Effects of diffraction and R5/R6 sensor on resolution of f/5.6, f/7.1 and f/11 lenses and TCs

[AlanF]

Another of my geek articles, which does have some implications for actual use. What I do here is to calculate the contributions of diffraction and sensor Mpx size (R5 vs R6) to the resolving power of the 400mm f/5.6 and 500mm f/7.1 zooms and the 600mm and 800mm f/11 primes and how resolution is affected by 1.4x and 2x teleconverters.

The summary based on just the effects of diffraction and sensor Mpx, and ignoring optical aberrations is:

• On an unrealistically high-resolution sensor, the f5.6, f/7.1 zooms and the 800mm f/11 would all have the same resolution and outresolve the 600mm f/11, and the TCs would not increase resolution.
• As the number of pixels on the sensor decreases, increasing focal length becomes increasingly advantageous.
• The 100-400mm at 400mm f/5.6 is slightly outresolved by the 600mm f/11 on the R5 and more so on the R6.
• The100-400mm f/5.6 + 1.4xTC at 560mm f/8 outresolves the 600mm f/11 on the R5 and the two lenses are similar on the R6
• The 100-500mm f/7.1 outresolves the 100-400mm f/5.6 and 600mm f/11 on the R5 and R6
• The 800mm f/11 outresolves the 100-500mm f/7.1 on the R5 and R6, but adding the 1.4xTC to the zoom brings it closer to the 800mm prime.
• Using the 1.4x and 2xTCs on the f/11 lens hardly increases the resolution on the R5 and only a little more on the R6. Further, any additional aberration added by the TCs could well render them even worse.
• The f/11 lenses are more suited to the lower resolution R6, and this makes it an attractive lower priced alternative.

Those who are interested can read further – I spent hours doing the calculations and graphs to confirm what I knew intuitively for my own satisfaction. But, my calculations are steering me to what lenses to use with the R5 and where the R6 is appropriate.

 

[AlanF]

In an earlier geek thread

https://www.canonrumors.com/forum/threads/diffraction-airy-disks-and-implications.36639/

there is an introduction to diffraction for non-experts, which explains what is going on here. There is a basic rule that the resolving power of a lens is limited by diffraction to being able to detect two lines separated by the diameter of an Airy Disk, d, or more, where D is the diameter of the front element and f is the focal length according to the equation:

d = 1.22*(wavelength of light)*D/f (1)
or in its usual form:
d = 1.22*(wavelength of light)*f-number (2)

This equation is used to calculate the diffraction-limited aperture of a lens, DLA, the widest f-number that makes best use of the size of pixels on the sensor. Equation 2 is independent of the focal length of the lens. But, that is not the end of the story. The size of the image on the sensor increases with f. So, if, say, you double the focal length of the lens and keep the same D, like adding a 2xTC to the lens, you double the size of the Airy Disk, so halving the resolving power. But, as the image is twice as large, you should see exactly the same amount of detail. In fact, the resolving power of the lens that is determined by diffraction depends simply on the wavelength of light and the diameter D:

Resolution(diffraction limited) varies as (wavelength of light)*D. (3)

But we know that 2x extenders give us more 2x2 more pixels on a target and usually work so we spend hard-earned money on them. So, what is going on? Equation 3 is just for the diffraction of the lens. We have to consider the diffraction of the system as a whole. To do this, we need to know the resolution of the different components, using MTF, which goes from 1, being 100%, down to zero, with MTF = 0.09 being the vale for “just resolvable” (the Rayleigh Criterion).

MTF(system) = MTF(diffraction)*MTF(sensor)*MTF(lens aberration) (4).

I have calculated the contributions of the MTF(diffraction) and the MTF(sensor) pixel size. The effects of Bayer and AA-filters on sharpness have been ignored for the sensor and won’t matter too much when comparing lenses. These are plotted versus lp/mm resolution. I have allowed for the different focal lengths, f, by reference to a 400mm lens as standard and multiplying on the x-axis the lp/mm by a factor of f/400 for each lens (i.e., x2 for 800mm etc).

The diameter of the front element (entrance pupil) for 400mm/ f5.6 is 71.4mm, 500mm f/7.1 is 70.4mm, 800mm f/11 is 72.7mm, and the 600mm f/11 is 54.4mm. The first plot shows the MTF(diffraction) lenses with D ~ 71mm have very similar curves and are better than the 600mm f/11. The line across the graph at MTF ~ 0.09 shows the Rayleigh limit where resolution is completely lost. The points where the curves intersect this line shows the relative resolution of the different lines.

