Can a lens affect noise?

Just a question which came up recently and I am wondering if bad glass can cause extra noise?

No.

Neilyb wrote in post #8340064
I am wondering if bad glass can cause extra noise?

Only if you drop it.

A poor-quality lens can be extremely soft, show distortion, or have poor contrast. It is the last shortcoming that can make sensor noise appear worse, although the noise is intrinsic to the sensor. So my answer to your question is..."sort of." It's not going to cause extra noise, but it could make existing noise appear worse as the signal (the actual image) is not as strong due to light losses/artifacts as the image passes through the lens.

A valid point was made about signal-to-noise ratio, the sensor depends on the light but the lens affects the light entering the camera...but I would argue an f5.6 lens should allow the same amount of light whatever the grade of lens? Cant imagine a lens being so bad that the light would be bad....

I see the point about contrast, I guess after PP it would become worse - but then the same would go for sharpening.

no but poor exposure can

Neilyb wrote in post #8340096
A valid point was made about signal-to-noise ratio, the sensor depends on the light but the lens affects the light entering the camera...but I would argue an f5.6 lens should allow the same amount of light whatever the grade of lens? Cant imagine a lens being so bad that the light would be bad....

I see the point about contrast, I guess after PP it would become worse - but then the same would go for sharpening.

You're confusing aperture with MTF. While MTF is affected by f-number, it is not merely the amount of light falling on the sensor that determines the quality of the signal.

I think I understand your reasoning--you're thinking in terms of noise as created as the uncertainty in the actual value of a pixel due to non-photonic effects (e.g., heat) outweighing the signal intensity received at that photosite. So for example, when too few photons hit a pixel, the signal/noise ratio is too low. This is why we see noise in dark regions of an image regardless of ISO sensitivity. Thus when you refer to f-number, you're thinking that two lenses, identical in aperture but of different optical quality, should furnish the same amount of light to the sensor. And this is true.

But what I refer to as "signal" in this case is the actual image--i.e., what we should see if the lens and sensor were ideal. Since no lens is ideal we see degradation due to things like internal reflection and other kinds of optical aberrations. This reduces things like contrast and sharpness accordingly. So a good lens might show better contrast in areas and a poor lens might show low contrast, even though both admit the same amount of light. If a lens is a poor differentiator of dark and bright areas, then the noise that shows up in the sensor regardless of what light is cast upon it will appear more obvious in the recorded result, as opposed to a lens that shows good contrast, even though the sensor's noise is the same in both instances. That's why my answer was "sort of"--because perceived noise levels are relative to the quality of the incoming light, not just the amount but also how precisely representative it is of the actual image. Poor contrast results in higher perceived noise because now the variation due to noise is more apparent relative to variations due to the actual image.

Perhaps a simpler way to think about it would be to look at a noisy image that contains areas of low detail and high detail. The low detail areas, like out-of-focus bokeh, will show noise patterns more clearly than areas where detail is high, because the high detail "overpowers" the perceived noise, even though in actuality it is more or less constant throughout the image.

Thus if a lens cannot register high detail, more of the underlying sensor noise becomes evident.

wickerprints wrote in post #8340156
Perhaps a simpler way to think about it would be to look at a noisy image that contains areas of low detail and high detail. The low detail areas, like out-of-focus bokeh, will show noise patterns more clearly than areas where detail is high, because the high detail "overpowers" the perceived noise, even though in actuality it is more or less constant throughout the image.

Thus if a lens cannot register high detail, more of the underlying sensor noise becomes evident.

Which was what I was trying to figure out, would it be severe enough to make a difference? Thankyou for your explanation.

Neilyb wrote in post #8340064
Just a question which came up recently and I am wondering if bad glass can cause extra noise?

Just one point to clarify, there are actually some older Canon lenses that will generate 'noise' in an image. I think a search on the topic will find the list. These film era EOS lens' processors / AF motors generate electromagnetic noise that shows up on digital EOS cameras.

I think Canon offered to make a change to the electronics of these lenses if you happen to have one.

wickerprints wrote in post #8340140
You're confusing aperture with MTF. While MTF is affected by f-number, it is not merely the amount of light falling on the sensor that determines the quality of the signal.

Please define what MTF means. I could not find the meaning of the acronym.

In the future, it might be a good idea to define such acronymns in the same post as where it's used.

^ (modulation transfer function) but that probably helps about as much as the acronym....

http://www.normankoren​.com/Tutorials/MTF.htm​l

TeamSpeed wrote in post #8340617
^ (modulation transfer function) but that probably helps about as much as the acronym....

http://www.normankoren​.com/Tutorials/MTF.htm​l

Thanks for the link.