fredmiranda.com

Moderated by: Fred Miranda

New fredmiranda.com Mobile Site

New Feature: SMS Notification alert

New Feature: Buy & Sell Watchlist

FM Forums | Leica & Alternative Gear | Join Upload & Sell

1 2 3 … 6

8 … 11 12 end

Archive 2011 · Switched entirely to Micro4/3

kwalsh
Offline
• • • •
Upload & Sell: Off

p.7 #1 · Switched entirely to Micro4/3

denoir wrote:
If that is the case, then I really do need a reality check! That's pretty amazing if correct. In that case they've come a long way since the GF1. The D3 and similar pro DSLR systems have significantly better AF systems than the 7D and other systems in the class. Still, definitely not orders of magnitude better.

I do have some reservations about the test as the lenses they used on the Canonikon systems are not exactly high end and it seems to me that in most cases there the system is limited by the lens AF rather than the camera...Show more →

You are on to something with the lens AF vs. camera AF. That is really how the m43 CDAF systems are getting such good performance, they are restricting their lens designs to have very small internal focusing groups so they can move them quickly enough to deal with the frequent direction changes needed in CDAF. They aren't being entirely "fair" by design. Of course good signal processing and a 120Hz sensor read out is a big part too, but in my mind it is really the lens design being optimized for CDAF that is key to them being able to take on many PDAF systems in the single shot case. Put a lens on there with a high mass focusing group and physics is quickly going to intervene!

Ken

Aug 08, 2011 at 08:28 AM

denoir
Offline
• • • •
Upload & Sell: Off

p.7 #2 · Switched entirely to Micro4/3

kwalsh wrote:
Read what I wrote carefully. Not just image circle, image circle and focal length so that each format has the same FoV. And yes, go ahead and try it with your software...

And yes, it is pretty basic - in fact so you don't have to go to the library here is a Google books result on scaling of optical designs:

http://books.google.com/books?id=qHJnPuc6dzYC&pg=PA33&lpg=PA33&dq=optical+design+scaling&source=bl&ots=yRRoUzYLKx&sig=T7IfLmNk7trL8CtcL4cu7afVIqE&hl=en&ei=Zds_TtP7L5HC0AH2vpShDg&sa=X&oi=book_result&ct=result&resnum=3&ved=0CC4Q6AEwAg#v=onepage&q=optical%20design%20scaling&f=false

That addresses the point I was trying to make, which may be a bit different than what you are referring to. I'm not trying to say a LF lens can't be made higher resolution than a m43 lens - I'm just...Show more →

It does not address the point you made. What you said was:
So the fundamental relationship with resolution while scaling focal length and image circle size (that is taking a given lens prescription for one image format and perfectly scaling it to another image format by multiplying all dimensions) is that the relationship for resolution scales perfectly linearly.

Which is I'm sorry to say as wrong as it can be. Resolution as in %MTF at a given spatial frequency as a function of distance to the image center does in no way depend linearly as you scale the lens design.

Also, you are using the term 'prescription' incorrectly. That's just one parameter of a single lens element defining the refractive power of it. What you mean is the 'optical design' of the lens, not' prescription'.

Did you bother looking at the prescriptions for the two Distagons?

Both textbook examples of a retrofocus design. Remarkably similar actually as usually MF retrofocus lenses are more simple. 'Distagon' by the way means simply that it is a retrofocus design. And what of it? I could have equally well used a Tessar to refute the claim that just because a lens covers a larger image circle it has to have inferior MTF curves.

I think I'm talking about fundamental optical scaling laws and you are better addressing what is available in the market place.

No, I'd say it's precisely the other way around. In theory there is no reason why a large format lens should have weaker MTF curves than a small format lens. In practice there are a many reasons, as I outlined in my post to Bob.

kwalsh wrote:
You are on to something with the lens AF vs. camera AF. That is really how the m43 CDAF systems are getting such good performance, they are restricting their lens designs to have very small internal focusing groups so they can move them quickly enough to deal with the frequent direction changes needed in CDAF. They aren't being entirely "fair" by design. Of course good signal processing and a 120Hz sensor read out is a big part too, but in my mind it is really the lens design being optimized for CDAF that is key to them being able to...Show more →

That's a good point. I had not thought about that the number of lens elements and their weight will unavoidably have an effect on focusing speed. But of course you are right, that definitely does make a difference.

Edit: Oops, I saw now that I forgot to post the first part of my response to Bob. Fixed now.

