Wide angle lens desgined to be sharp at low f numbers?

As Hoppy says diffraction is a fact of life with all lenses. The quality of the glass or make or design of lens is not a factor. It is a function of the size of the hole produced by the aperture.

before the days of pixel peeping It was hardly seen as a problem by most workers.

You can find a good discussion on the subject and a calculator here http://www.tawbaware.com/forum2/viewtopic.php?t=5527&highlight=diffraction
 
Flash In The Pan said:
I think this is the main problem, Stuart. A 30 second exposure on a cheap (and presumably not so sturdy) tripod is never going to be sharp, no matter how fancy/expensive a lens is on your camera.....
Click to expand...

I know your right. I was just wondering if Defraction was constant or if there was a clever way to avoid it.

I think I have a firm grasp of tripod mechanics! :D

Stuart
 
HoppyUK said:
However, different exposure levels do produce quite dramatically different results, depending on where each particular tone sits on the histogram. The further to the right they are, the more accurate and pure they will be. Digital sensors are not linear (they are not technically digital!) and there is far more data recorded in the highlight end. This is one of the reasons to 'expose to the right'.
Click to expand...

This is getting off topic, but actually the working range of the sensor is linear, it's your perception of light that isn't. Note that the histogram is on a logarithmic x-axis.
 
Kaouthia said:
Please show me some evidence of this "fact". I'm not trying to dispute your claim, I just have never experienced it myself.

Whether we need to understand the underlying physics behind it or not, I'd like to see visual results... and then I'd like to go research the underlying physics behind it. :)
Click to expand...

I think its accepted that, for example, the colours of the Northern Lights appear more saturated on film than they do in reality, although I've never had the chance to prove it myself....and I'm not sure about digi....

And re: diffraction.....

I'd never heard of it until a couple of years ago, but again it has always been accepted that lenses are not at their best at either end of the aperture scale.
 
jerry12953 said:
I think its accepted that, for example, the colours of the Northern Lights appear more saturated on film than they do in reality, although I've never had the chance to prove it myself....and I'm not sure about digi....
Click to expand...

That's due to the fact that when you're shooting the Northern Lights, you're doing a long exposure. You're not shooting at 1/25th of a second are you? (guesstimate comparison to the eye's rate given that we can watch smooth motion on the TV @ 25fps). Even at f/2.8 on ISO100 you're look at about a 90 second exposure.

But, that's not what's in debate.

What I asked for evidence of was that the colours of a scene would somehow be different on an image shot at 15 seconds f/5.6 vs 60 seconds @ f/11. Which is what was being claimed.
 
HoppyUK said:
It is nowhere near linear, nothing to do with perception. That's the whole theory behind 'expose to the right'. Good explanation here http://www.luminous-landscape.com/tutorials/expose-right.shtml
Click to expand...

This is from your link: "Why? Because CCD and CMOS chips are linear devices."

I'm not sure what your oversight is, but it basically works like this. You can perceive roughly a doubling in light intensity. Consider this number sequence: 1, 2, 4, 8, 16, 32, 64...see how the numbers quickly increase at larger and larger steps? Consider an 8 stop range (In 8-bit, 256 discrete levels). In each successive stop you'd have this many discrete levels: 1, 2, 4, 8, 16, 32, 64, 128. Add those numbers up...256. See how half of the range is in the last stop?

The sensor response is linear, your perception isn't.
 
Mud said:
This is from your link: "Why? Because CCD and CMOS chips are linear devices."

I'm not sure what your oversight is, but it basically works like this. You can perceive roughly a doubling in light intensity. Consider this number sequence: 1, 2, 4, 8, 16, 32, 64...see how the numbers quickly increase at larger and larger steps? Consider an 8 stop range (In 8-bit, 256 discrete levels). In each successive stop you'd have this many discrete levels: 1, 2, 4, 8, 16, 32, 64, 128. Add those numbers up...256. See how half of the range is in the last stop?

The sensor response is linear, your perception isn't.
Click to expand...

I think we are arguing semantics here. I agree with your explanation, but that's not what I would describe as linear. Progressive maybe, but not linear. I also accept your comment about perception too, but I don't think it's relevant here - we perceive pretty much whatever we like according to any number of factors. It's a largely mental interpretation and to a certain extent even a personal and subjective one.

The point is that there is massively more data recorded on the right hand side of the histogram than the left. As illustrated by the table in the link above - something like twenty times more data and more, depending on the dynamic range. Furthermore, at the bottom end the signal is also diluted by noise so even progressive isn't a description that strictly holds up in practise.

This is the whole point behind exposing to the right of the histogram, in order to drag important tones out of the murk and noise to a position where there is better tone separation. I don't think we're disagreeing on that point though!
 
HoppyUK said:
I think we are arguing semantics here. I agree with your explanation, but that's not what I would describe as linear. Progressive maybe, but not linear. I also accept your comment about perception too, but I don't think it's relevant here - we perceive pretty much whatever we like according to any number of factors. It's a largely mental interpretation and to a certain extent even a personal and subjective one.

The point is that there is massively more data recorded on the right hand side of the histogram than the left. As illustrated by the table in the link above - something like twenty times more data and more, depending on the dynamic range. Furthermore, at the bottom end the signal is also diluted by noise so even progressive isn't a description that strictly holds up in practise.

This is the whole point behind exposing to the right of the histogram, in order to drag important tones out of the murk and noise to a position where there is better tone separation. I don't think we're disagreeing on that point though!
Click to expand...

The above is somewhat correct, but every time you say the sensor is non-linear you're incorrect. The sensor is a quantum detector with an effectively linear response. "not what I would describe as linear" is the bit I keep tripping over - linear is a mathematical word with a specific meaning, and though engineers will tend to slightly *******ise the meanings of mathematical words, over the working range a CCD or CMOS sensor the output scales linearly with the input. You're chewing on my words a bit...to be precise I said the sensor is linear and your perception is roughly logarithmic.

I'm not sure if you're alluding to brightness when you mention subjectivity. Yes, brightness is subjective. No, luminance isn't.
 
"colours of a scene would somehow be different on an image shot at 15 seconds f/5.6 vs 60 seconds @ f/11. Which is what was being claimed."

Eh? Who claimed that??????? I certainly didn't.

What I actually said was "With a night time shot a longer exposure can bring out more colour in the scene"

Other people have noted this too and the tendency to colour shift. If other people haven't seen this I can't do anything about that. It's a photographic effect that is available to me and I'm happy to make use of it.
 
woof woof said:
What I actually said was "With a night time shot a longer exposure can bring out more colour in the scene"

Other people have noted this too and the tendency to colour shift.
Click to expand...

If you do a longer exposure and don't change the aperture, then you will let more light into the shot.

If you adjust the aperture to compensate, you should get identical images.

You've said there's a colour shift, I want to see evidence, as I've never seen this.
 
You must log in or register to reply here.
Back
Top