Any telephoto zoom 4/3 Lenses for E-M1 that are much better than similar m4/3?

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

Sergey_Green wrote:

Anders W wrote:

DonSC wrote:

I've never heard this before and based on my experience with TCs it seems unbelievable. Even a 2X TC wouldn't degrade the image by that much. Just thinking about it, the 50-200mm is only ten or fifteen percent better in the center than the m.43 75-300mm, and when you add the 1.4 TC to the 50-200mm I don't see the image being any worse than what you get from the 75-300mm.

It's the laws of physics. The only thing the TC can do is to magnify the original image circle. When you magnify the image circle (and do it perfectly, all TCs are not perfect regrettably), the resolution of the lens as measured in lp/mm falls in direct proportion to the magnification (and stays constant for the image circle as a whole, but we are using only part of it now). What helps a bit in counteracting the problem is that the crop of the lens image that you are effectively dealing with utilizes the full resolution of the sensor and the image resolution is of course a product of lens as well as sensor resolution.

Many lenses out-resolve sensors today, so magnifying the projection (by less than two) hardly tells on the quality of the image. From my experience at least, with TC's and lenses I used.

Even if the lens "outresolves" (has higher resolution than) the sensor (which is a tricky thing to decide, especially since the lens doesn't have any resolution unless you specify a contrast criterion), a loss of lens resolution affects the image resolution negatively. A good approximation of how lens and sensor resolution actually combine into image resolution is

1/i = sqrt(1/l^2 + 1/s^2)

where i is image resolution, l is sensor resolution, and s is sensor resolution.

I have posted two pictures of dandelion above, one taken with 50-200 with ec14 and the other without, otherwise as close as I could make them. Anders has asked me in a previous post which is which. I couldn't tell myself if I didn't know, and though this kind of test is obviously limited, as a practical photographer I find it more useful than theory.

Before I comment on that, could you please answer the questions I asked, including the question about the testing procedure used. See here:

http://www.dpreview.com/forums/post/53284442

If you want a more detailed answer, I respectfully request a more detailed question. The larger, more zoomed-in file is taken with the teleconverter, but now that you know that, I'm afraid you can no longer judge the two images without bias.

There obviously are times when you need a longer or faster lens and you will choose whether or not to use a teleconverter on that basis. What this informal test tells me is that if I happen to have the teleconverter on this lens, I needn't bother removing it unless I need the extra stop.

I am sorry,

You don't act like it.

but this is hardly what I'd call a meaningful test.

Suit yourself

I can't see the file size, the originals are not available to me (they may be to you but I am not granted permission to see them),

My sympathies, DPR seems to be squirrely that way. I posted full size.

You need to grant permission for others to see what you post full size via your DPR preferences.

the framing is not identical, and the subject unsuitable for judging sharpness, especially sharpness across the frame. Were they shot handheld or on a tripod?

Handheld, IS1. note shutter speed 4x fl.

Why the slight difference in FL?

The zoom indications on that lens don't allow for exact division by 1.4 at the focal length most suitable for the subject.

What was the exact focus point? How did you assure exact the same focus on this problematic subject?

Single, small centerpoint af. Aperture f/8. Pretty much how I always shoot under similar circumstances.

Above all: Why not shoot what we really need in order evaluate the sharpness: A flat target perfectly parallel to the sensor

I have never, under the preconditions you describe, found an interesting or meaningful photographic subject. Knowing exactly how my gear performs at a task I never set it to is of no interest to me. YMMV.

Why do you think sites that test the sharpness of lenses shoot test charts in spite of the fact that no one find this an interesting or meaningful subject for any other purpose than that particular one?

I won't speculate on why they publish them. I only read them for a preliminary assessment of gear I don't own. Once I own them, I test them further, under the conditions I will use them. This is what I did, in response to a personal request from you; if my response is unsatisfactory to you please think twice before making another such request.

Sorry for sounding harsh, but for a while I suspected that you were playing games with me. I do appreciate the fact that you did try to do a test in response to a personal request from me. I am afraid I can't appreciate the results quite as much for reasons already mentioned but that's another matter.

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

gandalfII wrote:

Anders W wrote:

Sergey_Green wrote:

Anders W wrote:

DonSC wrote:

I've never heard this before and based on my experience with TCs it seems unbelievable. Even a 2X TC wouldn't degrade the image by that much. Just thinking about it, the 50-200mm is only ten or fifteen percent better in the center than the m.43 75-300mm, and when you add the 1.4 TC to the 50-200mm I don't see the image being any worse than what you get from the 75-300mm.