Aug 08, 2011 at 08:41 AM

Tomser
Offline
• •
Upload & Sell: Off

p.7 #3 · Switched entirely to Micro4/3

But does lens resolution vs. format matter at all ?
Each sensor size comes with its own set of lenses, doesn't it ?
And the bigger the sensor, the better the file, I think, just like it used to be with film .

Aug 08, 2011 at 04:07 PM

Jman13
Offline
• • • • • •
Upload & Sell: On

p.7 #4 · Switched entirely to Micro4/3

Ok...some sort of direct comparisons with Full Frame vs Micro 4/3. These are taken at the same spots, though two years apart in both cases, though with somewhat different light in one and vastly different foliage and water flow in another, but still...the crop comparison is valid at least.

So, first image: Hocking Hills, OH. The full frame pic was taken in October of 2009 (with fantastic colors due to the fall foliage and heavier water flow to lend some blues and such, plus a longer shutter speed since it was taken with a smaller aperture.).

The m4/3 pic was taken yesterday. Almost identical framing...I was maybe a foot to the right in the m4/3 pic. Used f/8, since that gives the same DOF as f/16 on full frame.

17mm with the 17-40L on the 1Ds Mark II on the full frame pic; Panasonic 7-14mm at 8.5mm on the GH2 in the m4/3 pic. The FF pic has much better colors, though this is primarily due to the better colors available at the time of shooting, and it's obviously the better shot as a result, but hey, you take what you can get. Processing was identical.

1Ds II with 17-40 f/4L @ 17mm, f/16, 25 seconds, ISO 100.
http://www.jordansteele.com/2011/ff-2.jpg

GH2 with 7-14mm f/4 @ 8.5mm, f/8, 10 seconds, ISO 160
http://www.jordansteele.com/2011/m43-2.jpg

Now on first glance, the FF shot is a much better shot, and I agree wholeheartedly, but not due to any technical superiority. Let's take a look at 100% crops, shall we?

Center frame...m4/3 is on the left, full frame on the right:
http://www.jordansteele.com/2011/m43-ff-3.jpg

To my eyes, VERY similar.

Now it gets really interesting...upper right corner at the bridge (again m4/3 left, FF right):
http://www.jordansteele.com/2011/m43-ff-4.jpg

The m4/3 shot is significantly sharper and contains a lot more detail. IMO the overall rendering here is also clearly better than the FF shot. Both were of course tripod mounted, mirror Lockup on the 1Ds II and both shot remotely.

On to test two. Very similar light, but not exact, one was in May, the other was today. The Full frame shot was with the 1Ds Mark II and the Sigma 15mm f/2.8 EX Fisheye at f/11. The M4/3 shot was with the GH2 and Panasonic 8mm f/3.5 fisheye at f/6.3.

1Ds II, 15mm, f/11, 1/320s, ISO 100:
http://www.jordansteele.com/2011/ff-1.jpg

GH2, 8mm, f/6.3, 1/1600, ISO 160:
http://www.jordansteele.com/2011/m43-1.jpg

So, to my eyes, the Sigma 15 has better flare control on the sun, and it gets nice rays due to the fact that the sun isn't partially obscured by a cloud at the time. A little richer in the blues perhaps. and perhaps a nicer overall tonal response from the FF shot due to better dynamic range. However, I don't think the global differences are huge.

To the crops, again, m4/3 on the left, FF on the right for both crop sets:
Center:
http://www.jordansteele.com/2011/m43-ff-1.jpg

Bottom Left:
http://www.jordansteele.com/2011/m43-ff-2.jpg

Look similar in tonal range and such to my eyes, with the m4/3 shot being vastly superior in the corners both in sharpness and in CA/color rendition. I will say that the m4/3 shot is more accurate in color rendition as well, especially in the reds, as the red in the m4/3 shot is dead on what the actual color of the bridge is.

Now, these are two examples with four lenses total, but it gives some insight into why those of us who shoot m4/3 aren't put off by the minor image quality things. Sure, a modern FF sensor will have much better noise control and better dynamic range. However, the m4/3 system is much more compact, has some amazing glass (as shown here, two wide options pretty much spank the 17-40 f/4L and the Sigma fisheye, which is widely considered as good or better than the Canon 15mm), and provides excellent overall image quality...in many cases equal to FF equivalents (and, at least in these tests, superior in terms of overall detail)

Perhaps the next time I go out, I'll lug along my FF kit (what remains of it) and shoot some more direct comparisons in the same light, which I'm sure would be more comparable for the overall shooting experience.

