Canon 100mm f/2.8 IS on 5D III (left) and Sony 90mm f/2.8 on A7r (right). Lensrentals.com, 2015
Roger Cicala from Lensrentals tested the resolution performance of the new and yet non shipping Sony 90mm f/2.8 G OSS macro lens and compared it with the yet best macro lens on market (Canon 100mm f/2.8 IS on 5D III):
At both focusing distances the Sony system is clearly out resolving the Canon system. We would expect that, to some degree, given the higher resolution Sony sensor. But the difference was, quite frankly, surprising to me. It would seem to indicate the new Sony 90mm Macro lens performs at least as well as the Canon 100mm f/2.8 IS L, from a resolution standpoint, anyway.
See the MTF chart comparison:
That sounds terrific! If you want to see a few image samples jump over at Gustavs images on SAR.
The 90mm lens will ship from July 7 at those stores:
In USA at Amazon, Adorama, BHphoto.
In Europe at Sony Germany, Sony UK, Sony FR, Sony IT, Sony ES, Sony NL, Sony BE, Sony AT, Sony CH, Sony SE, Sony FI, Sony NO, Sony PT.
In Asia at Digitalrev, Sony Australia.
Zeiss Batis preorder in USA and Europe (UK/Germany)
Zeiss Batis 85mm f/1.8 lens in US at Adorama, Bhphoto. In Eu at Amazon DE, WexUK. In Asia at Digitalrev.
Zeiss Batis 25mm f/2.0 lens at Adorama, Bhphoto. In EU at Amazon DE, WexUK. In Asia at Digitalrev.
Lens shade for 85mm at Adorama, Bhphoto.
Lens shade for 25mm at Adorama, Bhphoto.
130 Responses
That difference is stagering! Even though the Sony 90mm lens is tested with a higher resolution sensor, I wouldn’t expepected to see such a difference.
He should have adapted the Canon lens to the a7r to eliminate the sensor variable.
That’s what I was thinking too
Yes, why not just do that, that’s a no brainer
Good point. But Roger already wrote in another article why he doesn’t.
Because then people will complain that it isn’t fair, because of “adapter tolerances”.
Or use the A7II as well as a third measurement
So there is my macro lens FE!!! Damn i think i won’t have money enough! 25mm batis and this one! Maybe i m’ gonna pass on the batis 85.Hope the new camera won’t be too sexy!!!
I am considering this lens, but I do not own an E mount camera.
If Sony comes out with a 50ish mp E mount camera and an updated lea4 I might get the lens, camera , and adapter.
If Sony makes the camera fully high speed af compatable with Canikon lenses and A mount without mirror…there is absolutely no doubt I would get them.
I expect Nikon to anonce a ff mirrorless F mount camera end of this or begining of next year.
Assuming it uses the same sensor and Sony does nothing about the compatability issue this the camera I will get as well as a few F mount macros to go with it.
I hope to get my Sigma 150-600 sport for Nikon fairly soon, so with two systems I am unlikely to get E mount unless compatability issues are addressed.
At this point I have no idea what Sony will do and until I do I have not been buying any new Sony equipment.
I think Sony has scared many potential buyers away with thier lack of any discernable roadmap.
>I think Sony has scared many potential buyers away with thier lack of any discernable roadmap<
Rubbish, only in your mind! Canon have zero, zip, nada roadmaps and they're the top selling DSLR for new buyers bar none. BTW, no Nikon mirrorless is on the cards with an AIS mount……..
As with many e-mount lenses, the edges look very poor. Wonder how much they improve on stopping down.
Yeah, I was surprised by that. There’s almost a 50% drop in resolution from the center to the corners.
Which e-mount lenses? Same goes for the Canon there, btw.
The Sony corners are on par with the Canon average, I don’t find the 90/2.8 “very poor”.
No….. The A7r has nearly double the resolution of the 5diii — and therefore, if it was “on par,” one would then expect MUCH higher resolution numbers out of the Sony. They look to be “on par” in the center, not in the corners.
On the Canon, the corners have about a 30% drop off from the center, in resolution.
On the Sony, the corners have a drop off of about 40%. In other words, the Canon corners hold up much better than the Sony corners.
It’s a shame I can’t find objective testing on the Sony/Minolta 100/2.8, which may be the best of the bunch.
The myth…the legend.
Looks like you stole my copy, lol. I had the case, etc. Really a fantastic old lens.
Just bought one of these a couple months ago – great lens.
Actually you wouldn’t expect that.
The better the lens, the more improvement you get with a higher MP centre.
