DXO tests the Canon 500 II and 600 II lenses

[garyknrd]

Sorry if this has been listed before. I just saw it.
Interesting to look at anyway.
Looks like both are not near as sharp as the 300mm II lens. Which surprises me.

 

[Loswr]

Link.

Conclusion: they're good. :

 

[RGF]

Sorry if this has been listed before. I just saw it.
Interesting to look at anyway.
Looks like both are not near as sharp as the 300mm II lens. Which surprises me.

What is the relationship between DxO testing and in the field results?

Are they measuring what is important to getting a sharp, crisp picture?

 

[garyknrd]

Sorry if this has been listed before. I just saw it.
Interesting to look at anyway.
Looks like both are not near as sharp as the 300mm II lens. Which surprises me.

What is the relationship between DxO testing and in the field results?

Are they measuring what is important to getting a sharp, crisp picture?

I know they are not liked by many. But I personally really like there testing. I have found they are spot on most of the time with the few lenses and cameras I own.
After that it is whether the features are what you want on a lens or camera. Make it worth it or not.
AF, IS or what ever.
Both of those lenses are the best in class IMO.
www.flickr.com/photos/avianphotos

 

[Loswr]

Looks like both are not near as sharp as the 300mm II lens. Which surprises me.

I don't find that particularly surprising - the 300/2.8 MkI was also sharper than the 500 and 600 MkI lenses. I do find it a bit surprising that the 600 II is slightly sharper than the 500 II - that's a reversal from the MkI rank order.

Also, I wouldn't say 'not near as sharp' - the difference between the 300 II and the 600 II is 2 P-Mpix. Keep in mind that the DxOMark Score isn't a very useful metric, much like the Overall Sensor Score. DxO weights that score based on their priorities. The lens overall score is based on a low-light situation, so the 3.2 Tstops of the 300 II vs. the 4.5 Tstops of the 600 II gives the f/2.8 lens a boost in the overall score.

 

[garyknrd]

Looks like both are not near as sharp as the 300mm II lens. Which surprises me.

I don't find that particularly surprising - the 300/2.8 MkI was also sharper than the 500 and 600 MkI lenses. I do find it a bit surprising that the 600 II is slightly sharper than the 500 II - that's a reversal from the MkI rank order.

Also, I wouldn't say 'not near as sharp' - the difference between the 300 II and the 600 II is 2 P-Mpix. Keep in mind that the DxOMark Score isn't a very useful metric, much like the Overall Sensor Score. DxO weights that score based on their priorities. The lens overall score is based on a low-light situation, so the 3.2 Tstops of the 300 II vs. the 4.5 Tstops of the 600 II gives the f/2.8 lens a boost in the overall score.

Good to know, thanks for the head up on the way they measure.

Same here on the switch. 600 wins over the 500. Never would of guessed that one either after looking at the MTF charts?
Both top shelf for sure.

www.flickr.com/photos/avianphotos

 

[insanitybeard]

Looking at Canon's own MTF charts for the bare 300, 500 and 600mm lenses (the mk2 versions) the 500mm appears to have the least drop off towards the corners (in sagittal resolution if I am reading the charts correctly), less than the 300 and 600mm. I realise Canon's MTF charts are theoretical, but based on this, how come the 300mm is sharpest? It appears to have the largest drop off of the three lenses.

300mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_300mm_f_2_8l_is_ii_usm
500mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_500mm_f_4l_is_ii_usm
600mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_600mm_f_4l_is_ii_usm

 

[garyknrd]

Looking at Canon's own MTF charts for the bare 300, 500 and 600mm lenses (the mk2 versions) the 500mm appears to have the least drop off towards the corners (in sagittal resolution if I am reading the charts correctly), less than the 300 and 600mm. I realise Canon's MTF charts are theoretical, but based on this, how come the 300mm is sharpest? It appears to have the largest drop off of the three lenses.

300mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_300mm_f_2_8l_is_ii_usm
500mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_500mm_f_4l_is_ii_usm
600mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_600mm_f_4l_is_ii_usm

One of the main reasons I picked up the 500 II lens. Here is an example. I have another 500 mm prime that cannot get close to the edge sharpness of the 500 II. It really is amazing IMO.
This is on a 1.3 crop sensor. But still shows the power of the new lenses.
http://www.flickr.com/photos/avianphotos/8853424460/#sizes/h/in/photostream/

 

[Mt Spokane Photography]

Here is their explanation of the score for the 500. The 600 is a tiny bit higher.
As far as real world photos, that depends on the skill of the photographer. This is a measurement of a tool, and, doesn't measure every possible situation, nor can it anticipate the skill of a photographer to make the best of a scene, or the converse.

The 400mm f/2.8 II has a even higher score.
http://www.dxomark.com/index.php/Publications/DxOMark-Reviews/Canon-EF500mm-f-4L-IS-II-USM-and-EF600mm-f-4L-IS-II-USM-review-Giants-in-the-range/Canon-EF500mm-f-4L-IS-II-USM-and-EF600mm-f-4L-IS-II-USM-versus-competition

An overall score of 25 may seem like this lens falls short of excellence; however this is partly down to the nature of super-telephoto lenses compared to shorter telephoto lenses like an 85mm. In reality, this is a fantastic score. A sharpness score of 19-P-Mpix relative to the 22megapixels of the EOS 5D Mark III shows the lens is resolving incredibly well and crucially, there is less than 5% variation from centre to edge between f/4 and f/5.6 and less than 3% variation from f/5.6 to f/32.

 

[Loswr]

Looking at Canon's own MTF charts for the bare 300, 500 and 600mm lenses (the mk2 versions) the 500mm appears to have the least drop off towards the corners (in sagittal resolution if I am reading the charts correctly), less than the 300 and 600mm. I realise Canon's MTF charts are theoretical, but based on this, how come the 300mm is sharpest? It appears to have the largest drop off of the three lenses.

DxO measurements are not the same as an MTF curve. Also, Canon's MTF curves are theoretical, based on the lens design, whereas DxO is empirically measuring a real lens.

 

[Random Orbits]

Looking at Canon's own MTF charts for the bare 300, 500 and 600mm lenses (the mk2 versions) the 500mm appears to have the least drop off towards the corners (in sagittal resolution if I am reading the charts correctly), less than the 300 and 600mm. I realise Canon's MTF charts are theoretical, but based on this, how come the 300mm is sharpest? It appears to have the largest drop off of the three lenses.

300mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_300mm_f_2_8l_is_ii_usm
500mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_500mm_f_4l_is_ii_usm
600mm IS II: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_600mm_f_4l_is_ii_usm

The 300 f/2.8 IS II would be at a disadvantage for the thin lines because that's at max aperture, so you're comparing f/2.8 versus f/4 for the other two lenses. MTFs curves also tend to get higher for longer focal lengths as well.

But you are right in that the MTFs for the 500 are better than 600. I wonder if DxO's reversal is due to copy-to-copy variation...

 

[insanitybeard]

The 300 f/2.8 IS II would be at a disadvantage for the thin lines because that's at max aperture, so you're comparing f/2.8 versus f/4 for the other two lenses. MTFs curves also tend to get higher for longer focal lengths as well.

But you are right in that the MTFs for the 500 are better than 600. I wonder if DxO's reversal is due to copy-to-copy variation...

But as I understand it the thin dotted blue line is sagittal resolution af f/8, the 300mm still has a larger drop off towards the edges than the 500mm in that regard.

 

[Random Orbits]

The 300 f/2.8 IS II would be at a disadvantage for the thin lines because that's at max aperture, so you're comparing f/2.8 versus f/4 for the other two lenses. MTFs curves also tend to get higher for longer focal lengths as well.

But you are right in that the MTFs for the 500 are better than 600. I wonder if DxO's reversal is due to copy-to-copy variation...