Aug 08, 2011 at 07:42 PM

sirimiri
Offline
• • • • •
Upload & Sell: Off

p.7 #5 · Switched entirely to Micro4/3

Jman, that's pretty impressive to get it lined up like that again, after time passes. But, mobile devices can help alot, if you had in with you. Or maybe you just had it framed in your head - I wish I could do that.

One thing- and thanks for the comparison - I wonder about, is if we're shooting at the wide end of the focal length spectrum, how well is a retrofocal wide-angle lens for an SLR body, going to compare to a system that doesn't have such a restriction?

The flip side of it might be, and again this is me just musing, but does the apparent differential even out, or even invert, once one gets to the full-frame equivalent of say 90mm or so?

What I'm getting at, is that in terms of the focal length for an SLR system, isn't this a little bit putting it at a disadvantage, for the sake of argument?

Aug 08, 2011 at 08:17 PM

mh2000
Offline
• • • • •
Upload & Sell: Off

p.7 #6 · Switched entirely to Micro4/3

Jman13, nice post! Thanks!

Aug 08, 2011 at 08:19 PM

denoir
Offline
• • • •
Upload & Sell: Off

p.7 #7 · Switched entirely to Micro4/3

Interesting comparisons Jordan, thanks for doing the test.

A couple of observations though:

1) Image quality in both cases, especially when looking at the 100% crops is not exactly top end. Yes, those are zoom lenses being used, but on FF135 you can choose among a large number of high quality ultra wide angle and wide angle lenses. For m4/3 your choices are far fewer.

2) Like all Canon UWA zooms, the 17-40 is a weak lens, especially in the corners. That the m4/3 is much better in the corners is no surprise. Designing a FF lens that is perfect in the corners is very difficult and Canon have failed miserably with its UWA zooms. I have the 17-40's big brother the 16-35/II and it was what got me to switch to using prime lenses.

Having said that, the m4/3 corners are better, but hardly perfect. Somebody said that the 8-14 was in the same league as the Nikon 14-24 and this is clearly not the case. Considering that the optical strain on the lens is twice of that of one on an FF camera it is nevertheless impressive.

3) Color separation seems to be much better on FF, but it could be to different lighting or RAW processing.

4) Center crops - in the first case the Canon is a lot better - the m4/3 crop just seems sharpened to death. In the second case the m4/3 is good but the FF shot is terrible. If you say that it's a good lens then I'd guess that the shot is somewhat out of focus.

5) The GH2 is a modern camera, while the 1DsII is almost eight years old which is a lot for a digital camera. There have been very significant sensor improvements in the mean time. I would actually be surprised if the GH2 did not have superior high ISO performance and better dynamic range compared to the 1DsII.

Aug 08, 2011 at 08:40 PM

kwalsh
Offline
• • • •
Upload & Sell: Off

p.7 #8 · Switched entirely to Micro4/3

denoir wrote:
It does not address the point you made. What you said was:

Which is I'm sorry to say as wrong as it can be.

Except of course it is exactly what my cited reference says, scale all the lens elements (resulting it identical increases in focal length and image circle size, which is exactly what you'd do for a change in format size while preserving the FoV) and the *angular resolution* stays the same. That is all I'm saying, and it is a fundamental scaling law in optics design and clearly described in the cited reference. I think you are reacting to something different that I'm not saying or writing. Perhaps I'm failing to write clearly.

Resolution as in %MTF at a given spatial frequency as a function of distance to the image center does in no way depend linearly as you scale the lens design.

That is not what I said, and of course I agree with you that it is wrong. Every MTF graph on earth illustrates it is wrong. Read what I said, carefully, and then actually read the link I posted if you care to. If you don't care to, that's fine too, we are getting further from the point! You keep arguing about something I'm not saying. Getting very old. Apologies if I'm failing to make myself understandable.

Also, you are using the term 'prescription' incorrectly. That's just one parameter of a single lens element defining the refractive power of it. What you mean is the 'optical design' of the lens, not' prescription'.

My apologies if I wasn't using the term to your liking. It is the term the two optics designers I've worked with on seekers use to describe the arrangement and specifications of all elements in a compound lens and I picked up the habit from them. My copy of "Modern Optical Engineering" (4th Edition) begins to describe a Cooke triplet anastigmat on page 117 and following much text and figures the author writes on page 119:

"The prescription for this lens is as follows:"

Following which is a table with 7 entries listing the positions, radius, semi-diameter and glass of every element as well as the specifications of the stop.

From all this I'm going to conclude my use of the word "prescription" to describe more than just the specifications of a single lens element but rather the specifications and arrangement of all lens elements in a compound lens is born out in both the literature and is used in this context by many practicing optical engineers.

