[Rondec]

I think there are a lot of different things you can look at. One is peak performance. A DA *55 has its peak sharpness at f4, while a lens like the DA *60-250 peaks at f5.6. The DA * 55 is decent by f2. Obviously if you need to shoot at 250mm and f4 there is no comparison at the same time if you need a lens that you can use at 55mm and f2 and is relatively small sized, you won't find that in a zoom.

To me, zooms and primes serve different purposes and there are few situations where you would substitute one for the other.

---

Last edited by Rondec; 08-07-2015 at 04:40 PM.

[clackers]

Which model?

Oh, the 2.8.

Here's the first IS version, lovingly reviewed by ... Ken Rockwell!

http://www.kenrockwell.com/canon/lenses/70-200mm-f28-is.htm

[fwcetus]

in my short time on this forum, I've only seen two zooms referred to as 'stack of primes': the SMC A 35-105mm f3.5 and the SMC A 35-70mm (but which version, I'm not sure)...

It must be the A 35-70/4.

[.a.t.]

Oh, the 2.8.

Here's the first IS version, lovingly reviewed by ... Ken Rockwell!

Canon 70-200mm f/2.8 IS

I appreciate the link. Thank you.

[normhead]

Sorry, I was out, I used the Pentax 100 macro for comparison because it was there. The only Pentax 135 prime tested was the K-135 2.5 and the DA* 60-250 mops the floor with it. I used the FDA 100 because it's a lens that a lot of people use.

But if you have to have a 135 here ya go. The FDA 100 macro does much better against the 60-250 by the way.



I'm not sure what mTF is but lw/ph, is Line Width per Picture Hieght. So if you have an lw/ph of 2350, you can produce a test pattern of 2350 distinct lines on your picture no matter what sensor size or the image height is.

My Tamron 90 macro is another top prime i use. It's very highly rated on every platform it's been tested on. The difference is indestinguishable at first glance but if you really stare for a long time pixel peeping side by side, the consensus is the 90 is a little bit better micro contrast, probably due to less CA.

[Nicolas06]

Sorry, I was out, I used the Pentax 100 macro for comparison because it was there. The only Pentax 135 prime tested was the K-135 2.5 and the DA* 60-250 mops the floor with it. I used the FDA 100 because it's a lens that a lot of people use.

But if you have to have a 135 here ya go. The FDA 100 macro does much better against the 60-250 by the way.



I'm not sure what mTF is but lw/ph, is Line Width per Picture Hieght. So if you have an lw/ph of 2350, you can produce a test pattern of 2350 distinct lines on your picture no matter what sensor size or the image height is.

My Tamron 90 macro is another top prime i use. It's very highly rated on every platform it's been tested on. The difference is indestinguishable at first glance but if you really stare for a long time pixel peeping side by side, the consensus is the 90 is a little bit better micro contrast, probably due to less CA.

MTF50 mean you accept to consider that the pattern is still visible if it keep only 50% from original contrast (you have lost 100-50= 50% of contrast).
MTF30 is a common measure and would mean you accept to keep only 30% of original constrast and so you accept a lost of 70% contrast.

Basically the original signal was say black and whites line and the lense resolve light gray vs dark gray lines. You'd need to boost contrast extensively to get back to original contrast. MTF50 tend to be the standard of many review sites and many manufacturers use MTF30 on their spec charts.

We have also to keep in mind this is for a given subject distance (lenses do have quite different sharpness characteristic at different subject distance) and that the number are very specific to a camera sensor. 2 lenses can appear almost identical on an old 10MP sensor but quite different on a 24MP sensor without low pass filter that can go up to 3500-3700 while a old 10MP might not get more than 2000-2100 and a K5 would go up to 2700. Typicall a weaker lense migtht manage to show quite good performance on an old sensor but might struggle with border performance with a sensor with much more resolution.

I know it is a bit sad we don't have better number, but when you show us theses photozone results norm you show us how many dark grey vs light grey lines you can count on a quite old camera sensor. This say nothing on how the lense perform without degrading the signal, how it would perform on a more modern sensor, how it perform for landscape where the subject can be far away or how it perform in contra light.

Counting that we typicall display our picture at quite small size and we can't really see that much detail to begin with, the tendancy to use the resolution metric with MTF50 as the key indicator of lense performance might not be that relevant in the end.

