Comparing MTF between FF and M43

[NigelC]

In assessing whether I could get suitable lenses to consider ditching my residual FF kit and consolidating on M43, I looked at the MTF curve at wide open for the Leica/Panasonic 45/2.8 (I took these from DPR test results). I then compared to those, on the same website, for my existing Canon 100/2.8 IS and the Canon is considerably better. Is there some scaling impact from the coverage of the lens that I don't understand or is the Canon just a better lens?

[ErikKaffehr]

Hi,

The LW/PH figures that DPReview and Photozone present are characterizing system performance. Fullframe 135 twice the dimensions of 4/3 so 4/3 needs to perform twice as well in absolute terms (MTF at lp/mm) as 135.

I guess the Canon 100/2.8 IS is a very good lens, so it could be very difficult to build a 4/3 that is twice as good. That said, some 4/3 lenses seem to be remarkably good.

In practice, it may be a bit different. You need to stop down for DoF, focusing may not be exact, there is camera vibration and diffraction if you stop down to much. Tests show the maximum performance a system is capable of under optimal circumstances.


Best regards
Erik

In assessing whether I could get suitable lenses to consider ditching my residual FF kit and consolidating on M43, I looked at the MTF curve at wide open for the Leica/Panasonic 45/2.8 (I took these from DPR test results). I then compared to those, on the same website, for my existing Canon 100/2.8 IS and the Canon is considerably better. Is there some scaling impact from the coverage of the lens that I don't understand or is the Canon just a better lens?

[Vladimirovich]

In assessing whether I could get suitable lenses to consider ditching my residual FF kit and consolidating on M43, I looked at the MTF curve at wide open for the Leica/Panasonic 45/2.8 (I took these from DPR test results). I then compared to those, on the same website, for my existing Canon 100/2.8 IS and the Canon is considerably better. Is there some scaling impact from the coverage of the lens that I don't understand or is the Canon just a better lens?

do you need it for macro ? most tests do not test lenses @ macro distances...

[NigelC]

There may be something I'm failing to grasp here. Firstly the difference in resolution arising from the camera is down to number of pixels, not the frame size - I'm comparing GH2/GX7 16 MP to 5D2 21MP and also the " per pixel sharpness" although I have no idea what factors shape that. The frame difference just results in bigger pixels which are one determinant  of overall signal/noise, i.e. how noisy at higher gain settings. Right so far?

Therefore a M43 lens needs to resolve more lines per mm to compensate for deficit in pixels, not size of pixels - Yes?

So the fact that the Leica/Panny 45 peaks just over 1500 and falls to about 1200 at edges at 50% contrast whereas the Canon starts at over 2200 and declines to about 1700 at edges suggests it's a lot better?

However, in the real world, the Leica Panny may be good enough. In reality in 3 and a half years I've really only used the Canon as a portrait lens - I think what I plan for indoor still life more equates to product shots, in which case the Sigma DP3M is a better tool anyway. The 45mm would, I think, be a good field macro lens though in conjunction with the GX7's tilting EVF.

I think I may do this and go with the Olympus 12mm at the wide end. If I find I still crave an ultra wide occasionally , there's always the Olympus 9-18

Sorry for boring you folks.

[ErikKaffehr]

There may be something I'm failing to grasp here. Firstly the difference in resolution arising from the camera is down to number of pixels, not the frame size - I'm comparing GH2/GX7 16 MP to 5D2 21MP and also the " per pixel sharpness" although I have no idea what factors shape that.

The factor that limits resolution is the lens, not the sensor. If you make a 24 MP sensor that is half size another 24MP the lens needs to be twice as good. That is hard to achieve.

The frame difference just results in bigger pixels which are one determinant  of overall signal/noise, i.e. how noisy at higher gain settings. Right so far?

Well, it is not dependent on gain/ISO. A larger sensor has always less noise, other factors kept equal.

Therefore a M43 lens needs to resolve more lines per mm to compensate for deficit in pixels, not size of pixels - Yes?

Don't think amount of data but quality of data. Making pixels smaller means that lens needs to transfer finer details, it is called MTF (Modulation Transfer Function), it is the ability of a lens to transfer contrast for fine details. A 4/3 lens needs to transfer contrast at the double rate compared with a FF lens.

So the fact that the Leica/Panny 45 peaks just over 1500 and falls to about 1200 at edges at 50% contrast whereas the Canon starts at over 2200 and declines to about 1700 at edges suggests it's a lot better?

However, in the real world, the Leica Panny may be good enough. In reality in 3 and a half years I've really only used the Canon as a portrait lens - I think what I plan for indoor still life more equates to product shots, in which case the Sigma DP3M is a better tool anyway. The 45mm would, I think, be a good field macro lens though in conjunction with the GX7's tilting EVF.

Yes to both.

Check this link: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/76-my-medium-format-digital-journey?start=12

39 MP Medium format, 24 MP FF and 24 MP APS-C. FF and APS-C use the same lens.

Best regards
Erik

Sorry for boring you folks.
[/quote]

[Vladimirovich]

However, in the real world, the Leica Panny may be good enough. In reality in 3 and a half years I've really only used the Canon as a portrait lens - I think what I plan for indoor still life more equates to product shots, in which case the Sigma DP3M is a better tool anyway. The 45mm would, I think, be a good field macro lens though in conjunction with the GX7's tilting EVF.