Both textbook examples of a retrofocus design. Remarkably similar actually as usually MF retrofocus lenses are more simple. 'Distagon' by the way means simply that it is a retrofocus design. And what of it? I could have equally well used a Tessar to refute the claim that just because a lens covers a larger image circle it has to have inferior MTF curves.

Ah, my retrofocus example is a ridiculous comparison even though it is a practical example of fixed FoV design on different formats of very similar optical design. Your different retrofocus designs, which element for element are far more different than my examples if you'd bother looking at the "optical designs" ("prescriptions" is apparently verboten), are somehow completely justifiable and I'm a fool to challenge them. Interesting approach to rhetoric. Have you considered entering American politics? You might do really well!

Seriously, I think we are talking about different things. I was giving an example of two lenses with similar optical designs that have the same FOV on two different formats that shows the smaller format having higher resolution as a practical market example of how resolution can be relatively format independent for similar classes of lenses.

You seem to be illustrating a different (but also interesting point) that at a given focal length you can expand the image circle without suffering a loss of MTF by altering the lens design. I think you are trying to draw invalid parallels between my example and your example when we aren't trying to illustrate the same point.

In theory there is no reason why a large format lens should have weaker MTF curves than a small format lens. In practice there are a many reasons, as I outlined in my post to Bob.

I agree completely. You can make a very sharp large image circle lens and there are many examples of them - many are expensive or heavy or bulky but no theory prevents their construction if you give the designer unlimited resources. I was only pointing out an optical scaling law which shows that bigger does not automatically equal better all things being equal. Somehow this has offended you, I think because you are misinterpreting what I am saying or I'm failing to be clear. All I said was a scaled "optical design" will produce the identical resolution per format dimension, and that is an inarguable fact unless you are prepared to duke it out with Maxwell and his equations and overturn a century of optical theory and practice. I also said as a practical matter that isn't the way lenses are actually designed for different formats, in reality no one just scales a design from one format to another - and so as a practical matter the market does not always illustrate the fundamental scaling law simply because most lens designs aren't necessarily approached by scaling.

That's a good point. I had not thought about that the number of lens elements and their weight will unavoidably have an effect on focusing speed. But of course you are right, that definitely does make a difference.

It is really obvious if you put an older 4/3 lens meant for PDAF onto an m43 camera, things get a lot slower!

Ken

Aug 08, 2011 at 09:08 PM

denoir
Offline
• • • •
Upload & Sell: Off

p.7 #9 · Switched entirely to Micro4/3

Ken what you are trying to do is to back away from incorrect statements that you made - and by now realize are wrong - by adding other incorrect statements. But I don't see any point in a continuation of the discussion as when I point out some error you'll read up on it, misunderstand it and come with another incorrect statement just to save face. It's pointless and it's not working as you are showing over and over again that you are not very familiar with lens design beyond what you've managed to google during this discussion.

You have, to take a simple example, not realized that comparing simulated MTF charts from two different manufacturers is utter folly. Anybody vaguely familiar with the topic would know that there are only two major manufacturers that publish actually measured MTF curves - Zeiss and Leica (and here are even some question marks regarding Leica reliability). Pansonic and Nikon give out quasi-MTF charts that are based on computer simulations of the lens designs and without providing any data about their methodology. Nobody familiar with the field would make such an elementary mistake as to try to compare a Panasonic and a Nikon MTF chart.

I'm by the way no expert in the field - there are those in this forum that know this stuff much better than I do. My philosophy however is that it's pointless to try to cover up a lack of knowledge by verbal sparring and that I that I should learn from my mistakes rather than trying to defend them ad absurdum. My incorrect statement about contrast detection AF speed and you correcting me is a good example of my approach.

If you (or anyone else) are really interested in optical design and you don't have the time to plow through a whole book, I can highly recommend the following document: Fundamental optics. It covers all the basics without becoming too technical and detailed.

Anyway, that's it for me as far as this discussion goes. I'll assume that you'll disagree with everything I've said above, so I can save you the time of posting a refutation and denouncement of my post.

Aug 08, 2011 at 10:28 PM

kwalsh
Offline
• • • •
Upload & Sell: Off

p.7 #10 · Switched entirely to Micro4/3

I certainly agree about the unreliability of calculated MTF charts, it is a well known problem and you're right to call it out. On the other hand, the resolving power of those two lenses has been measured on both photozone.de and slrgear.com and those measurements match the MTF comparisons. To save you some breath, their testing methods are pretty unreliable, even more so when comparing systems so perhaps that ancillary data isn't any better than the calculated MTF charts. If that was what you were objecting to I concede that, the Zeiss data is far more reliable. I failed to catch that was what you were fussed about. If you are saying the 7-14 vs 14-24 MTF charts are so unreliable as to be irrelevant I can't honestly object to that, I've seen past articles (one I think from Ziess actually) that showed how large the errors can be. Mea culpa.