It's the laws of physics. The only thing the TC can do is to magnify the original image circle. When you magnify the image circle (and do it perfectly, all TCs are not perfect regrettably), the resolution of the lens as measured in lp/mm falls in direct proportion to the magnification (and stays constant for the image circle as a whole, but we are using only part of it now). What helps a bit in counteracting the problem is that the crop of the lens image that you are effectively dealing with utilizes the full resolution of the sensor and the image resolution is of course a product of lens as well as sensor resolution.

Many lenses out-resolve sensors today, so magnifying the projection (by less than two) hardly tells on the quality of the image. From my experience at least, with TC's and lenses I used.

Even if the lens "outresolves" (has higher resolution than) the sensor (which is a tricky thing to decide, especially since the lens doesn't have any resolution unless you specify a contrast criterion), a loss of lens resolution affects the image resolution negatively. A good approximation of how lens and sensor resolution actually combine into image resolution is

1/i = sqrt(1/l^2 + 1/s^2)

where i is image resolution, l is sensor resolution, and s is sensor resolution.

I have posted two pictures of dandelion above, one taken with 50-200 with ec14 and the other without, otherwise as close as I could make them. Anders has asked me in a previous post which is which. I couldn't tell myself if I didn't know, and though this kind of test is obviously limited, as a practical photographer I find it more useful than theory.

Before I comment on that, could you please answer the questions I asked, including the question about the testing procedure used. See here:

http://www.dpreview.com/forums/post/53284442

If you want a more detailed answer, I respectfully request a more detailed question. The larger, more zoomed-in file is taken with the teleconverter, but now that you know that, I'm afraid you can no longer judge the two images without bias.

There obviously are times when you need a longer or faster lens and you will choose whether or not to use a teleconverter on that basis. What this informal test tells me is that if I happen to have the teleconverter on this lens, I needn't bother removing it unless I need the extra stop.

I am sorry,

You don't act like it.

but this is hardly what I'd call a meaningful test.

Suit yourself

I can't see the file size, the originals are not available to me (they may be to you but I am not granted permission to see them),

My sympathies, DPR seems to be squirrely that way. I posted full size.

You need to grant permission for others to see what you post full size via your DPR preferences.

the framing is not identical, and the subject unsuitable for judging sharpness, especially sharpness across the frame. Were they shot handheld or on a tripod?

Handheld, IS1. note shutter speed 4x fl.

Why the slight difference in FL?

The zoom indications on that lens don't allow for exact division by 1.4 at the focal length most suitable for the subject.

What was the exact focus point? How did you assure exact the same focus on this problematic subject?

Single, small centerpoint af. Aperture f/8. Pretty much how I always shoot under similar circumstances.

Above all: Why not shoot what we really need in order evaluate the sharpness: A flat target perfectly parallel to the sensor

I have never, under the preconditions you describe, found an interesting or meaningful photographic subject. Knowing exactly how my gear performs at a task I never set it to is of no interest to me. YMMV.

Why do you think sites that test the sharpness of lenses shoot test charts in spite of the fact that no one find this an interesting or meaningful subject for any other purpose than that particular one?

I won't speculate on why they publish them. I only read them for a preliminary assessment of gear I don't own. Once I own them, I test them further, under the conditions I will use them. This is what I did, in response to a personal request from you; if my response is unsatisfactory to you please think twice before making another such request.

Sorry for sounding harsh, but for a while I suspected that you were playing games with me. I do appreciate the fact that you did try to do a test in response to a personal request from me. I am afraid I can't appreciate the results quite as much for reasons already mentioned but that's another matter.

Pax

Agreed!

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

50-200 @ 200mm

50-200 + EC-14 @283mm

Panasonic 100-300 @ 300mm

Notice as a bonus that the 50-200 + EC-14 doesn't lose out to the 100-300 in magnification. In fact, at closest focus, the Pany seems to have serious focus breathing and the Oly is much better (i.e. larger magnification).