If we take the 60-250 case:
- photozone find the borders are a bit behind and noticably so at f/4 60mm on an K10 but overall that the performance is really, really great.
- e-photozine working on a K5-IIs find that while the lense is very sharp at 60mm f/4, it need f/5.6 by 135mm and f/8 by 250mm to keep this sharpness... And the borders at 135mm and 250mm never catch up the center
- looking at user reviews here, many think the performance is great at f/4 but is really better at f/5.6 so that seems to match.

Still it seems that the bokeh is truely great, as are the colors... And for sure the performance even from the e-photozine review show is certainly good enough for as you are likely to care mostly of the center and you will close down to f/5.6 in many situation when that possible.

Can we say it is stack of prime?
Would it compare favorably to DA*300 or DA*200 ? Difficult to say but maybe we are not that far!
Would it compare favorably to a DA50 f/1.8 (one of the cheapest prime available) or DA70 f/2.4 or worse FA77 f/1.8? I'd say no way it could do that. The wider apperture are really key for portraiture and even if maybe the DA50 is no sharper, what it can offer would look dramatically different than what the 60-250 can do. FA77 is going to have better bokeh too, even at same apperture and DA70 is likely to be sharper, corner to corner.

In the end saying if 60-250 is a stack of prime will mainly depend where you focus is... If it is for landscape closed down or wildlife, that may be very well true. For portraiture and low light that would be quite wrong.

[normhead]

I know it is a bit sad we don't have better number, but when you show us theses photozone results norm you show us how many dark grey vs light grey lines you can count on a quite old camera sensor. This say nothing on how the lense perform without degrading the signal, how it would perform on a more modern sensor, how it perform for landscape where the subject can be far away or how it perform in contra light.

Actually, it's not sad at all. You take what you have, you look at a pile of images and you see what they mean, if there is any correlation between MTF and lens performance.

To say they say mean nothing... sorry to be contradictory, but that's just wrong. The two biggest factors in lens accuracy, are measured by lw/ph and control of CA, because these are the numbers that tell you how good a lens is at stopping light coming from different parts of the scene affecting pixels that they should not be affecting. A lens that only gets 1000 lw/ph is scattering light all over the image degrading the image. I'ts either poorly designed or not manufactured to high enough tolerances. CA similarly spreads light from different spectrums onto neighbouring pixels, again compromising subject integrity. In my experience after many comparisons, a lens that isn't high up the lw/ph charts has no hope of achieving stellar micro-contrast that will lead to realistic looking images. Again based on my own experience. With two lenses within say 100 lw/ph in resolving power, which is functionally about the same given lens sample variation etc. the lens with the best control of CA will produce the best image.

Can we say it is stack of prime?
Would it compare favorably to DA*300 or DA*200 ? Difficult to say but maybe we are not that far!
Would it compare favorably to a DA50 f/1.8 (one of the cheapest prime available) or DA70 f/2.4 or worse FA77 f/1.8? I'd say no way it could do that.

This is where you and I differ. I look at my Tamron 90 images and my 60-250 images and say "this is how it is" and report my results. You look at a bunch of numbers and say things like the above. IN my mind, the only people worth listening to are people doing some real world comparisons. I know what to expect based on the numbers, been there done that, my area of interest is "how does this play out in the real world?" I don't have a lot of time for speculations. SO are the mtf charts a valuable tool? Combined with CA charts and distortion characteristics, yes they are. Is every difference they point out relevant? Hardly, in many of my test images it's hard to tell the difference. Have I ever had test results comparing lenses that went against the MTF and CA numbers? No, and I've tested close up, far away, medium subjects. That being said, I think most would be surprised at how good a lens a level down on the charts can do compared to your ace lens. As in the last photo test i did, where every lens that was functioning properly did an acceptable job, including my old FA 35-80 kit lens.

so to summarize. You learn things about a lens looking at MTF and CA numbers. There's precious little difference between a good lens and a great lens in every day shooting. The biggest difference in great lenses, well corrected and sharp is how they render out of focus areas. And that's the difference between great test chart lens and a great lens.The great lens will have smooth OoF areas.

For portraiture and low light that would be quite wrong.

There is no requirement that a prime lens has to be faster than ƒ4. You are attributing speed to primes, and that's an error. Zooms can be fast, primes can be slow. You need to rethink this. You need to test lenses for the purposes. With the Sigma 18-35 ƒ1.8, and more lenses of that type probably in the works, it's likely that speed, the last resort of primes will disappear as well. My 21 is ƒ3.4, my 35 is ƒ2.4, neither is as good as the Sigma 18-35 ƒ1.8 for portraits and low light as you define it. You're talking about individual lens characteristics, not characteristics of zooms or primes. So yes, the DA*60-250 is a stack of ƒ4 primes. That it's not a stack of ƒ2.8 primes is irrelevant.