I was focusing on the fact you seemed to be claiming basic optical scaling theory was wrong - and that is a fundamental of lens design which it is difficult to imagine someone objecting to. Maybe I was misunderstanding what you were objecting to, or maybe you weren't understanding what I meant by scaling.

I'd recommend Modern Optical Theory if you are interested in photographic optics at a more detailed level. It isn't expensive and is more relevant to photographic optics than the Fundamental Optics link you provided (thanks by the way, it is a nice concise reference but doesn't discuss resolution or MTF - Modern Optical Theory is more quantitative and has example designs). Applied Photographic Optics is excellent, but I've only browsed for a few hours in a university library - it is rather pricey and I've resisted purchasing as it isn't my area of expertise either.

Really though, from my end, I've tried to stop spending much time reading on optical theory - it really doesn't help my photography and I've got plenty of other technical literature to read that is relevant to my job. I should probably add stopping entering optics discussions to the list as well.

Thanks for the discussion, even if it was abrasive. Your examples culled from the Zeiss products were very illuminating. I really do fundamentally agree with most everything you've been saying about the benefits of larger formats, I got wrapped around the axle on the optical scaling point which I'm not sure you understand - and that is OK, sticking to practical measured data rather than theory is probably more relevant to the discussion anyway (and you've made your point well on that data alone).

I'll sign out of the discussion here too. Apologies that it was acrimonious and I didn't do much of anything to make it less so.

Ken

Aug 08, 2011 at 11:18 PM

CKrueger
Offline
• • • •
Upload & Sell: Off

p.7 #11 · Switched entirely to Micro4/3

I don't understand why anybody gives a damn about MTF. It's a synthetic benchmark and is irrelevant if your goal is photography.

First of all it doesn't measure my lens or your lens, it measures a lens design; an ideal lens. A slightly decentered element or a slight vibration or a bit of haze in the air and your MTF is out the window. And perhaps more importantly, the lens' performance is only one part of what makes a high-quality image, and it's getting less important every day. We also have to consider the resolution of sensor, the camera's imaging engine, the software RAW converter (if any), and the processing done by the photographer. All of these are at least as important as the difference between a coveted Leica lens and a garden-variety Canon lens.

What I have seen in this thread is a number of great photographs with very high image quality. And I've seen some stats saying that these images aren't as good as I think. What should I believe? My own eyes, or equations? Forgive me, but I find MTF charts absolutely useless. If you want to learn which system is better, take them both out shooting. The resulting photos are the ONLY thing that matters.

denoir wrote:
If that is the case, then I really do need a reality check! That's pretty amazing if correct. In that case they've come a long way since the GF1. The D3 and similar pro DSLR systems have significantly better AF systems than the 7D and other systems in the class. Still, definitely not orders of magnitude better.

My experience is anecdotal, but having used a 1D2N, 5D2, 7D, and EP3, the EP3's CDAF is as fast as any DSLR PDAF I've ever used in single-shot AF mode. I don't know which camera on the market is the fastest at the moment, but it's neck and neck.

Clearly CDAF has come of age, at least for single-shot AF. For continuous AF and burst firing single-shot AF, PDAF is still vastly superior.

Aug 09, 2011 at 12:29 AM

Jman13
Offline
• • • • • •
Upload & Sell: On

p.7 #12 · Switched entirely to Micro4/3

denoir wrote:
1) Image quality in both cases, especially when looking at the 100% crops is not exactly top end. Yes, those are zoom lenses being used, but on FF135 you can choose among a large number of high quality ultra wide angle and wide angle lenses. For m4/3 your choices are far fewer.

Well, yes, it's likely not going to hold up against $2,000 wide angle lenses, but I wouldn't expect a $900 zoom to do that.

2) Like all Canon UWA zooms, the 17-40 is a weak lens, especially in the corners. That the m4/3 is much better in the corners is no surprise. Designing a FF lens that is perfect in the corners is very difficult and Canon have failed miserably with its UWA zooms. I have the 17-40's big brother the 16-35/II and it was what got me to switch to using prime lenses.