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

Compare, for example, with the two shots below, both with the 100-300 at 300 and both available at full resolution. These are both at f/7.1 (which is close to the optimum aperture for the Pany at 300) rather than f/6.3 but although f/7.1 is better, the performance difference between the two f-stops is not all that great. The first shot is in excellent light and optimally exposed. The second is in very low-contrast light and a bit underexposed so noisier.

50-200 @ 200mm

50-200 + EC-14 @283mm

Panasonic 100-300 @ 300mm

Notice as a bonus that the 50-200 + EC-14 doesn't lose out to the 100-300 in magnification. In fact, at closest focus, the Pany seems to have serious focus breathing and the Oly is much better (i.e. larger magnification).

Sergey_Green wrote:

Anders W wrote:

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

Compare, for example, with the two shots below, both with the 100-300 at 300 and both available at full resolution. These are both at f/7.1 (which is close to the optimum aperture for the Pany at 300) rather than f/6.3 but although f/7.1 is better, the performance difference between the two f-stops is not all that great. The first shot is in excellent light and optimally exposed. The second is in very low-contrast light and a bit underexposed so noisier.

I would think the most you can compare is shoot images from different distances for the same framing, and then see if TC really degrades the results. I would expect to see some very minor differences, nowhere big enough to if you simply up-sample (or crop) smaller magnification to the larger one. So yes, it is worth using teleconverters when you need a longer reach, but the true lens may often be better than a shorter lens of the same class with teleconverter in front of it.

I think these images show quite a lot!

I didn't buy a teleconverter so that I can can shoot from further away. I bought it because I wanted more magnification. A teleconverter will almost always degrade the resolution of the original lens. The only bizarre circumstance where it might not, would if if it somehow corrected flaws in the original lens. Think Hubble telescope optics! Very unlikely.

The teleconverter is worthwhile if it gives better image quality than cropping and, in this case, if it gives better results than my 300mm lens.

I fully hope and expect the Oly 300mm f/4 to be better than the 50-200 + EC-14 and I will buy it if it is. A harder decision will be whether to replace the 50-200 with the new Oly 40-150 F/2.8, possibly with teleconverter. The new lens will have to be very good to make me get it. One of my biggest concerns is will it do C-AF as well as the 50-200.

Anders W wrote:

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

...

I agree. I have always thought that part of the issue with the 100-300 and m43 CDAF in general is that the AF performance in terms of accuracy degrades with deteriorating conditions or other sources of low contrast. So in lower light, or at higher Iso, or with a low contrast target, I suspect the AF always misses a bit. With the tree bud, this may have been a tough target for the 100-300's CDAF.

I am always a bit cautious about judging image quality from images of gulls or animals in zoos because they tend to let you get closer than with many other wildlife subjects. And in the case of gulls, they can be stationary in the air but still look like they are flying! That can be very flattering to C-AF tests.

This is an image that I took recently with the E-M1 and 100-300. It is probably the sharpest bird image I have ever taken with the lens. The conditions were perfect. I was in a hide about 8 feet from the bird. The evening sun was low in the sky and coming in from the side accentuating all the fine detail contrast in the feathers. I think this allowed the AF to work at its best. Most of the time, I find this doesn't happen.

Robin. E-M1 with Pany 100-300 @ 300mm

I'm quite impressed with how much easier it is to get good images with the 50-200. I have not had it very long but this was one of my first attempts. Check out the bird on the left.

Goldfinches. E-M1 with Oly 50-200 SWD @ 200mm

I was about 12-15 feet away from the goldfinches.

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

...

I agree. I have always thought that part of the issue with the 100-300 and m43 CDAF in general is that the AF performance in terms of accuracy degrades with deteriorating conditions or other sources of low contrast. So in lower light, or at higher Iso, or with a low contrast target, I suspect the AF always misses a bit. With the tree bud, this may have been a tough target for the 100-300's CDAF.

Getting the focus just right can certainly be a problem with then 100-300, yes. And then you have shutter-shock issues (and/or just ordinary camera shake) on top of that. But are you saying that there are bigger problems in this regard than with the 50-200 and E-M1 PDAF or FT PDAF? If so, I'd find that surprising. After all, one major strength of CDAF more generally is AF precision.

I am always a bit cautious about judging image quality from images of gulls or animals in zoos because they tend to let you get closer than with many other wildlife subjects. And in the case of gulls, they can be stationary in the air but still look like they are flying! That can be very flattering to C-AF tests.