---

Last edited by normhead; 08-08-2015 at 07:41 AM.

[Bagga_Txips]

Everyone knows Norm loves his 60-250! (j/k)

[Rondec]

The DA *60-250 is a very nice lens. Stellar for a zoom. I just don't know that I would refer to it as a stack of primes. As Nicolas says, the question really is what you are comparing the lens to. I just don't know of any time that I would be excited to shoot with an FA 77 (for instance) that I would substitute a 60-250. They are such different lenses.

However, just to get back to the initial question, my perception is that when folks say that a zoom is like a "stack of primes," they usually mean that a zoom has a relatively fast aperture and is usable close to wide open. Lenses like the 50-135, DA 20-40, and Sigma 18-35 f1.8 certainly seem to fit that bill.

That doesn't mean that there aren't primes that have better sharpness at f2.8 or bokeh than these lenses, but they are certainly pretty good.

[aleonx3]

The way I understand it, "a stack-of-prime" is usually associated with zoom lens in a short range (for example, the 20-40). Prime lens design are optimized in image rendering and less distortion (CA, barrel etc.), and to lesser degree widest aperture. I have the 20-40 and it is exactly what I expect, very little distortion comparing to the equivalent prime lens at the same focal lengths.

Rondec had a good point, a "stack-of-prime" lens also must have characteristic of fast aperture (less then f4). Hence, the old F35-70 lens (though short range) but usually not considered as in "stack-of-prime" category.

[normhead]

The DA *60-250 is a very nice lens. Stellar for a zoom. I just don't know that I would refer to it as a stack of primes. As Nicolas says, the question really is what you are comparing the lens to. I just don't know of any time that I would be excited to shoot with an FA 77 (for instance) that I would substitute a 60-250. They are such different lenses.

I would be comparing it to any ƒ4 prime in it's class, and many ƒ 2.8 lenses when they aren't being shot at ƒ2.8. I'm not sure why this is so hard to comprehend. It seems like everyone wants to say "my lens is a stack of primes and yours isn't." Calling a Sigma 18-35 ƒ 1.8 a stack of primes would be an insult to zooms everywhere.

So my definition would be,
1. Excellent or very near excellent everywhere in it's range.
2. Excellent or near excellent borders every where in it's range.
3. Prime like control of CA every where in it's range.

Except for 60-70mmm which is a very small portion of it's range the DA*60-250 will achieve that.

It seems to be the consensus among some that zooms can't compete. Well, a lot of them can. They may not be able to compete with every prime in their whole range... but they will also be able to surpass and actually be the best lens averrable in other focal lengths. If there are three focal lengths it can cover that are better than any prime available, I would argue it's a stack of primes. For Pentax mount, as far as I can tell, the 60-250 cover 90-100mm ƒ4 and up as good or better than whatever is available, it owns the 135-180 range. It's equivalent to the DA*200 ƒ2.8, ƒ4 and up and it owns 210-250. If it's as good as the 90-100 primes, better than anything 135-180 and equal to or better than any prime 200-250, how is it not a stack of primes?

And the Sigma 18-35 ƒ1.8, definitely a stack of primes, better than a stack of primes, the 31 ltd being the only lens in it's FL that exceeds it's performance. The big lie that's being upheld here is the supremacy of primes. There are a few primes that are excellent. There are a few zooms that are excellent. But if you go through the charts you realize, there are a lot of bad primes out there, and no one even wants a stack of those. Calling these amazing lenses a "stack of primes" is an insult to zooms, and a throwback to the archaic days when "prime" meant better. That was around 1972 if memory serves me well. And even then, zooms were way more useful.

[Rondec]

I would be comparing it to any ƒ4 prime in it's class, and many ƒ 2.8 lenses when they aren't being shot at ƒ2.8. I'm not sure why this is so hard to comprehend. It seems like everyone wants to say "my lens is a stack of primes and yours isn't." Calling a Sigma 18-35 ƒ 1.8 a stack of primes would be an insult to zooms everywhere.

So my definition would be,
1. Excellent or very near excellent everywhere in it's range.
2. Excellent or near excellent borders every where in it's range.
3. Prime like control of CA every where in it's range.

Except for 60-70mmm which is a very small portion of it's range the DA*60-250 will achieve that.