If you think that the 17-40L is a 'weak' lens, then m4/3 is not for you. While not perfect (it does soften in the corners), it and the 16-35 II are used by thousands of professionals across the globe to make pretty amazing, saleable images. I see shots from the 17-40L in photography magazines every month. It is not a 'weak' lens. It's not a perfect lens, and it's not as good as a Zeiss 21 or something, but it's not some crappy consumer zoom. Also realize that the corner crops are about 15 feet into the DOF (the focus point is just below the middle of the shot, before the water takes the first drop).

Having said that, the m4/3 corners are better, but hardly perfect. Somebody said that the 8-14 was in the same league as the Nikon 14-24 and this is clearly not the case.

Again, they're a good bit into the DOF. I have not used the Nikon 14-24, but while the 7-14 is an oustanding lens (especially given its price and extremely small size...less than 1/3 the weight), I doubt it's quite its equal in the extreme corners. Still...it's 1/2 the price as well.
3) Color separation seems to be much better on FF, but it could be to different lighting or RAW processing.

4) Center crops - in the first case the Canon is a lot better - the m4/3 crop just seems sharpened to death. In the second case the m4/3 is good but the FF shot is terrible. If you say that it's a good lens then I'd guess that the shot is somewhat out of focus.

I guess I'm not seeing what you're seeing at all. In the first case, I applied the same sharpening to both images, and they are not oversharpened. (I don't see a single sharpening artifact in the m4/3 shot...there's a little bit of noise in the water, but that's there in the FF shot too) As to the second shot, I also am not seeing what you're seeing. The shot is not OOF. I did not apply any sharpening to the second shot, so perhaps that's what you're reacting to, but there is plenty of fine detail in the wood grain in both shots. In the corners, the 8mm fisheye is superior...the best fisheye I've ever used in the corners (I've owned 3 others).

5) The GH2 is a modern camera, while the 1DsII is almost eight years old which is a lot for a digital camera. There have been very significant sensor improvements in the mean time. I would actually be surprised if the GH2 did not have superior high ISO performance and better dynamic range compared to the 1DsII.

I guess that's my point. First off, the 1Ds II was current until 2007, even though it was released in 2004. And until 2007 it was the absolute best DSLR you could buy. That's only 4 years. And many people, both serious amateur and working pros, still use the 1Ds II for their shooting...it's image quality is still outstanding and the camera is considered by many to be one of the best cameras ever made. If a $900 micro 4/3 body can equal it (or come darn close) in image quality, then that's by far good enough for me. What the heck else do we need? We've got a tiny system that equals (and in some ways betters) cameras and lenses that have been used by the best professionals for years... And yet many people continue to dog micro 4/3 as having 'inferior' image quality.

Aug 09, 2011 at 05:25 AM

denoir
Offline
• • • •
Upload & Sell: Off

p.7 #13 · Switched entirely to Micro4/3

kwalsh wrote:
Thanks for the discussion, even if it was abrasive. Your examples culled from the Zeiss products were very illuminating. I really do fundamentally agree with most everything you've been saying about the benefits of larger formats, I got wrapped around the axle on the optical scaling point which I'm not sure you understand - and that is OK, sticking to practical measured data rather than theory is probably more relevant to the discussion anyway (and you've made your point well on that data alone).

I'll sign out of the discussion here too. Apologies that it was acrimonious and I didn't do...Show more →

Oh well, we'll just have to try do better next time. People with an interest in lens design are few enough that it's a pity to ruin a good discussion by letting disagreements become personal. Let's hope we can both manage to avoid that in the next discussion.

CKrueger wrote:
I don't understand why anybody gives a damn about MTF. It's a synthetic benchmark and is irrelevant if your goal is photography.

Your second sentence is pretty much an indirect explanation of you first sentence.

An MTF chart shows you certain very important aspects of the rendering style of a lens. If you care about the final look of the image knowing the MTF of a lens will help you predict what that look will be like. It's of course not the only aspect of the rendering style of a lens and certainly not the only thing that determines the final look. The aspects of a lens that an MTF chart shows do however have a significant impact on the final image.

I've by the way already given you an extensive reply to your previous anti-MTF statement here:
https://www.fredmiranda.com/forum/topic/1033249/3#9821610

If you think that's just theory, think again. For instance a blind experiment we did a while ago showed that people here in the forum could with 90% accuracy tell a shot taken with a Canon lens vs a shot taken with a Zeiss lens. All shots were taken stopped down so there were no obvious clues such as bokeh - the rendering differences can accurately be predicted by looking at the MTF charts. What's more the experiment was repeated over at the POTN forums where the majority of the members are beginners and they did not do it much worse - around 80% accuracy.