A corollary, is that the better the resolution at a location the more improvement you get.
Any lens is better centrally than in the corners, so you should expect a larger imprvement centrally than in the corners with a boost in sensor resolution
No.
First, you’re confusing resolution and pixel count. One is linear, the other is the total of both axis mutiplied. The sensor resolution of the A7R is 28% higher, measured corner resolution is 25% higher, measured center resolution a whopping 44% higher and average over the frame it is measured to be 37% higher.
Second, measured resolution often does not scale up linearly with the sensor resolution increase, in part due to optical limitations.
Like you said the Sony sensor is bigger, so only 10% difference is not that huge, is it? (Just asking)
The Sony sensor isn’t larger. The a7r is much higher resolution than the 5diii. The numbers would be totally different if you compared the 5ds to the a7s– then the Canon numbers would be tremendously higher than the Sony numbers. So you need to equalize the resolution to really determine the quality of the lens.
Whether it’s a big difference or not… Hard to tell from the limited testing here. Probably not a huge deal. Just slightly softer corners, which isn’t a big deal in most cases.
My 200,000 plus Canon files speak differently. Sony FE glass (with the exception of the FE 4/24-70) is on a par. Sony’s FE 1.8/55 is an oder of magnitude better than any Canon L prime under 85mm.
Roger just can’t bring himself to say the Sony is the better lens, old biases die hard.
How exactly would he even be able to back up a statement like that when the lenses were tested on two different sensors?
If you read the article, Rodger started with the false assumption that the Canon lens is so much better that even using the inferior camera/sensor that the readings should be about even. Since it didn’t work out that way he was forced to report the bad news that the Sony camera/sensor combination exhibited far superior results. The test methodology is in fact flawed, so why did he publish it in the first place. I trust legitimate testing will show the Sony lens to be superior, but we will have to wait to confirm that.
“If you read the article, Rodger started with the false assumption that the Canon lens is so much better that even using the inferior camera/sensor that the readings should be about even.”
That is an extremely dishonest characterization of what he actually said.
“Since it didn’t work out that way he was forced to report the bad news that the Sony camera/sensor combination exhibited far superior results. The test methodology is in fact flawed, so why did he publish it in the first place. I trust legitimate testing will show the Sony lens to be superior, but we will have to wait to confirm that.”
There’s no possible way you can ascertain that the Sony lens “exhibited far superior results”, there’s simply not enough data. The only thing this test could really tell would be that if the two lenses gave similar MTF values then we might know that the Sony version was lagging behind the Canon.
That you think Roger is the biased one here is actually quite hilarious. You should read what you’ve written.
Another brokenhearted fan boy grasping for straws, get over it.
“Another brokenhearted fan boy grasping for straws”
I’m glad you were able to see your own posting for what it was.
Now dry your tears, it’s only a lens.
I don’t really expect someone like you to be able to be able to grasp these concepts, but the difference he’s presented between the two lenses, ~40% sharpness advantage in the center for the Sony, is almost identical to what he measured when he put the Zeiss 35 f/2 on both the 5D2 and A7R. The Zeiss may very well be the better lens, but the fact is that there’s not enough data available to say anything other than that it’s not worse than the Canon.
The data was enough to invalidate his false assumption, which in its self indicates a bias. My point is Roger likely knows the Sony lens is superior, but being a Canon fan boy couldn’t get the words out. That’s my interpretation of his testing, you may disagree; that’s your prerogative. EOF.
It seems pretty clear at this point that you simply don’t understand what’s in the article so I’m only going to say this one more time. The new Sony macro may very well be superior to the Canon version, but there’s simply not enough data to support that conclusion yet. Given the data available, there’s no way Richard could honestly say that the Sony is superior, only that it’s not worse than the Canon in terms of center resolution.
No, he doesn’t. He just said that in the case the Sony were inferior to the Canon, similar performance would be seen, which is not the case. He never assumed the Sony to be inferior per definition.
The Canon 100 Macro is probably one of the best macro lenses for any money, in simply getting close Sony are excelling!
I think Roger is too old to give a shite about “brand loyalty.” 🙂
Run this test again when the 5DS/R comes out.
…and Canon will come first. We all know how to calculate optical system resolution and this test is not of much value.
as i told before , this lens is really suberb
I remember Otus 55+D800E only got around 1300 lp/ih in Roger’s test and now both Canon and Sony can easily kick Otus’s ass.
The better way to test would have been to do so on the same a7 camera.
Indeed! They should have used at least cameras with similar sensor resolving power A7 II vs 5d mkIII would be closer.