It seems to be the consensus among some that zooms can't compete. Well, a lot of them can. They may not be able to compete with every prime in their whole range... but they will also be able to surpass and actually be the best lens averrable in other focal lengths. If there are three focal lengths it can cover that are better than any prime available, I would argue it's a stack of primes. For Pentax mount, as far as I can tell, the 60-250 cover 90-100mm ƒ4 and up as good or better than whatever is available, it owns the 135-180 range. It's equivalent to the DA*200 ƒ2.8, ƒ4 and up and it owns 210-250. If it's as good as the 90-100 primes, better than anything 135-180 and equal to or better than any prime 200-250, how is it not a stack of primes?

And the Sigma 18-35 ƒ1.8, definitely a stack of primes, better than a stack of primes, the 31 ltd being the only lens in it's FL that exceeds it's performance. The big lie that's being upheld here is the supremacy of primes. There are a few primes that are excellent. There are a few zooms that are excellent. But if you go through the charts you realize, there are a lot of bad primes out there, and no one even wants a stack of those. Calling these amazing lenses a "stack of primes" is an insult to zooms, and a throwback to the archaic days when "prime" meant better. That was around 1972 if memory serves me well. And even then, zooms were way more useful.

Let me be clear that I have never used the term "stack of primes" when referring to a zoom. It just doesn't make much sense to me to refer to a zoom this way, but I'm sort of guessing what people mean when they say that.

As to whether zooms are better than primes, it depends on what you want. If you want a small lens that has minimal distortion, a faster than f2 aperture and is flare resistant, you won't find many zooms that fit that bill. On the other hand, my most frequently used lens is the 16-50 purely because if offers flexibility that I don't get when I shoot with only a DA 15. I know that's why you use your 18-135 a lot. It is handy to be able to zoom, even if your lens isn't as strong at some focal lengths.

[normhead]

Let me be clear that I have never used the term "stack of primes" when referring to a zoom. It just doesn't make much sense to me to refer to a zoom this way, but I'm sort of guessing what people mean when they say that.

As to whether zooms are better than primes, it depends on what you want. If you want a small lens that has minimal distortion, a faster than f2 aperture and is flare resistant, you won't find many zooms that fit that bill. On the other hand, my most frequently used lens is the 16-50 purely because if offers flexibility that I don't get when I shoot with only a DA 15. I know that's why you use your 18-135 a lot. It is handy to be able to zoom, even if your lens isn't as strong at some focal lengths.

In any case, I've said enough on the topic.... everyone thinks these things through differently.

[Nicolas06]

@norm,

A 18-55 kit lenses get same MTF results at 35 and 55mm than your 60-250 get. We can see the lense is not that good at 18mm but that's all. One would need at least 2 primes to cover the 35-55mm range so the 18-55 is surely a stack of primes too by your standard. So saying a zoom is stack of prime has just no value at all if all we have for reference are MTF50 on a K10D !

I agree with you on real world use, but make no mistake... On real world use lenses have much more difference than resolution. The way they render colors, their contrast/micro contrast, the in-out focus transistion, performance in contra light... and of course at that for all appertures. There no way an FA render like a DA and even lenses differ between models.

If I put your 60-250 on a istDL the border MTF will be arround 1400 and the center will be 1600 but the pictures it take will continue to look great anyway and you'll be able to print them to 8x12" without reservation! A great lense make the picture look better at full screen or even at websize... And for sure a larger appertures make quite a visible difference that don't need a magnifying glass to check !

But people actually buy primes for many reasons:
- some because they are smaller/lighter
- some because they have larger apperture and better performance at such large apperture... And f/4 can't be considered fast bellow 300mm honestly.
- some because they just perform better in real life: better rendering, 3D pop, resistance to flare, better contrast, bokeh quality
- for all the reason combined.

So well when you get one of the FA ltd you get quite reasonnably small lenses (even if the FA31 is not that small), you get lot of resolution and you get large appertures. You also get very special, better rendering than most other lenses and only a few lenses like high end 35mm f/1.4 or 85mm f/1.4 will get this kind of results. And already many would argue the FA ltd are slow because they could be f/1.4 and some people are really after like 135mm f/2 or 85mm f/1.4.

I'd grant you the 60-250 has great center sharpness wide open and is very sharp stopped down a little. I'd grant you the picture it give are constrasty and with great colors and bokeh is nice... At least when the light is good. But this last part that play a much more important part on how your picture look in practice than ultimate sharpness, MTF50 say nothing of it.

[normhead]

Please, just stop talking about the DA*60-250. It's pretty clear you don't own the lens. When you stick to what you know, I usually agree with what you say.