These curves would be seen if the MTF(sensor) = 1.0, i.e, extremely high Mpx. There is no gain in resolution on using TCs for such an extreme sensor, and the only important factor is the diameter of the entrance pupil.

 

[AlanF]

The MTF(sensor) vs lp/mm for R5 and R6 sensors.

These curves show that the loss of resolution is progressive and gradual as the lp/mm increase, and the R6 hits the buffer of 0.09 earlier.

 

[AlanF]

Plots of MTF(diffraction)*MTF(sensor) vs lp/mm for different lenses on R5 and R6

1st Plot

• The 100-400mm at 400mm f/5.6 is slightly outresolved by the 600mm f/11 on the R5 and more so on the R6.
• The100-400mm f/5.6 + 1.4xTC at 560mm f/8 outresolves the 600mm f/11 on the R5 and the two lenses are similar on the R6
• The 100-500mm f/7.1 outresolves the 100-400mm f/5.6 and 600mm f/11 on the R5 and R6

2nd Plot

• The 800mm f/11 outresolves the 100-500mm f/7.1 on the R5 and R6, but adding the 1.4xTC to the zoom outresolves the 800mm prime.

 

[AlanF]

Adding TCs to the f/11 lenses

• Using the 1.4x and 2xTCs on the f/11 lens hardly increases the resolution on the R5 and only a little more on the R6. Further, any additional aberration added by the TCs could well render them even worse.

 

[Jonathan Thill]

@AlanF thank you for putting in the effort to do this, this is extremely valuable information for me.

I own the R5 and have the R6 on order and will be getting the 600 and the 800. My wife and I shoot together and she loves all things wildlife and really wants the R6 and the 600 and 800 because she is not the strongest little thing.

She struggles now with 100-400 and a TC on the EOS R

 

[BeenThere]

Another of my geek articles, which does have some implications for actual use. What I do here is to calculate the contributions of diffraction and sensor Mpx size (R5 vs R6) to the resolving power of the 400mm f/5.6 and 500mm f/7.1 zooms and the 600mm and 800mm f/11 primes and how resolution is affected by 1.4x and 2x teleconverters.

The summary based on just the effects of diffraction and sensor Mpx, and ignoring optical aberrations is:

• On an unrealistically high-resolution sensor, the f5.6, f/7.1 zooms and the 800mm f/11 would all have the same resolution and outresolve the 600mm f/11, and the TCs would not increase resolution.
• As the number of pixels on the sensor decreases, increasing focal length becomes increasingly advantageous.
• The 100-400mm at 400mm f/5.6 is slightly outresolved by the 600mm f/11 on the R5 and more so on the R6.
• The100-400mm f/5.6 + 1.4xTC at 560mm f/8 outresolves the 600mm f/11 on the R5 and the two lenses are similar on the R6
• The 100-500mm f/7.1 outresolves the 100-400mm f/5.6 and 600mm f/11 on the R5 and R6
• The 800mm f/11 outresolves the 100-500mm f/7.1 on the R5 and R6, but adding the 1.4xTC to the zoom outresolves the 800mm prime.
• Using the 1.4x and 2xTCs on the f/11 lens hardly increases the resolution on the R5 and only a little more on the R6. Further, any additional aberration added by the TCs could well render them even worse.
• The f/11 lenses are more suited to the lower resolution R6, and this makes it an attractive lower priced alternative.

Those who are interested can read further – I spent hours doing the calculations and graphs to confirm what I knew intuitively for my own satisfaction. But, my calculations are steering me to what lenses to use with the R5 and where the R6 is appropriate.

Great work! So, best is zooms on the R5, with TC if you need it to fill the frame.

 

[AlanF]

Great work! So, best is zooms on the R5, with TC if you need it to fill the frame.

My reaction is for me to get the R5 and use the 100-400mm II +TCs as I love that lens so much. But, the R6 with the 600 f/11 does look a nice light affordable package. But, I'd avoid the RF TCs as they are very expensive and may even make things worse. I want to try the R5 with my Sigma 150-600mm C. I think the IQ will be very good but the AF is an unknown for speed.