I didn't mean it as a focus test really. I showed the pictures as decent examples of what the lens is optically capable of, when you get everything right.

This is an image that I took recently with the E-M1 and 100-300. It is probably the sharpest bird image I have ever taken with the lens. The conditions were perfect. I was in a hide about 8 feet from the bird. The evening sun was low in the sky and coming in from the side accentuating all the fine detail contrast in the feathers. I think this allowed the AF to work at its best. Most of the time, I find this doesn't happen.

Very nice image and far more representative of what the lens can when everything works out right (which I agree it does only sometimes ).

Robin. E-M1 with Pany 100-300 @ 300mm

I'm quite impressed with how much easier it is to get good images with the 50-200. I have not had it very long but this was one of my first attempts. Check out the bird on the left.

I agree that's very nice too.

Goldfinches. E-M1 with Oly 50-200 SWD @ 200mm

I was about 12-15 feet away from the goldfinches.

Anders W wrote:

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

...

I agree. I have always thought that part of the issue with the 100-300 and m43 CDAF in general is that the AF performance in terms of accuracy degrades with deteriorating conditions or other sources of low contrast. So in lower light, or at higher Iso, or with a low contrast target, I suspect the AF always misses a bit. With the tree bud, this may have been a tough target for the 100-300's CDAF.

Getting the focus just right can certainly be a problem with then 100-300, yes. And then you have shutter-shock issues (and/or just ordinary camera shake) on top of that. But are you saying that there are bigger problems in this regard than with the 50-200 and E-M1 PDAF or FT PDAF? If so, I'd find that surprising. After all, one major strength of CDAF more generally is AF precision.

Yes, I am suggesting that focus will be better on a sharper lens. If the 100-300 can't resolve the fine hairs on that tree bud or the fine quills (?) on the bird's feathers, it won't be able to use them in its contrast detect algorithm. If the image has lots of random noise in it, that will obscure the detail needed for CDAF. So as light levels drop and/or Iso increases, I think focus accuracy degrades.

The 50-200 PDAF seems to be at least as good (accurate) for static subjects as the 100-300 in CDAF. Out of the five shots I took with this lens + EC-14, three were like the best with the 100-300 and two were better. It makes sense to take a few shots if you can because one may be much better. I suspect this variability may have something to do with IBIS. If I focus in S-AF and then view the subject at 14x magnification before taking the shot, there is definitely some variation in clarity/focus. It wavers in and out. It is an effect a little bit like looking at C-AF with my old G3 only not nearly as obviously focus jittering and very minor by comparison. If I get a chance, I'll repeat the test on a tripod with IBIS off to see if it goes away.

The PDAF with the 50-200 is very good in C-AF too even when the subject is not moving. There is no hunting once initial focus is acquired and that initial focus seems to be as good as if you put the lens in S-AF. I often use it this way with focus activated by a button other than the shutter button. The subject stays in good focus until you are ready for the shot and then it fires quickly. I think C-AF on moving subjects is better with the 50-200 than with the 100-300.

There are a couple of situations (at least) where focusing on the 50-200 is not as good as native m43 lenses. One is certain types of detail and another is very low light.

I have a sheet of paper printed all over with very fine hexagons (actually the inside of an envelope) which I have used in the past for shutter shock tests with lots of lenses. The 50-200 refuses to focus on this at all. Like never! Similarly, below a certain light level (e.g. at dusk) it just fails to lock at all. I know that CDAF on the 100-300 works in these conditions.

Actually, I wish that the E-M1 gave manual control over PDAF/CDAF because I think there are cases where m43 lenses would work better with PDAF (say birds in flight) and 43 lenses would work better with CDAF (like the low light and hexagon detail situations described above). I would be happy to use PDAF most of the time unless m43 lenses have inherent issues with it (maybe m43 focus motors have poor positioning repeatability for example).

I just took this one today. The Blue Tit was about 20 feet away and this is a heavy crop. Image is now 2685 x 1790. I always found it hard to get the feather detail on these.

I should have set F/6.3 but it's pretty sharp anyway. You can see the original in my gallery. I am so glad I bought this lens!

Blue Tit. E-M1 with 50-200 SWD + EC-14 @ 283mm

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

...