So the aspects of the rendering that an MTF chart shows does make a noticeable difference even when one is limited to smaller web sized shots and at least photographers have no trouble spotting those differences. Whether the differences matter to you or not is a different question.

Aug 09, 2011 at 05:40 AM

denoir
Offline
• • • •
Upload & Sell: Off

p.7 #14 · Switched entirely to Micro4/3

Jman13 wrote:
Well, yes, it's likely not going to hold up against $2,000 wide angle lenses, but I wouldn't expect a $900 zoom to do that.

Of course, but the point was that you don't have the choice of getting an ultra high quality lens for m4/3 as it doesn't exist. And it would be unusual if it did given the market segment the systems are aimed at and the overall design goals of the system (i.e keeping it compact and light while providing convenient functions such as AF).

If you think that the 17-40L is a 'weak' lens, then m4/3 is not for you. While not perfect (it does soften in the corners), it and the 16-35 II are used by thousands of professionals across the globe to make pretty amazing, saleable images. I see shots from the 17-40L in photography magazines every month. It is not a 'weak' lens. It's not a perfect lens, and it's not as good as a Zeiss 21 or something, but it's not some crappy consumer zoom.

No disagreement there. You have to however take into context that this is a lens and image quality obsessed forum and I'm probably even more obsessed about it than most. The standards of what makes a lens 'good' or 'bad' are relative.

Having said that, for many applications the 16-35 & 17-40 are widely considered inadequate. A lot of landscape photographers have traditionally chosen Nikon over Canon because of the lack of good options in the wide end. Canon has partially fixed that with the excellent TS-E lenses, but the wide angle zooms remain distinctly weaker than the Nikon counterparts.

I guess that's my point. First off, the 1Ds II was current until 2007, even though it was released in 2004. And until 2007 it was the absolute best DSLR you could buy. That's only 4 years. And many people, both serious amateur and working pros, still use the 1Ds II for their shooting...it's image quality is still outstanding and the camera is considered by many to be one of the best cameras ever made. If a $900 micro 4/3 body can equal it (or come darn close) in image quality, then that's by far good enough for me. What the...Show more →

Inferior is a relative term and any criticism of the system should be taken as such. It is in comparison with the same generation of camera systems. All the things we've discussed here and the benefits & disadvantages of m4/3 vs FF could also be applied to FF vs medium format. A $2,000 135FF 5DII will for the most part produce better results than a $20,000 digital medium format Hasselblad that is a couple of years old. A modern digital medium format camera on the other hand will make the 5DII look like a cheap P&S.

That's the way things are when it comes to technology, and usually the concept 'good enough' can't be applied - only 'better' exists. This is of course not limited to cameras. If you today look at the best mobile phone from 2004 it will look rather pathetic - even though it was considered to be a brilliant phone when it was released and nobody felt it had any shortcomings.

Aug 09, 2011 at 06:06 AM

jj birder
Offline
• •
Upload & Sell: Off

p.7 #15 · Switched entirely to Micro4/3

denoir wrote:
Hmm. I'll have to disagree with that. Look at the moss on the left side of the shot and look at the grooves in the rock. It certainly has texture, it's just that the details are so low contrast that it all sort of blends together.

While I would agree with the premise that the combination of that prime and larger sensor outperforms the zoom and smaller format, you are clearly wrong about the geology. One photo is of a water-smoothed channel and the other is a rough rock wall formed by landslip. Yes the Zeiss would almost certainly have extracted more detail from the smooth rock, but it is not textured like your rough rock example.

Aug 09, 2011 at 07:15 AM

telyt
Offline
• • • •
Upload & Sell: Off

p.7 #16 · Switched entirely to Micro4/3

This comparison was very valuable for me. It convinced me that the R8+DMR isn't too big and heavy.

Aug 09, 2011 at 08:00 AM

Jman13
Offline
• • • • • •
Upload & Sell: On

p.7 #17 · Switched entirely to Micro4/3

Would you care to elaborate? From the samples I've seen, the DMR clips highlights pretty badly and has and nothing particularly special in terms of detail, color or noise compared to micro 4/3 chips, let alone a modern full frame chip. Heck, if you put stock in DxO (which I think is of limited use, but YMMV), the GH2 and Leica M8 have essentially identical capabilities. Nearly identical noise profiles, nearly identical color depth (21.2 bits (GH2) to 21.1 bits (M8), exactly identical dynamic range (11.3 stops)...and from what I've read, the DMR is nearly identical to the M8 in image quality (if not slightly behind the M8).

