What is the maximum point of MP when the digital sensors start to outresolve the optical quality of current canon lenses. In other words, what is the ideal amount of MP to match the resolving power of current Canon available lenses?
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Cureent Canon lenses vs resolving power of sensor
- Thread starter Searching
- Start date
Frank Dernie
Forum Enthusiast
That depends on the lens and where in the frame you mean. All lenses vary in performance from center to edge, wide angles much more than telephotos because many aberations increase very rapidly with angle of view.
On the full size sensors the outer zone of the lens' coverage is being used so the 5D and 1DS are more lens critical than the small sensor cameras.
With the 1Ds and 5D only the best lenses approach the potential of the sensor, on the small sensor cameras more of the lenses give good results as far as resolution is concerned.
Resolution is by no means the only important lens characteristic though and flare resistance and contrast benefits show up with any sensor. If you only print up to A4 size any of the current digital SLRs are overkill.
On the full size sensors the outer zone of the lens' coverage is being used so the 5D and 1DS are more lens critical than the small sensor cameras.
With the 1Ds and 5D only the best lenses approach the potential of the sensor, on the small sensor cameras more of the lenses give good results as far as resolution is concerned.
Resolution is by no means the only important lens characteristic though and flare resistance and contrast benefits show up with any sensor. If you only print up to A4 size any of the current digital SLRs are overkill.
Steve H
Veteran Member
Many L primes can exceed 100lp/mm in resolution, so a (bayer type) sensor with a resolution of around 300pixels/mm would be required to match this. In other words around 30mp for a 1.6 crop body and 80mp for full-frame.
It's not quite so easy, though. At these sorts of resolutions, lenses start to become diffraction limited at relatively large apertures - around f/5.6. So you only get any real benefit with optically excellent lenses which can also maintain their performance below f/5.6. The big super-teles can, but not much else
Steve H
--
It's not quite so easy, though. At these sorts of resolutions, lenses start to become diffraction limited at relatively large apertures - around f/5.6. So you only get any real benefit with optically excellent lenses which can also maintain their performance below f/5.6. The big super-teles can, but not much else
Steve H
--
Steve H
Veteran Member
Hmmm. I'm not sure what I was thinking when I wrote that load of cr@p, but obviously it's wrong. Most good primes give their best around f/4-5.6 anyway. Having said that, it's still true that the resolution of a 30mp 1.6 crop sensor would only be realised at apertures below f/5.6....optically excellent lenses which can also maintain their
performance below f/5.6. The big super-teles can, but not much
else
Steve H
--
It seemed to me in reading another article that at 16MP many current Canon lenses can't out resolve the sensor of a 1 series camera. I was thinking FF here.
I was trying to get to the MP number that is an ideal match to lenses. 30 MP seems high.
I was trying to get to the MP number that is an ideal match to lenses. 30 MP seems high.
ChrisDM
Veteran Member
It depends on the lens. I'm sure that a 1DsMkII would reveal any EF lens' limitation, but most L prime lenses would hold up fine to this resolution and much more.
Chris
http://www.imagineimagery.com
Chris
http://www.imagineimagery.com
Chris
http://www.imagineimagery.com
--
Chris
http://www.imagineimagery.com
Steve H
Veteran Member
As another poster just pointed out, there is an enormous variation in the resolving power of different lenses. A lot of wides would be hard pressed to resolve 50lp/mm in the centre and a lot less in the corners. For this, 7mp on a cropper and less than 20mp on full-frame is more than enough. Most zooms are in the same ballpark. But put a 100mm macro or a 300mm/2.8L on it, and it's a whole different story.
Steve H
--
Steve H
--
MitchAlsup
Veteran Member
I would have put the mark at 24MP for the cropped camera...but we can agree that we are in the same ballpark. 24MP is just above the 100lp/mm Nyquist sampling criterion. so information at the resolution limit of the lens can be captured and recovered by sharpening.
An interesting advantage shows up when the sensor is guarenteed to outresolve the lens. That is when the lens cannot produce any feature with a spatial resolution smaller thant eh sensor, one no longer needs the AntiAlias filter. This will imporve photon throughput (efficiency) and decrease cost.It's not quite so easy, though. At these sorts of resolutions,
lenses start to become diffraction limited at relatively large
apertures - around f/5.6. So you only get any real benefit with
optically excellent lenses which can also maintain their
performance below f/5.6. The big super-teles can, but not much
else
Another advantage is that when you have a plethora of pixels, one can trade resolution for noise-supression by various binning techniques.
--
Mitch
Steve H
Veteran Member
Well, say you want a body to match a 100mm macro, 300/2.8L or something along those lines. You'll need around 25-30mp for a 1.6 cropper (assuming they still make them!) and around 70-80mp for full frame. And as Mitch points out above, it will be the first DSLR that doesn't need an AA filter to cope with the effects of the lens outresolving the body
Steve H
--
Steve H
--
dbuhanan
Leading Member
Very interesting observations, Mitch. I had not thought of that before you brought it up, but it does make very good sense. In fact, that is one of the best reasons I have heard to increase pixel density.
kindest regards
Dale
It is possible to make a small fortune in photography, you just have to start with a large fortune.
gml
Veteran Member
I don't have any scientific data to prove it but my observations are that the 20D already outresolves most lenses. So, I’d say that the maximum meaningful resolution is 8-10MP for 1.6x sensors and 20-24MP for FF sensors. Increasing sensor resolution beyond that will not result in improved image quality.