I agree. I have always thought that part of the issue with the 100-300 and m43 CDAF in general is that the AF performance in terms of accuracy degrades with deteriorating conditions or other sources of low contrast. So in lower light, or at higher Iso, or with a low contrast target, I suspect the AF always misses a bit. With the tree bud, this may have been a tough target for the 100-300's CDAF.

Getting the focus just right can certainly be a problem with then 100-300, yes. And then you have shutter-shock issues (and/or just ordinary camera shake) on top of that. But are you saying that there are bigger problems in this regard than with the 50-200 and E-M1 PDAF or FT PDAF? If so, I'd find that surprising. After all, one major strength of CDAF more generally is AF precision.

Yes, I am suggesting that focus will be better on a sharper lens.

At least in theory, you might be right about that. The question is how much it matters in practice given the variation in sharpness we are effectively dealing with here.

If the 100-300 can't resolve the fine hairs on that tree bud or the fine quills (?) on the bird's feathers, it won't be able to use them in its contrast detect algorithm. If the image has lots of random noise in it, that will obscure the detail needed for CDAF. So as light levels drop and/or Iso increases, I think focus accuracy degrades.

Yes, but that is true about PDAF too, not least the on-sensor PDAF of the E-M1 (only some of the pixels serving as PDAF sensors).

The 50-200 PDAF seems to be at least as good (accurate) for static subjects as the 100-300 in CDAF. Out of the five shots I took with this lens + EC-14, three were like the best with the 100-300 and two were better. It makes sense to take a few shots if you can because one may be much better. I suspect this variability may have something to do with IBIS. If I focus in S-AF and then view the subject at 14x magnification before taking the shot, there is definitely some variation in clarity/focus. It wavers in and out. It is an effect a little bit like looking at C-AF with my old G3 only not nearly as obviously focus jittering and very minor by comparison.

You mean the clarity varies when you are at sufficient distance from the subject that your own body movements toward/away from the subject won't matter? I've seen what I think might be the kind of "jittering" you are describing too but chalked it up to atmospheric conditions. At a distance with the 100-300, that's a factor to reckon with, I would think.

If I get a chance, I'll repeat the test on a tripod with IBIS off to see if it goes away.

Yes, please do.

The PDAF with the 50-200 is very good in C-AF too even when the subject is not moving. There is no hunting once initial focus is acquired and that initial focus seems to be as good as if you put the lens in S-AF. I often use it this way with focus activated by a button other than the shutter button. The subject stays in good focus until you are ready for the shot and then it fires quickly. I think C-AF on moving subjects is better with the 50-200 than with the 100-300.

In what way is C-AF on the 50-200 better? If I understand things correctly, both lenses will use PDAF in AF-C on the E-M1 (although CDAF may be involved for touching-up in either of both cases).

There are a couple of situations (at least) where focusing on the 50-200 is not as good as native m43 lenses. One is certain types of detail and another is very low light.

I have a sheet of paper printed all over with very fine hexagons (actually the inside of an envelope) which I have used in the past for shutter shock tests with lots of lenses. The 50-200 refuses to focus on this at all. Like never! Similarly, below a certain light level (e.g. at dusk) it just fails to lock at all. I know that CDAF on the 100-300 works in these conditions.

Worse in low light makes sense for reasons already mentioned above. The thing about the pattern is interesting. Could be something with the spacing of the pattern relative to the spacing of the PDAF pixels. Does the ability to focus on this pattern vary with the distance/magnification?

Actually, I wish that the E-M1 gave manual control over PDAF/CDAF because I think there are cases where m43 lenses would work better with PDAF (say birds in flight) and 43 lenses would work better with CDAF (like the low light and hexagon detail situations described above). I would be happy to use PDAF most of the time unless m43 lenses have inherent issues with it (maybe m43 focus motors have poor positioning repeatability for example).

I can see the point about more control. And I wouldn't think MFT lenses have issues with PDAF, especially not of the kind you mention. On the contrary, their motors/interface better control over positioning and thus also more repeatability.

Tony Rogers wrote:

I just took this one today. The Blue Tit was about 20 feet away and this is a heavy crop. Image is now 2685 x 1790. I always found it hard to get the feather detail on these.

I should have set F/6.3 but it's pretty sharp anyway. You can see the original in my gallery. I am so glad I bought this lens!

Yes. Very nice.

Blue Tit. E-M1 with 50-200 SWD + EC-14 @ 283mm

Anders W wrote:

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

...