But as I understand it the thin dotted blue line is sagittal resolution af f/8, the 300mm still has a larger drop off towards the edges than the 500mm in that regard.

Yes, you are correct. I do not know how MTFs translate to resolution, although higher curves should produce better resolution results. I've always been wary of trying to compare MTFs across different focal lengths and among different companies for this very reason.

 

[Loswr]

A line on the MTF curve represents data from a single spatial frequency, which isn't directly comparable to the way DxO measures sharpness.

 

[Random Orbits]

" MTFs curves also tend to get higher for longer focal lengths as well."

no they don't, longer focal lengths are measured around 10lp/mm which gives higher higher curves, for example Canons 300mm has higher resolution than 500 or 600 who are optimized for high contrast and do well at lower Lp/mm
300/2,8 and shorter lenses has higher resolution and are often measured around 30-50Lp/mm when they are compared to each other

So, there are nothing strange about that 300/2,8 are a sharper lens regarding resolution and contrast

Please explain. The Canon MTF charts have both thin and thick lines. I thought the thick lines corresponded to 10 lp/mm and that the thin lines corresponded to 30 lp/mm. We're not comparing thick from one lens to thin on the other. Unless you are saying that what the thick and thin lines represent (x lp/mm) varies depending on the Canon lens focal length. Does it?

If it does not, then what Insanitybeard points out is correct. The 300/2.8 has lower 30 lp/mm curves than the 500/4, yet it measures having higher resolution. That is the crux of the question. Neuro points out that MTF curves/sharpness do not have a 1:1 relationship, which is fine. However, we often get MTF curves in advance of lens testing results, so we often use them as proxies for sharpness, and most of time, it works well. Noting something that does not follow the pattern is interesting.

And my statement that longer focal lengths tend to produce higher MTF curves is a general trend born out by Canon's MTFs. See the F/8 lines for the 24L II, 35L, 85L II and then the telephoto primes (200 and up).

 

[ddashti]

I still feel like DxO's tests are geared more toward Nikon stuff.
They have good information, though.

 

[Loswr]

Mikeal, you were asked a direct, specific question:

Please explain. The Canon MTF charts have both thin and thick lines. I thought the thick lines corresponded to 10 lp/mm and that the thin lines corresponded to 30 lp/mm. We're not comparing thick from one lens to thin on the other. Unless you are saying that what the thick and thin lines represent (x lp/mm) varies depending on the Canon lens focal length. Does it?

You respond with:

If you look at common motive the 10Lp/mm frequency makes up 57% of the total assessment, while 20 lp/mm counts for 29%, and 40 lp/mm 14%.
So I think both Canon and Nikon knows what they are doing when they calculate what a 500mm or 600mm often be used to.

...which is a tangential response that basically repeats an earlier post but does not answer the question being asked.

 

[Don Haines]

I don't trust DxoMark and the scores that they assign to lenses....

Look at the Canon 600mm F4 while mounted on a 5D3.... Dxo rating is 26.

There are higher rated Canon lenses you can buy....

Canon 50mm F1.8 while mounted on a 5D3..... Dxo rating is 28.

That's right.... the cheapest lens Canon makes is rated higher than their most expensive lens.... OK people, who wants to swap their 600F4 for my 50F1.8? Even though the 50 is a better lens I'll swap you even....

 

[Loswr]

I don't trust DxoMark and the scores that they assign to lenses....

Look at the Canon 600mm F4 while mounted on a 5D3.... Dxo rating is 26.

There are higher rated Canon lenses you can buy....

Canon 50mm F1.8 while mounted on a 5D3..... Dxo rating is 28.

That's right.... the cheapest lens Canon makes is rated higher than their most expensive lens.... OK people, who wants to swap their 600F4 for my 50F1.8? Even though the 50 is a better lens I'll swap you even....

This is my earlier point - their Scores are useless. However, their Measurements are generally useful.

 

[Apop]

http://www.cambridgeincolour.com/tutorials/lens-quality-mtf-resolution.htm