RainerRaffalski
Forum Enthusiast
According to this site
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
the sensor of a 1DSII is essentially diffrection limited around f11. The sensor of a D2X (equaling to 27 MP on FF) is diffraction limited already at f8. So a FF sensor of around 40 MP should be diffrection limited at f5.6, where the very best Canon lenses reach their maximum. (That means 16 MP for 1.6 crop.)
To reach that resolution you need about 110 lp/mm, what may be possible in the center for example with a 100macro, a 2/135 or a 2.8/300. But that's it. No more. And at f5.6 only.
OTOH in the corners already the 5D's 12.8 MP sensor seems to outresolve even the best of today's Canon wide angle lenses. And Canon's own MTF charts support that assumption.
So, to answer your question, that magic point where sensor resolution and lens resolution perfectly match is anywhere between 10 and 40 MP. It just depends on the lens you use, and whether you mean center or corner resolution.
Frank Dernie
Forum Enthusiast
I am not following your maths here - surely a FF camera would not need more than twice the pixel density than the cropped one for resolution limit. Even if lenses resolved as well into the corners as the central portion, which very few do, the requirement would be the area ratio surely?
Steve H
Veteran Member
You're right - my maths was a bit sloppy
For a given pixel density, a FF sensor will have 1.6 squared times as many pixels as a 1.6 cropper - in other words 2.56 times as many pixels. So if we agree on 25mp of resolution for a 1.6 crop, then the equivalent FF figure is 25*2.56 or 64mp. 30mp on a 1.6 crop translates to around 77mp FF.Steve H
--
where did you read that L primes surpass 100lp/mm?
the MTFs that canon supplies are done at 30lp/mm, and none of them surpass 1 (or 100% accurate reproduction; i.e. the can't reproduce, all 30lp/mm).
not calling you wrong or a liar, just asking where you got that number.
--
ed murphy ----------- AIM: monky9000
the MTFs that canon supplies are done at 30lp/mm, and none of them surpass 1 (or 100% accurate reproduction; i.e. the can't reproduce, all 30lp/mm).
not calling you wrong or a liar, just asking where you got that number.
--
ed murphy ----------- AIM: monky9000
how did you arrive at 24MP?
i am getting the same number steve arrived at (29.6 actually), assuming i use the '3' for an interpolated bayer sensor insead of the customary '2' for nyquist (with which i arriave at just 13.1MP).
and this is all assuming i use 100lp/mm of course, and i am still wondering exactly where that number came from.
thanks
ed
--
ed murphy ----------- AIM: monky9000
i am getting the same number steve arrived at (29.6 actually), assuming i use the '3' for an interpolated bayer sensor insead of the customary '2' for nyquist (with which i arriave at just 13.1MP).
and this is all assuming i use 100lp/mm of course, and i am still wondering exactly where that number came from.
thanks
ed
--
ed murphy ----------- AIM: monky9000
Steve H
Veteran Member
I guess this is the magic number which represents the extinction resolution of the very sharpest current lenses - the figure you have to beat if you don't want sensor resolution to be the limiting factor. To be honest, I've done a quick search and can't find any tests with resolutions this high, although William Castleman has tested several Canon lenses which exceed 80lp/mm.
As regards the comparison with MTF figures, the two really tell different stories. Line pair per millimetre figures represent the maximum resolution at which the lens can still reproduce a discernible pattern and probably equates to only about 5% MTF. MTF graphs on the other hand normally use a maximum 30lp/mm figure, as a lot of lenses (especially wides) just can't reproduce anything except grey mush (0% MTF) at spatial frequencies beyond say 50lp/mm. It's therefore been adopted as a sort of standard at which even less expensive lenses can also return useful figures.
Steve H
--
All the L-primes do well with a 2x TC. A 2x TC is basically equivalent to multiplying MP by 4.
--
Lee Jay
(see profile for equipment)
the way i understand MTF is that the left hand side is the "%" at 30lp/mm (or 10, depending on the line you look at). for example:
this is the MTF for the 24L. the 50% MTF should be almost exactly 30lp/mm at f/1.4. this is pretty much in line with castleman's findings... around 25lp/mm.
i guess then my issue is that i find it odd to use a figure of 100lp/mm when the best lenses are already down to 50% at 45lp/mm. by 100lp/mm those lenses will be around 10%... to me, that doesn't seem like a limit to work towards for the sensor when the lenses are already severly limited in microcontrast in that range.
ed
--
ed murphy ----------- AIM: monky9000
Keyboard shortcuts
- f
- Forum