Diwa labs tested a few ZD's

[Goffen]

Unfortunately they used the E-3, so the test is probably badly flawed, but anyway, here it is:

http://www.diwa-labs.com/wip4/detail.epl?id=216287

Just figures though, (just) no recommendations

Goffen

--
http://www.flickr.com/goffen

 

[John King]

Hi Goffen

Thanks for the links. I have bunged them into my bookmarks

Regards, john from Melbourne, Australia.
-----

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

http://canopuscomputing.com.au/gallery2/main.php



Bird Control Officers on active service.

 

[Louis_Dobson]

How could that be?

It isn't jut figures, there is a summary on the page you linked to.
--
http://www.flickr.com/photos/acam

 

[SphericalAberration]

"The outstanding Zuiko Digital 150mm f/2 is the best lens ever tested at DIWA Labs, showing modest quality reduction with the EC-14 tele converter added. A top performing 300mm f/2 or 420mm f/2.8 (35mm eqv.) in such a light weight and compact design is truly remarkable."

 

[Jeff]

There's an interesting comparison. After looking over the ZD results, go to the 1Ds page and look at lenses on the 1Ds. The increased resolution of the 1Ds does result in higher lp/mm at intermediate f stops.

Would the ZD lenses give a better showing on a higher res body? Probably. Does AA filter make a difference? Maybe, but can't know for sure from these particular lens tests.

How could that be?

It isn't jut figures, there is a summary on the page you linked to.
--
http://www.flickr.com/photos/acam

--
Jeff

 

[Goffen]

The 150mm is tested without the TC on another body before. I guess that's what they are referring to.

Goffen

--
http://www.flickr.com/goffen

 

[Goffen]

I was thinking of no recommendation/evaluation for each lens.

(And my comment rgd testing on the e3 was meant to be irony)

So i guess we'll never know how much of the degradation on the 150+1.4x is from the TC and how much is from the e-3.

Goffen

--
http://www.flickr.com/goffen

 

[Jan3x5]

There's an interesting comparison. After looking over the ZD results,
go to the 1Ds page and look at lenses on the 1Ds. The increased
resolution of the 1Ds does result in higher lp/mm at intermediate f
stops.

Yes, but there are lot of comparisons possible - just look at the MTF results of Zuiko 50-200 on E-3 and Canon 70-200/4L (many times described as outstanding lens) on 40D ...

 

[Jonas B]

Thank you for the link.

When looking at the MTF graphs I got a bit puzzled by this:

"This table contains the values of the MTF of the green channel, expressed in cycle/pixel, and in lp/mm (24x36mm eq)"

Can some tech savvy wxplain what this really means? When they say, for example, 50 lp/mm is it then really 100lp/mm they have measured and then divided by two to get the 35mm FF equivalents??

And why is that interesting? Does it make for a comparison number in the same way we get with IMA-test (but then expressed as LP/PH)?

thank you,

--
Jonas

 

[Jeff]

I was thinking of no recommendation/evaluation for each lens.

(And my comment rgd testing on the e3 was meant to be irony)

So i guess we'll never know how much of the degradation on the
150+1.4x is from the TC and how much is from the e-3.

The results for the 150 are shown both with the TC (marked as 212mm) and without the TC (marked as 150mm, natch).

Goffen

--
http://www.flickr.com/goffen

--
Jeff

 

[Jeff]

It's kind of explained here: http://www.diwa-labs.com/wip4/test_result.epl?cat=11122

While not explicit, I read their explanation to mean essentially what you have described. This would allow a simple comparison of resolution among different brands and formats.

Thank you for the link.

When looking at the MTF graphs I got a bit puzzled by this:

"This table contains the values of the MTF of the green channel,
expressed in cycle/pixel, and in lp/mm (24x36mm eq)"

Can some tech savvy wxplain what this really means? When they say,
for example, 50 lp/mm is it then really 100lp/mm they have measured
and then divided by two to get the 35mm FF equivalents??

And why is that interesting? Does it make for a comparison number in
the same way we get with IMA-test (but then expressed as LP/PH)?

thank you,

--
Jonas

--
Jeff

 

[D Q E]

Unfortunately they used the E-3, so the test is probably badly
flawed, but anyway, here it is:

http://www.diwa-labs.com/wip4/detail.epl?id=216287

Just figures though, (just) no recommendations

Not just figures, but INCREDIBLY BADLY REPRESENTED figures.

Bad graphing.

--
--
mumbo jumbo

 

[Jeff]

There's an interesting comparison. After looking over the ZD results,
go to the 1Ds page and look at lenses on the 1Ds. The increased
resolution of the 1Ds does result in higher lp/mm at intermediate f
stops.

Yes, but there are lot of comparisons possible - just look at the MTF
results of Zuiko 50-200 on E-3 and Canon 70-200/4L (many times
described as outstanding lens) on 40D ...

Wow -- that's dramatic. Not so dramatic but quite interesting is to compare the E-3 w/ 50 and the D300 w/ 60mm. The E-3 (AA filter and all) holds up quite well.

--
Jeff

 

[Jonas B]

It's kind of explained here:
http://www.diwa-labs.com/wip4/test_result.epl?cat=11122

While not explicit, I read their explanation to mean essentially what
you have described. This would allow a simple comparison of
resolution among different brands and formats.

Hmm, maybe I should have looked around before posting. I'll click that link.