I think I may do this and go with the Olympus 12mm at the wide end. If I find I still crave an ultra wide occasionally , there's always the Olympus 9-18

Sorry for boring you folks.

why do you need to buy expensive 45/2.8 though ? when you can do w/ cheaper O60/2.8 for macro (and it is weathersealed, so you might later get some m43 sealed body too for outdoors)

[thefl]

I wouldn't try to over analyze and just test the lens or system you're interested in in order to make up your mind.

All those nice numbers and charts often don't really mean much in reality.
E.g. the Zuiko 50/2.0 Macro also gets a lot worse numbers than mentioned Canon lens. But in fact even when using the Zuiko on the highest resolution sensor in (m)FT world at the moment, this lens easily outresolves said sensor in a relatively wide aperture range. So the lens couldn't be optically any better for the system, not even theoretically, still it fares worse in the numbers than the Canon lens. Which also is a very good lens of course.
You can't judge the quality of a lens without a sensor behind it and with good lenses the sensors' capability is where the numbers come from, not the lenses'.

Also having the 45/2.8 (with quite some other macro lenses for other systems) I can't find an optical fault in it. Sharp (though not as pixel-perfect as the Zuiko 50/2.0 in the extreme corners - which I would call one of the best lenses I ever had), contrasty and with a beautifully soft bokeh.
Most tele macros belong to a system's highest resolution lenses and the little Pana is no exception. In case you don't get a bad copy...

[Vladimirovich]

E.g. the Zuiko 50/2.0 Macro also gets a lot worse numbers than mentioned Canon lens. But in fact even when using the Zuiko on the highest resolution sensor in (m)FT world at the moment, this lens easily outresolves said sensor in a relatively wide aperture range.

do you happen to point to a test w/ it on 16mp sensor or it is just your opinion ? because I can't recall such test.

[Ken Bennett]

My wife has the Olympus 60mm f/2.8 macro lens for m4/3. It's a fine lens, and we bought it for $400 when Oly was offering discounts earlier this summer. I've never seen the Panasonic 45mm macro, so I can't compare, but the Oly does make some very nice photos.

I would not choose it for portraits, however, when the Oly 45/1.8 is available for less than $400. The extra DOF control is well worth it, especially on the smaller m4/3 sensor, and the extra 1.3 stops of light gathering helps for indoor available-light portraits. It's an excellent lens.

[thefl]

do you happen to point to a test w/ it on 16mp sensor or it is just your opinion ? because I can't recall such test.

Hi,

What do you mean?
The statement that the Zuiko got worse numbers than the Canon or that it outresolves a 16MP sensor?

[Vladimirovich]

Hi,

What do you mean?
The statement that the Zuiko got worse numbers than the Canon or that it outresolves a 16MP sensor?

that it outresolves a 16MP sensor?

[Vladimirovich]

I would not choose it for portraits, however, when the Oly 45/1.8 is available for less than $400. The extra DOF control is well worth it, especially on the smaller m4/3 sensor, and the extra 1.3 stops of light gathering helps for indoor available-light portraits. It's an excellent lens.

and Sigma 60/2.8 is barely above $200 then... but OP wants macro lens.

[thefl]

that it outresolves a 16MP sensor?

That's my own conclusion based on several years of experience with this lens on (m)FT sensors of changing resolution and after a whole lot of comparisons with other lenses of known excellent resolution characteristics.

[EricV]

There may be something I'm failing to grasp here. Firstly the difference in resolution arising from the camera is down to number of pixels, not the frame size .... The frame difference just results in bigger pixels ... Therefore a M43 lens needs to resolve more lines per mm to compensate for deficit in pixels, not size of pixels - Yes?

No.  Frame size impacts resolution directly.  Pixel count or pixel size only limit resolution when the pixel size becomes too large to support the lens resolution.  The 4/3 lens needs to resolve more lines/mm primarily to compensate for the smaller frame size (mm not pixels).

So the fact that the Leica/Panny 45 peaks just over 1500 and falls to about 1200 at edges at 50% contrast whereas the Canon starts at over 2200 and declines to about 1700 at edges suggests it's a lot better?

Those values are lines per picture size, not lines per mm, so the frame size is already factored in.  This enables you to compare numbers directly across different formats.

[fike]

This topic again.  Wow!

As a generalization, full-frame cameras produce higher quality images--whether it is noise or resolution that floats your boat.

I worship quality.  I shoot panoramics. I buy good lenses.  I print giant images.  I shoot MFT.  It is good enough for me.

I can't tell you if it is good enough for you, but I can tell you that the current full frame cameras are better than the current MFTs. Visually the difference can be difficult to detect, but in the corner-cases of performance (low light, giant enlargements), full frame cameras are demonstrably better.

I am not sure why this reality is debated. 

[ErikKaffehr]

Hi,

I shoot APS-C, FF and medium format digital. Many of my best pictures were shot on 12 MP APS-C, which I feel is good enough for A2-size prints.

Best regards
Erik

This topic again.  Wow!

As a generalization, full-frame cameras produce higher quality images--whether it is noise or resolution that floats your boat.

I worship quality.  I shoot panoramics. I buy good lenses.  I print giant images.  I shoot MFT.  It is good enough for me.

I can't tell you if it is good enough for you, but I can tell you that the current full frame cameras are better than the current MFTs. Visually the difference can be difficult to detect, but in the corner-cases of performance (low light, giant enlargements), full frame cameras are demonstrably better.

I am not sure why this reality is debated.