If he did’t it has to have a reason why.
Perhaps to make things better as they are?
Or with an adapter, easy.
Canon owners can’t accept that sony is now making great lenses too.
Nikon owners also. XD
This is not true. We can….
where is the size advantage…
As i told several months ago, FE system will be the highest lens quality system in the world in 3-4 years. I was wrong only in that it will be much sooner than i predicted.
Thank you Nostradamus.
Then what?
Well, in that moment everybody was acting like that was absolute nonsense. Now a lot of those same people praise the system. Hope those recognize themselves.
On my A7R the Canon 100/2.8 is an extremely good lens
Agreed, I shoot the Canon 100L lens on my a7 and it is actually my favourite lens in terms of colour and rendering. Not just for macro either but for general purpose uses as well. When looking through my pics after a days shooting, I can always instantly tell when the 100L pics start as they have such a distinctive look to them. Regardless of sharpness or auto focus, if this new Sony macro doesn’t render as good as the Canon then I’ll stick with the old.
I don’t believe that these numbers are correct.
It would result that this lens would have poorer corners than almost any other macro in this focal length range. A macro lens with a drop of roughly 40% to the corners has a serious problem.
So apply a 40% blur filter to the centre of the image – that will get rid of your horrible problem that the centre is sharper than the corners, and it will still come out ahead of the Canon across the frame.
I can’t believe that you think that a system that gives you a ton more resolution is the centre and “only” a lot more resolution in the corners is a bad thing. Some whiners are never happy.
You can’t just think in percentages: when the absolute centre resolution is as high as this, you can’t expect the centre to be remotely as high – that’s just not going to happen in a lens of this order. The absolute corner resolutions are great!
Usually I don’t give much on pure technical evaluations.
But then it still is a hugh drop compared to the center which will be quite visible in situations where you need the corners to be as good as possible.
But percentage is useless if you can’t perceive the drop. For example if the bottleneck is in the sensor. The Sony G lens would appear more uniform and still have vastly higher absolute numbers.
If the farthest corner of the Sony lens is sharper than the entire average of the Canon lens, it should say a lot by itself…
Agree.
However it is more an assumption at this time.
I have pre-ordered long time ago, and will test for myself. Not with charts but by real life usage. 🙂
I really doubt it will be visible; those corner resolutions are at the centre levels of some very fine lenses! So the whole field is at a level that’s high enough for the human eye not to be able discern differences even at large print sizes. So the system resolution is above what you need or want for ordinary purposes. However if you want to use a crop mode for reach, then the extra central resolution could come in handy and be visible…
I would normally agree with you, but to be honest, you shouldn’t care about “corners” on a Macro Lens… unless you shoot flat subjects like stamps or something… most likely the corners will be Bokeh anyways…
I would gladly trade center & Mid area sharpness over corner performance on any Macro Lens.
Roger saw almost the exact same difference, a bit less than 40%, when he tested the Zeiss 35 f/2 on both the 5D2 and A7R last year so it’s actually quite consistent with previous findings.
There is a clear size and weight advantage on the right side. Unless you are blind.
Does anyone know if the Sony FE 90mm G Macro lens is sharper than the Zeiss FE 1,8/55 ZA lens with Macro tubes?
The FE 1.8/55 is a superb lens used as it was designed at normal distances, but stick some tubes on it and it’s not too hot. I’d suspect most any lens designed for flat-field resolution would perform better at close distances.
Please compare this to the Minolta 100mm 2.8. That lens is a legend and deserves to go against the modern big boys.
VOLUME, not size! Body PLUS lens’ = total volume. In layman’s terms, see how much Canon crap you can shove in your Sony pack.
The 16-50 on aps-c, horrid corners. The 24-70. The 70-200 has horrid corners at 200/4. The 35/2.8. Etc.
As we know, the e-mount lenses have more distortion than traditional mounts. I could see how this could be causing corner softness
Canon’s 70-200 ll has terrible corners at 70, 105 & 200mm. I think you’ve not gotten enough time in with the other guys glass. The only let-down in Sony’s FE glass is FE 4/24-70 @ 24mm, that’s it, you’re stretching a point too far here.
What are you talking about?
http://www.photozone.de/canon_eos_ff/431-canon_70200_4is_5d?start=1
The Canon 70-200/4 corners hold up quite well at all focal lengths.
Why must fan boys claim total superiors across the board in their beloved brand.
Fact is, weaker corners is a common theme across many Fe/e lenses I see tested. Of course, ALL lenses have worse corners than centers, but this seems more pronounced across most FE/E lenses.