I agree. I have always thought that part of the issue with the 100-300 and m43 CDAF in general is that the AF performance in terms of accuracy degrades with deteriorating conditions or other sources of low contrast. So in lower light, or at higher Iso, or with a low contrast target, I suspect the AF always misses a bit. With the tree bud, this may have been a tough target for the 100-300's CDAF.

Getting the focus just right can certainly be a problem with then 100-300, yes. And then you have shutter-shock issues (and/or just ordinary camera shake) on top of that. But are you saying that there are bigger problems in this regard than with the 50-200 and E-M1 PDAF or FT PDAF? If so, I'd find that surprising. After all, one major strength of CDAF more generally is AF precision.

Yes, I am suggesting that focus will be better on a sharper lens.

At least in theory, you might be right about that. The question is how much it matters in practice given the variation in sharpness we are effectively dealing with here.

If the 100-300 can't resolve the fine hairs on that tree bud or the fine quills (?) on the bird's feathers, it won't be able to use them in its contrast detect algorithm. If the image has lots of random noise in it, that will obscure the detail needed for CDAF. So as light levels drop and/or Iso increases, I think focus accuracy degrades.

Yes, but that is true about PDAF too, not least the on-sensor PDAF of the E-M1 (only some of the pixels serving as PDAF sensors).

The 50-200 PDAF seems to be at least as good (accurate) for static subjects as the 100-300 in CDAF. Out of the five shots I took with this lens + EC-14, three were like the best with the 100-300 and two were better. It makes sense to take a few shots if you can because one may be much better. I suspect this variability may have something to do with IBIS. If I focus in S-AF and then view the subject at 14x magnification before taking the shot, there is definitely some variation in clarity/focus. It wavers in and out. It is an effect a little bit like looking at C-AF with my old G3 only not nearly as obviously focus jittering and very minor by comparison.

You mean the clarity varies when you are at sufficient distance from the subject that your own body movements toward/away from the subject won't matter? I've seen what I think might be the kind of "jittering" you are describing too but chalked it up to atmospheric conditions. At a distance with the 100-300, that's a factor to reckon with, I would think.

Yes, I think so. The subject is at about 20 feet and I've got my elbows resting on a table. I think the depth of field should be about 3 or 4 inches at this range (that's a whole bird! Focusing should be easy!). Too close for atmospheric effects I think.

If I get a chance, I'll repeat the test on a tripod with IBIS off to see if it goes away.

Yes, please do.

The PDAF with the 50-200 is very good in C-AF too even when the subject is not moving. There is no hunting once initial focus is acquired and that initial focus seems to be as good as if you put the lens in S-AF. I often use it this way with focus activated by a button other than the shutter button. The subject stays in good focus until you are ready for the shot and then it fires quickly. I think C-AF on moving subjects is better with the 50-200 than with the 100-300.

In what way is C-AF on the 50-200 better? If I understand things correctly, both lenses will use PDAF in AF-C on the E-M1 (although CDAF may be involved for touching-up in either of both cases).

In practice, trying to follow a bird in flight, the 50-200 seems to find focus better and keep locked on better. And the images are better. With the 100-300, it can be hard to track at all if the initial focus movements make the EVF view so out of focus you can't see the subject. My success rate is low.

It is possible that this could be because the 100-300 is rather slow to focus compared to some newer m43 lenses. Maybe it can't keep up. The 45-175 doesn't seem to be much better though. However, the 50-200 isn't fast. It is also possible that my technique has improved since I got the 50-200 and moved the AF to a separate button. I would have to try the 100-300 again to check.

There are a couple of situations (at least) where focusing on the 50-200 is not as good as native m43 lenses. One is certain types of detail and another is very low light.

I have a sheet of paper printed all over with very fine hexagons (actually the inside of an envelope) which I have used in the past for shutter shock tests with lots of lenses. The 50-200 refuses to focus on this at all. Like never! Similarly, below a certain light level (e.g. at dusk) it just fails to lock at all. I know that CDAF on the 100-300 works in these conditions.

Worse in low light makes sense for reasons already mentioned above. The thing about the pattern is interesting. Could be something with the spacing of the pattern relative to the spacing of the PDAF pixels. Does the ability to focus on this pattern vary with the distance/magnification?

