I had a look at the DXOMark website to find out pixel size of E-5 sensor versus D7000 one and just discovered their ‘lens test’ section. I got curious to find out about their reports on Olympus/Panasonic glass. There are only two lenses, the Oly ZD 35mm f/3.5 macro and the ZD 14-42 f/3,5-5,6.
Looking at the score of ZD 35mm, which is a respectable lens, they report a best resolution figure of only 36 lp/mm at MTF20 which would convert into a poor 936 LWPH (36x2x13) !!!! . The bottom glass of my wine bottle will probably do better than this. By comparison the best FF tested lens (Canon EF 85mm f/1.8) reaches a resolution of 67 lp/mm at MTF20 which equates 3216 LWPH (67x2x24); while the best APS-C lenses reach the 48-49 lp/mm mark or 1540 LWPH (49x2x15,7), according to them.
Looking at Photozone test of same lens, I can read “the resolution characteristic of the lens is generally on a very high level†and the graph shows a max resolution of 2278 LWPH at MTF 50 which is more stringent than the MTF20 used by DXO.
Where I am wrong?
Posts: 6,715
Threads: 236
Joined: Apr 2010
Reputation:
22
[quote name='pesitalia' timestamp='1304284574' post='7993']
Do you mean that their procedures are flawed by some sort of error?
[/quote]
I mean that I have not seen more useless tests... lenses that really are not great get super results, and lenses with great reputations score badly.
So.. either DXOmark's lens "tests" are flawed, or the rest of the world is wrong?
Posts: 8,000
Threads: 1,850
Joined: Apr 2010
Reputation:
49
05-02-2011, 05:18 AM
(This post was last modified: 05-02-2011, 09:00 AM by Klaus.)
[quote name='pesitalia' timestamp='1304262782' post='7989']
I had a look at the DXOMark website to find out pixel size of E-5 sensor versus D7000 one and just discovered their ‘lens test’ section. I got curious to find out about their reports on Olympus/Panasonic glass. There are only two lenses, the Oly ZD 35mm f/3.5 macro and the ZD 14-42 f/3,5-5,6.
Looking at the score of ZD 35mm, which is a respectable lens, they report a best resolution figure of only 36 lp/mm at MTF20 which would convert into a poor 936 LWPH (36x2x13) !!!! . The bottom glass of my wine bottle will probably do better than this. By comparison the best FF tested lens (Canon EF 85mm f/1.8) reaches a resolution of 67 lp/mm at MTF20 which equates 3216 LWPH (67x2x24); while the best APS-C lenses reach the 48-49 lp/mm mark or 1540 LWPH (49x2x15,7), according to them.
Looking at Photozone test of same lens, I can read “the resolution characteristic of the lens is generally on a very high level†and the graph shows a max resolution of 2278 LWPH at MTF 50 which is more stringent than the MTF20 used by DXO.
Where I am wrong?
[/quote]
I reckon that they are depend on the same constraints like us here.
I measured the performance of the E-5 at some time in order to check its effective resolution against the L10. It merely delivered 2000LW/PH. I reckon the full format tests are at least performed on the EOS 5D II over there and it can deliver the double (actually quadruple) resolution.
In any case - as I mention frequently here - cross-system comparisons based on different DSLRs are INVALID.
I think the resolution measured is accurate there because their reviews on Canon or Nikon lens agree well with PZ, lenstip, etc.....
[quote name='mousefire' timestamp='1304415085' post='8016']
I think the resolution measured is accurate there because their reviews on Canon or Nikon lens agree well with PZ, lenstip, etc.....
[/quote]
For FF lenses it could be, but APS-C and FT tests do not agree with other sites because they apply that stupid crop factor (I understood that they resize lens resolution to FF format, that is, they take measured resolution in lp/mm and divide it by the crop factor to extrapolate equivalent FF resolution or, in other words, they take into consideration the required enlargment ratio to get a given size print from all formats). Obviously, smaller the sensor worst the relative resolution, but this would also apply to FF when compared to medium format. Basically their measurements are NOT reflecting lens performance but system performance, which is questionable.
Really it would make sense if we only compare APS-C to FT as these two systems are competing each other. If so, considering that FT sensor height is 13mm and APS-C is about 15mm depending on manufacturer, we would have only 15% size advantage of APS-C over FT (15/13), which is in part balanced by the average higher performance of FT glass vs. Canikon and the rest of the APS-C band. According to DPReview Olympus E-5 raw resolution is around 2800LWPH, in line with Nikon D7000.
' />