Stop “quoting” charts. Some of us have extensive professional experience with Canon glass which is pretty much industry standard & superb.
Sony’s FE glass is on a par, to a point where amongst 200,000 plus Canon files, it’s indistinguishable. Like I said before, the FE 4/24-70 @ 24mm is lacking but the FE 1.8/55 is an order of magnitude better than any Canon prime < 85mm.
Takes your FE & L glass glass, goes shoots some stuffs and be prepared to be pleasantly surprised – that's all fffolks!
Yes…. Let’s ignore objective testing and just take your subjective word for it.
Not questioning that there are some exceptional e-mount lenses. But compared to traditional mounts, they seem inclined to show weaker corners. Unless you think the testing is just a huge conspiracy of lies.
You got any idea how objective scientific testing by non-photographers stacks up against real-world subjective analysis by time-served non-fanboy realists…………obviously not!
How many Landscapes will you shoot with this lens at f/2.8, where the result of this “objective test” would be valid? How many macros will you shoot at f/2.8 that will require absolute sharpness in the corner? Most likely zilch, Mr. Armchair Photographer.
Lol. So defensive. Did I say it was a big deal? In fact, I said it makes me curious to see the results when stopped down. I made a little observation, and the fan boys erupt in defensiveness.
“Very poor”, “horrid”. Why not choose appropriate words if you want to say “little observation, no big deal”? How can you say “the edges look very poor” without even seeing an actual image made with the lens?
Where did I say the 90 looked horrid???
We have little to go on so far. I commented based on the limited evidence we have thus far, and openly wondered how things would look when we get more evidence. If we aren’t going to comment on the incomplete evidence, then why post the lens rentals testing at all?
If the 35mm f2.8 has “horrid corners”, the Canon with the sharpest corners (35mm f/2) surely should be called a disaster, since it performed worse, even on the same sensor where the Sony peformed better in the corners wide open than the Canon at F4.
More objective measurements for you:
http://www.lensrentals.com/blog/2013/12/sony-a7r-a-rising-tide-lifts-all-the-boats
I’m not familiar with the 35/2.. But that’s not an apples to apples comparison, as you’re comparing f2 to 2.8. How does the Canon do at 2.8 compared to the Sony?
Regardless… I’m not here saying that every Canon lens is brilliant and better than every Sony lens. Far from it.
I made the very limited — and true point — that several e-mount lenses show weaker corners than comparable lenses.
As to the battle of the 70-200’s… I point you to the photozone charts that show a pronounced weakness at 200/4 in the corners compared to Canon and Nikon. Outside of that focal length and aperture, results are pretty comparable.
Considering the shorter flange distance, as well as efforts to keep the size down on many e-mount lenses, it shouldn’t come as a surprise that some optical compromises have been made.
Did you not read? The Sony showed significantly (24%) more resolution even in the corners at f/2.8, than the Canon at f/4, using the same sensor. Let alone using the same aperture (be it f/2.8 or both at f/4).
None of the other 35mm lenses from Canon does as well in the corners as the 35mm f/2 IS at equal apertures.
http://www.dxomark.com/Lenses/Compare/Side-by-side/Canon-EF-35mm-F2-on-Canon-EOS-5D-Mark-III-versus-Canon-EF-35mm-F2-IS-USM-on-Canon-EOS-5D-Mark-III-versus-Canon-EF-35mm-F14L-USM-on-Canon-EOS-5D-Mark-III___474_795_1086_795_797_795
Did you not read?
Example, the Nikon 70-200/4 vs Sony 70-200/4:
http://slrgear.com/reviews/zproducts/nikon70-200f4vr/ff/tloader.htm
http://slrgear.com/reviews/zproducts/sony70-200f4gfe/ff/tloader.htm
(They don’t have a recent full frame test of the Canon 70-200, trying to stick to apples v apples here. Comparing the Sony to the Nikon, you see at long focal length, the Nikon has much better corners)
Sony 24-70/4:
http://slrgear.com/reviews/zproducts/sony24-70f4zfe/ff/tloader.htm
Canon 24-70/4:http://slrgear.com/reviews/zproducts/canon24-70f4l/ff/tloader.htm
(No comparison)
Want to use reviews done by the same methodology… so switching to photozone:
Sony FE 35/2.8: http://www.photozone.de/sonyalphaff/865-zeiss35f28ff?start=1
Very weak corners… The closest Canon is probably their 40/2.8:
http://www.photozone.de/canon_eos_ff/752-canon_40_28_ff?start=1
Again, much more consistent across the frame.