The hexagon honeycomb pattern is made up of hexagons about 2.5mm across flats. The printed lines are dark and quite fine. I would guess 0.1mm wide. A quick check shows that it will focus at 283mm near to closest focus distance. However, it won't touch it at greater range or shorter focal length so it would appear that it is the size of the detail that matters or confuses it.

As the AF moves back and forth trying to get focus, I can momentarily see the pattern as plain as day through the EVF. Weird!

EDIT: I've just tried focusing on the edge of the pattern, where the border of the envelope is plain white. If I have half of the smallest focus box on the pattern and half on plain white it works! Works at 283mm and 70mm. That might imply that it is the repeating pattern that confuses it.

Actually, I wish that the E-M1 gave manual control over PDAF/CDAF because I think there are cases where m43 lenses would work better with PDAF (say birds in flight) and 43 lenses would work better with CDAF (like the low light and hexagon detail situations described above). I would be happy to use PDAF most of the time unless m43 lenses have inherent issues with it (maybe m43 focus motors have poor positioning repeatability for example).

I can see the point about more control. And I wouldn't think MFT lenses have issues with PDAF, especially not of the kind you mention. On the contrary, their motors/interface better control over positioning and thus also more repeatability.

Easy firmware update!

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

Anders W wrote:

Tony Rogers wrote:

For me, the practical question is: Is it better to use my Oly 50-200mm SWD without a teleconverter and crop harder or to use a teleconverter? And I bought this lens as a replacement for my Panasonic 100-300. Which is best?

Here's the answer.

I shot a tree bud out of the (open window). I took five shots with each setup and picked the best from each set.

The 200mm shot was Iso 200, 1/500 and F/4.5 for best sharpness. Image was processed in ACR using my defaults, upsized by 1.4x in Photoshop for convenience of comparison, de-noised and sharpened in my normal way (Neat Image + Smart Sharpen). Crop saved.

The 283mm (50-200 with EC-14) was Iso 400, 1/500 and F/6.3 for best sharpness, processed as above except that it is not up-sized.

The 300mm shot was the same settings and processing as the 283mm above.

These are setting choices I would have to make in practice. With the teleconverter or 100-300, I have to trade aperture for Iso.

And the results...

Thanks for doing the test. That the 50-200 with TC does better than a 50-200 shot upsampled to the same magnification is of course in line with expectations. The most interesting part is the comparison between the 50-200 with TC versus the 100-300. What surprises me here are two things. First that the shot with 50-200 plus TC looks noisier although the exposure is the same. Second, and more importantly, that the 100-300 doesn't do better in absolute terms.

...

I agree. I have always thought that part of the issue with the 100-300 and m43 CDAF in general is that the AF performance in terms of accuracy degrades with deteriorating conditions or other sources of low contrast. So in lower light, or at higher Iso, or with a low contrast target, I suspect the AF always misses a bit. With the tree bud, this may have been a tough target for the 100-300's CDAF.

Getting the focus just right can certainly be a problem with then 100-300, yes. And then you have shutter-shock issues (and/or just ordinary camera shake) on top of that. But are you saying that there are bigger problems in this regard than with the 50-200 and E-M1 PDAF or FT PDAF? If so, I'd find that surprising. After all, one major strength of CDAF more generally is AF precision.

Yes, I am suggesting that focus will be better on a sharper lens.

At least in theory, you might be right about that. The question is how much it matters in practice given the variation in sharpness we are effectively dealing with here.

If the 100-300 can't resolve the fine hairs on that tree bud or the fine quills (?) on the bird's feathers, it won't be able to use them in its contrast detect algorithm. If the image has lots of random noise in it, that will obscure the detail needed for CDAF. So as light levels drop and/or Iso increases, I think focus accuracy degrades.

Yes, but that is true about PDAF too, not least the on-sensor PDAF of the E-M1 (only some of the pixels serving as PDAF sensors).

The 50-200 PDAF seems to be at least as good (accurate) for static subjects as the 100-300 in CDAF. Out of the five shots I took with this lens + EC-14, three were like the best with the 100-300 and two were better. It makes sense to take a few shots if you can because one may be much better. I suspect this variability may have something to do with IBIS. If I focus in S-AF and then view the subject at 14x magnification before taking the shot, there is definitely some variation in clarity/focus. It wavers in and out. It is an effect a little bit like looking at C-AF with my old G3 only not nearly as obviously focus jittering and very minor by comparison.