Aug 09, 2011 at 08:53 AM

denoir
Offline
• • • •
Upload & Sell: Off

p.7 #18 · Switched entirely to Micro4/3

DxO is well... of debatable usefulness. They should get credit for trying to make proper test. The problem is that they summarize the data in a too simplistic way which often leads to rather ridiculous results. Their sensor tests are a bit better than their lens tests but also has a high frequency of very dubious results.

I've never worked with DMR files, but the M8 is indeed an interesting point of comparison. If you look at the files at 100% it outdoes everything other camera on the market. The M9 is very very close but the stronger IR filter makes the results ever so slightly softer. So if only looking at 100% crops (you often can't tell a 100% crop from a full & resized image), it's the best digital camera out there. The problem is that although the individual pixels are fabulous, there are very few of them - it's a 10 megapixel camera.

The GH2 is 16 megapixel so it may be equal or at least approaching the same resolution. The M8 RAWs are 14 bit while the GH2 raws are 12 bit. As for dynamic range, it depends on how you measure it. If we stick to latitude then the M8 significantly outperforms the 5DII in the shadows but is inferior in the highlights. In the noise department the GH2 & M8 are probably comparable although I'd expect the GH2 to have an edge there.

As for resolution, does the 16 megapixels of the GH2 provide more detail than the 10 megapixel of the M8? I don't know. What I do know is that I have to reduce my NEX-C3 16 megapixel files to about 6-8 megapixel to get the same per pixel quality as the M9. So if the GH2 has a similar quality to the NEX when viewed at 100% then the answer is no.

Aug 09, 2011 at 09:46 AM

telyt
Offline
• • • •
Upload & Sell: Off

p.7 #19 · Switched entirely to Micro4/3

Jman13 wrote:
Would you care to elaborate? From the samples I've seen, the DMR clips highlights pretty badly and has and nothing particularly special in terms of detail, color or noise compared to micro 4/3 chips, let alone a modern full frame chip. Heck, if you put stock in DxO (which I think is of limited use, but YMMV), the GH2 and Leica M8 have essentially identical capabilities. Nearly identical noise profiles, nearly identical color depth (21.2 bits (GH2) to 21.1 bits (M8), exactly identical dynamic range (11.3 stops)...and from what I've read, the DMR is nearly identical to the M8 in...Show more →

What samples have you seen, and what have you been reading?

What I see in the comparison in this thread is that the pictures from both the GH2 and 1DsII lack detail and color gradation and that the L designation of the Canon 17-40 is quite a stretch.

Aug 09, 2011 at 09:52 AM

Bifurcator
Offline
• • • • •
Upload & Sell: Off

p.7 #20 · Switched entirely to Micro4/3

Jman13 wrote:
FlyPenFly - I don't think you've had any experience with the GH1, GH2 or G3, though, have you? Their sensor performance is significantly better than the E-P1/P2 and GF/G1,G2 series.

I mean, I know that the new Sony/Nikon sensor are fantastic, but really, especially since I don't shoot in super dark places often, how much better do I really need?

This was today at ISO 5000 on my GH2, and it was pushed a third of a stop, so is effectively ISO 6400. Works for me. One of the very few times I find myself needing to go beyond ISO...Show more →

I remain somewhat torn on this issue. I mean, when an image is scaled for the web or printed at 8x10 or smaller even the noise from terrible ancient tiny point&shoot sensors set to high ISOs, don't show up much. Sometimes in very flat detailless areas it does but even then not always and most of that cleans right up: https://www.fredmiranda.com/forum/topic/895722/0&year=2010#8432910 So in that regard there isn't a camera being sold today (or in the past 5 years) in ANY form-factor, that has "too much noise".

However, when cropping at 85% and above, printing much larger than 8x10, or when pushing the exposure a stop or more then it sure shows up! And I can tell you that although they are the best M4/3 cameras, all of the Panasonic G cameras have too much noise! Even at 100 ISO! This includes their flagship GH2 (tho it's base is 160 ISO IIRC). Most APS-C recorded images are very noticeably superior in this regard! And of course all FF cameras will kick both their butts!

So it really kinda depends. I can't really say "it works for me" and just leave it at that. It only does when the shots are perfectly exposed and/or they're scaled to web sizes - or near there. Then again, WTH, some people like grain and even add it to their images so... I guess there's a personal choice factor in there somewhere too.

Edited on Aug 09, 2011 at 10:23 AM · View previous versions

Aug 09, 2011 at 10:00 AM

1 2 3 … 6 7 8 … 11 12 end
FM Forums \| Leica & Alternative Gear \| Join Upload & Sell 1 2 3 … 6 7 8 … 11 12 end