 

[BeenThere]

My reaction is for me to get the R5 and use the 100-400mm II +TCs as I love that lens so much. But, the R6 with the 600 f/11 does look a nice light affordable package. But, I'd avoid the RF TCs as they are very expensive and may even make things worse. I want to try the R5 with my Sigma 150-600mm C. I think the IQ will be very good but the AF is an unknown for speed.

Sounds like a plan. Also keep in mind that on occasion you may want to photograph something other than birds, and there are a lot of sharp large aperture lenses available for that.

 

[SteveC]

Sounds like a plan. Also keep in mind that on occasion you may want to photograph something other than birds, and there are a lot of sharp large aperture lenses available for that.

The holy trinity: 15-35 RF, 24-105 RF 4.0/L, and the EF 100-400 II L (with an adapter of course). The RF 100-500 L apparently is comparable, but if you've already got the EF, I see little reason to upgrade unless you must have 500mm.

(And yes, I'm still missing the 24-105, bought from the Canon store as a refurb with an RP thrown in for 300 bucks over 24-105 list price.)

 

[tron]

Alan congratulations and a big thank you for getting us valuable information about all these camera/lens/tc combinations.

 

[Click]

Alan congratulations and a big thank you for getting us valuable information about all these camera/lens/tc combinations.

+1 Thank you Alan.

 

[Stu_bert]

+1 Thank you Alan.

+2 Thanks for spending all the time to provide that info & education. Very much appreciated.

 

[justaCanonuser]

Another of my geek articles, which does have some implications for actual use. What I do here is to calculate the contributions of diffraction and sensor Mpx size (R5 vs R6) to the resolving power of the 400mm f/5.6 and 500mm f/7.1 zooms and the 600mm and 800mm f/11 primes and how resolution is affected by 1.4x and 2x teleconverters.

The summary based on just the effects of diffraction and sensor Mpx, and ignoring optical aberrations is:

• On an unrealistically high-resolution sensor, the f5.6, f/7.1 zooms and the 800mm f/11 would all have the same resolution and outresolve the 600mm f/11, and the TCs would not increase resolution.
• As the number of pixels on the sensor decreases, increasing focal length becomes increasingly advantageous.
• The 100-400mm at 400mm f/5.6 is slightly outresolved by the 600mm f/11 on the R5 and more so on the R6.
• The100-400mm f/5.6 + 1.4xTC at 560mm f/8 outresolves the 600mm f/11 on the R5 and the two lenses are similar on the R6
• The 100-500mm f/7.1 outresolves the 100-400mm f/5.6 and 600mm f/11 on the R5 and R6
• The 800mm f/11 outresolves the 100-500mm f/7.1 on the R5 and R6, but adding the 1.4xTC to the zoom brings it closer to the 800mm prime.
• Using the 1.4x and 2xTCs on the f/11 lens hardly increases the resolution on the R5 and only a little more on the R6. Further, any additional aberration added by the TCs could well render them even worse.
• The f/11 lenses are more suited to the lower resolution R6, and this makes it an attractive lower priced alternative.

Those who are interested can read further – I spent hours doing the calculations and graphs to confirm what I knew intuitively for my own satisfaction. But, my calculations are steering me to what lenses to use with the R5 and where the R6 is appropriate.

Alan, this is very interesting, thank you very much. Your last finding "the f/11 lenses are more suited to the lower resolution R6, and this makes it an attractive lower priced alternative" didn't surprise me. That's basically what I try to tell the high MP sensor fans since many years.

 

[LogicExtremist]

Another of my geek articles, which does have some implications for actual use. What I do here is to calculate the contributions of diffraction and sensor Mpx size (R5 vs R6) to the resolving power of the 400mm f/5.6 and 500mm f/7.1 zooms and the 600mm and 800mm f/11 primes and how resolution is affected by 1.4x and 2x teleconverters.

