Thanks Jeffrey.

This

I'd also like to say that all of my fast primes overexpose when using a controlled lighting set up to test and calibrate my incident meters, some as much as a stop. All fast L primes, 50L, 85L, 135L more than 24L and 35L.