Thank you Jeff,

--
Jonas

 

[Jeff]

I've had the 150/2 since it first came out. I bought for shooting indoor high school hockey. These measurements are completely in line with my experience. It's a lens you can freely use wide open with no evident softness, even in the corners. It really is a remarkable lens.

The only challenge is to get critical focus with such a shallow DOF, but that comes with the territory.

"The outstanding Zuiko Digital 150mm f/2 is the best lens ever tested
at DIWA Labs, showing modest quality reduction with the EC-14 tele
converter added. A top performing 300mm f/2 or 420mm f/2.8 (35mm
eqv.) in such a light weight and compact design is truly remarkable."

--
Jeff

 

[Rriley]

Thank you for the link.

When looking at the MTF graphs I got a bit puzzled by this:

"This table contains the values of the MTF of the green channel,
expressed in cycle/pixel, and in lp/mm (24x36mm eq)"

Can some tech savvy wxplain what this really means? When they say,
for example, 50 lp/mm is it then really 100lp/mm they have measured
and then divided by two to get the 35mm FF equivalents??

yes
they use 35mm equivalence which as the story goes
requires 4/3rds to have twice the resolution to compete

im sure you would have heard that statement bandied about when speaking of 4/3rds in the bashing sense

the lens charts then, instead of providing lpmm in 10lpmm and 30lpmm that you might recall seeing in documents from say Canon,



Olympus then use 20lpmm and 60lpmm for MTF charts like that shown for the 35-100/2 shown above. 20lpmm are in black, 60lpmm are in orange

APSC of course, are not quite that lucky

And why is that interesting? Does it make for a comparison number in
the same way we get with IMA-test (but then expressed as LP/PH)?

thank you,

--
Jonas

--
Riley

I'd give my right arm to be ambidextrous (just)

 

[Jeff]

With a 10MP sensor, my guess on the upper limit on resolution in these charts would be

sqrt(10,000,000/1.333) (36mm*2) = 38 lp/mm (24x36mm equiv).

If that's anywhere close to right, then what most of the charts are showing is a resolution limited by the sensor+AA rather than the lens. It's worthwhile keeping that in mind while looking over these charts.

Thank you for the link.

When looking at the MTF graphs I got a bit puzzled by this:

"This table contains the values of the MTF of the green channel,
expressed in cycle/pixel, and in lp/mm (24x36mm eq)"

Can some tech savvy wxplain what this really means? When they say,
for example, 50 lp/mm is it then really 100lp/mm they have measured
and then divided by two to get the 35mm FF equivalents??

And why is that interesting? Does it make for a comparison number in
the same way we get with IMA-test (but then expressed as LP/PH)?

thank you,

--
Jonas

--
Jeff

 

[D Q E]

Thank you for the link.

When looking at the MTF graphs I got a bit puzzled by this:

"This table contains the values of the MTF of the green channel,
expressed in cycle/pixel, and in lp/mm (24x36mm eq)"

Can some tech savvy wxplain what this really means? When they say,
for example, 50 lp/mm is it then really 100lp/mm they have measured
and then divided by two to get the 35mm FF equivalents??

First up, they should be measuring white light (ie luma) rather than just using the green channel. Secondly, that's NOT an MTF graph, it's a graph showing "resolution at MTF50", which is quite a different thing. Thirdly "cycles per pixel" is a slightly baffling measure, but it essentially one of spatial frequency, and yes, a cycle would be the sinusoidal equivalent of a line PAIR.

They also seem to be normalising the results to some 20x30cm 'print' size, which seems questionable in value.

And why is that interesting? Does it make for a comparison number in
the same way we get with IMA-test (but then expressed as LP/PH)?

The proprietary results formats spat out by DxO and Imatest do not make for good understanding of performance in my view, and are both hugely compromised by being SYSTEM results rather than component results.

--
--
mumbo jumbo

 

[Jonas B]

It's no doubt that the lenses outresolve the sensor/AA-filter. But I don't think the limit is at 76 lp/mm:



Above are a few samples taken with an E-510 and a standard USAF chart as target.

The max resolution is achieved at f/5.6 (not shown above) and as the distance is known (55xFL) in all cases I get resolution numbers somewhere around 90 lp/mm. If our sensors are 13mm (height, measuring the active area) that would equal 48-50 lp/mm in 35mm FF numbers.

And that is lucky. 76 lp/mm would mean very por resolution compared to a 5D for example.

(My experience during the last years tells me that the rough numbers read from these USAF charts corresponds pretty well with the test results I see at photozone and similar places.)

regards,

--
Jonas

With a 10MP sensor, my guess on the upper limit on resolution in
these charts would be

sqrt(10,000,000/1.333) (36mm*2) = 38 lp/mm (24x36mm equiv).

If that's anywhere close to right, then what most of the charts are
showing is a resolution limited by the sensor+AA rather than the
lens. It's worthwhile keeping that in mind while looking over these
charts.

 

[Jeff]

Thanks for the interesting followup. I've been interested in the Panasonic/Leica 25/1.4, so was very intrigued by the apparent sharpness of that lens. Is what you have seen in this testing? How does is campare to the 50/2 overall (since its a lens I have and am very fond of, this is helpful comparison).

My little calculation was based on the number of pixels on the sensor. If you're seeing up to 90 lp/mm, wouldn't that mean you're seeing more line pairs than there are pixels?
--
Jeff