The summary based on just the effects of diffraction and sensor Mpx, and ignoring optical aberrations is:

• On an unrealistically high-resolution sensor, the f5.6, f/7.1 zooms and the 800mm f/11 would all have the same resolution and outresolve the 600mm f/11, and the TCs would not increase resolution.
• As the number of pixels on the sensor decreases, increasing focal length becomes increasingly advantageous.
• The 100-400mm at 400mm f/5.6 is slightly outresolved by the 600mm f/11 on the R5 and more so on the R6.
• The100-400mm f/5.6 + 1.4xTC at 560mm f/8 outresolves the 600mm f/11 on the R5 and the two lenses are similar on the R6
• The 100-500mm f/7.1 outresolves the 100-400mm f/5.6 and 600mm f/11 on the R5 and R6
• The 800mm f/11 outresolves the 100-500mm f/7.1 on the R5 and R6, but adding the 1.4xTC to the zoom brings it closer to the 800mm prime.
• Using the 1.4x and 2xTCs on the f/11 lens hardly increases the resolution on the R5 and only a little more on the R6. Further, any additional aberration added by the TCs could well render them even worse.
• The f/11 lenses are more suited to the lower resolution R6, and this makes it an attractive lower priced alternative.

Those who are interested can read further – I spent hours doing the calculations and graphs to confirm what I knew intuitively for my own satisfaction. But, my calculations are steering me to what lenses to use with the R5 and where the R6 is appropriate.

Hi Alan, thanks for this information and the huge effort you put into it!

 

[stevelee]

Thanks. I found that very interesting, even though I have no plans to get either of those cameras or any of those lenses.

It seems like people tend to think of diffraction as an all-or-nothing thing, as if light suddenly falls off a cliff at a diffraction limit.

I have tried taking macro pictures at various apertures to see what is the smallest opening I can get away with. As the openings stopped down, I could see an increasing softness to the image that was not unpleasant. Up to a point getting another millimeter in depth of field is worth that compromise. Even f/32 wasn’t awful though the diffraction effect was quite obvious. I clearly am not going to run similar tests to yours on my equipment, but in controlled situations, mainly macro using a tripod, I will try to see what is the optimum compromise to make.

 

[LogicExtremist]

Thanks. I found that very interesting, even though I have no plans to get either of those cameras or any of those lenses.

It seems like people tend to think of diffraction as an all-or-nothing thing, as if light suddenly falls off a cliff at a diffraction limit.

I have tried taking macro pictures at various apertures to see what is the smallest opening I can get away with. As the openings stopped down, I could see an increasing softness to the image that was not unpleasant. Up to a point getting another millimeter in depth of field is worth that compromise. Even f/32 wasn’t awful though the diffraction effect was quite obvious. I clearly am not going to run similar tests to yours on my equipment, but in controlled situations, mainly macro using a tripod, I will try to see what is the optimum compromise to make.

With macro, it's a matter of finding the balance between of depth of field, available light and degradation due to diffraction. Many macro photographers settle for the sweet spot of f/11. For static objects, it's possible to use wider apertures when using focus stacking.

 

[LogicExtremist]

Alan, I've always wondered how the real world image quality from the RF 600mm f/11 compares to the Sigma 150-600 Contemporary, curious to know what you think.

 

[stevelee]

With macro, it's a matter of finding the balance between of depth of field, available light and degradation due to diffraction. Many macro photographers settle for the sweet spot of f/11. For static objects, it's possible to use wider apertures when using focus stacking.

Yes, if I have a setup that needs stacking, I will use a focusing rail. I might still try several sets of shots at different apertures, say 8, 11, and 16, and see which turns out the best. As long as I have that all set up, it is not a lot of trouble to do more runs. I would do better if I had much more experience and shot macro more often, so I do more redundant steps to make up for judgment and skill.

 

[AlanF]

Alan, I've always wondered how the real world image quality from the RF 600mm f/11 compares to the Sigma 150-600 Contemporary, curious to know what you think.

There aren't much data out there, and I have never handled an RF 600mm f/11. So, here is my indirect opinion. Canon's own MTF values and the-digital-picture show clearly that the RF 100-500mm is significantly sharper and with better contrast at 500mm than the RF 600mm. My 100-400mm II + 1.4x TC at 560mm is pretty close to the Rf 100-500mm in sharpness and contrast. My very good copy of the Sigma 150-600mm C at 600mm had similar resolution but poorer contrast than my 100-400mm II at 560mm, and was definitely inferior to my RF 100-500mm - the RF cropped from 500 was better than the 150-600mm C at 600mm. I would guess that the weaker performance due to diffraction at f/11 of the RF 600mm is not far from the performance of the 150-600mm at f/6.3, which is n ot optimised for optical quality at the longest end. I found the AF and the IS of the Sigma much below those of the Canon lenses on the R5.