You mean the clarity varies when you are at sufficient distance from the subject that your own body movements toward/away from the subject won't matter? I've seen what I think might be the kind of "jittering" you are describing too but chalked it up to atmospheric conditions. At a distance with the 100-300, that's a factor to reckon with, I would think.

Yes, I think so. The subject is at about 20 feet and I've got my elbows resting on a table. I think the depth of field should be about 3 or 4 inches at this range (that's a whole bird! Focusing should be easy!). Too close for atmospheric effects I think.

OK. I'll try to check this out for myself when I find the time.

If I get a chance, I'll repeat the test on a tripod with IBIS off to see if it goes away.

Yes, please do.

The PDAF with the 50-200 is very good in C-AF too even when the subject is not moving. There is no hunting once initial focus is acquired and that initial focus seems to be as good as if you put the lens in S-AF. I often use it this way with focus activated by a button other than the shutter button. The subject stays in good focus until you are ready for the shot and then it fires quickly. I think C-AF on moving subjects is better with the 50-200 than with the 100-300.

In what way is C-AF on the 50-200 better? If I understand things correctly, both lenses will use PDAF in AF-C on the E-M1 (although CDAF may be involved for touching-up in either of both cases).

In practice, trying to follow a bird in flight, the 50-200 seems to find focus better and keep locked on better. And the images are better. With the 100-300, it can be hard to track at all if the initial focus movements make the EVF view so out of focus you can't see the subject. My success rate is low.

But is this really with AF-C on the E-M1? In that case, PDAF should be used even with the 100-300 so the lens shouldn't go more out of focus than it initially is.

It is possible that this could be because the 100-300 is rather slow to focus compared to some newer m43 lenses. Maybe it can't keep up. The 45-175 doesn't seem to be much better though. However, the 50-200 isn't fast. It is also possible that my technique has improved since I got the 50-200 and moved the AF to a separate button. I would have to try the 100-300 again to check.

It could of course be that the 100-300 has more difficulties keeping up because it is slower, especially since the PDAF pixels cover only 1/16 of the sensor. Is this comparison with the 50-200 plus TC or the 50-200 alone? With the TC, there is not much of difference in max aperture compared to the 100-300 but without, it's slightly more than a stop.

There are a couple of situations (at least) where focusing on the 50-200 is not as good as native m43 lenses. One is certain types of detail and another is very low light.

I have a sheet of paper printed all over with very fine hexagons (actually the inside of an envelope) which I have used in the past for shutter shock tests with lots of lenses. The 50-200 refuses to focus on this at all. Like never! Similarly, below a certain light level (e.g. at dusk) it just fails to lock at all. I know that CDAF on the 100-300 works in these conditions.

Worse in low light makes sense for reasons already mentioned above. The thing about the pattern is interesting. Could be something with the spacing of the pattern relative to the spacing of the PDAF pixels. Does the ability to focus on this pattern vary with the distance/magnification?

The hexagon honeycomb pattern is made up of hexagons about 2.5mm across flats. The printed lines are dark and quite fine. I would guess 0.1mm wide. A quick check shows that it will focus at 283mm near to closest focus distance. However, it won't touch it at greater range or shorter focal length so it would appear that it is the size of the detail that matters or confuses it.

As the AF moves back and forth trying to get focus, I can momentarily see the pattern as plain as day through the EVF. Weird!

EDIT: I've just tried focusing on the edge of the pattern, where the border of the envelope is plain white. If I have half of the smallest focus box on the pattern and half on plain white it works! Works at 283mm and 70mm. That might imply that it is the repeating pattern that confuses it.

Yes, or the fact that the lines are very thin and rather sparse. That might mean that the PDAF pixels see only white, not the lines, in nearly all cases unless you go sufficiently close.

Actually, I wish that the E-M1 gave manual control over PDAF/CDAF because I think there are cases where m43 lenses would work better with PDAF (say birds in flight) and 43 lenses would work better with CDAF (like the low light and hexagon detail situations described above). I would be happy to use PDAF most of the time unless m43 lenses have inherent issues with it (maybe m43 focus motors have poor positioning repeatability for example).

I can see the point about more control. And I wouldn't think MFT lenses have issues with PDAF, especially not of the kind you mention. On the contrary, their motors/interface better control over positioning and thus also more repeatability.

Easy firmware update!

Absolutely.