Sony 16-35…. Very impressive results, overall.. but corner weakness, with very poor corners at 24mm, and horrible at 35
http://www.photozone.de/sonyalphaff/897-zeiss1635f4oss?start=1
Canon 16-35:
http://www.photozone.de/canon_eos_ff/877-canon_1635_4is?start=1
Similar.. but not as extreme as the Sony. Particularly, not nearly as bad at 35mm.
In fairness, the Nikon 16-35:
http://www.photozone.de/nikon_ff/492-nikkor_afs_1635_4_ff?start=1
I hate the Nikon 16-35… It’s corners are just as bad at the Sony, though at different focal lengths.
Switching to an APS-C example:
The current E-mount kit lens, the 16-50:
http://www.photozone.de/sony_nex/842-sony1650f3556oss?start=1
Dang those corners are horrible!
Compared to the current Canon kit lens:
http://www.photozone.de/canon-eos/181-canon-ef-s-18-55mm-f35-56-is-test-report–review?start=1
Now, these types of comparisons are NOT true of EVERY lens. As shown, the Nikon 16-35 corners are just as bad as the Sony corners. The Sony 55/1.8 is amazing in the limited testing I’ve seen, certainly far better than most (if not all) 50ish/1.8 lenses. The 35/1.4 appears to be comparable to the other 35/1.4 lenses on the market. And unfortunately, there is still a limited supply of lens testing out there.
But when looking at apples to apples comparisons, in most cases, we seem to see a pattern of weaker corners on the Sony E-mount lenses. It’s true in APS-C and in full frame. Not universal — not every lens. And not every focal length on every lens.
It’s cool that you still cherry pick your results to favor your agenda and categorically ignore what I presented, including tests on the same sensor (where the 35mm trashes all competitors except the much larger Sigma, which performs similar in corners).
Plus you’re still in denial about the fact that the 16-50 is an extra option that neither Nikon or Canon even offers, yet you do use the “apples vs apples” reference more than helps your argument. Ironic.
So yeah, I can read, but for as long as there are several sources contradicting or even debunking your general claims, you don’t seem to be capable of the same thing.
It’s also telling that you ignore results of other lenses such as the 28mm f/2, 35mm f/1.8 OS and 50mm f/1.8 OS, which all do as good or better in corners than anyting from the competition in the respective price ranges.
Id be happy to go through every test of every lens.
The e 18-55 also has horrible corners:
http://www.photozone.de/sony_nex/724-sony1855f3556oss?start=1
The 50/1.8… Horrible corners until stopped down to f4.
http://www.photozone.de/sony_nex/720-sony50f18nex?start=1
I stick to facts. I haven’t claimed this is universal — the 55/1.8 is stellar across the board. The 16-35 is very comparable to the competition.
But across many e-mount lenses, compromises have led to some corner weakness. You’re getting all defensive for absolutely no reason. I’m not claiming that Sony lenses stink, or anything like that. I made a simple and truthful observation about a minor shortcoming across many e-mount lenses (including the 50/1.8, the 18-55, and more).
Why would I be defensive about a system I do not own or have little interest in? You’re projecting too much.
I stick to the facts and point out your clear cherry picking, fallacies and incorrect statements. See my posts above. Two more links to Photozone (single lens tests) will not change the facts that I posted.
Because again, they can be easily contradicted:
http://www.dxomark.com/Lenses/Compare/Side-by-side/EF50mm-f-1-8-II-on-Canon-EOS-700D-versus-Sony-E-50mm-F18-OSS-versus-Nikon-AF-S-Nikkor-50mm-F18G-on-Nikon-D5300___187_870_745_0_435_919
Relatively good corner performance at f/1.8 already, much better than the Nikkor and Canon 50mm f/1.8 lenses.
Are you blind? Yes… Don’t let actual evidence interfere with your preconceived biases.
Ad hominem is all that is left when your argument fell flat a long time ago.
Are you now denying that Photozone test that I linked to shows that the 18-55 performed better in cornerd on a 16MP sensor than the Nikkor equivalent tested on a 16 MP sensor? Keep your commend above about blindness in mind.
Are you suddenly denying SLRGear results contradicting your earlier claim about the 50mm f/1.8? And the existence of DXOmark results doing the same?
All that you accuse others of, hits back like a boomerang.
What are you talking about? Photozone hasn’t even reviewed the current Nikon 18-55 yet, nor have they reviewed the current 50/1.8 on aps-c body. In terms of current kit comparisons, they have reviewed the current Canon 18-55, much much much more even performance than the e 18-55.
I stand by every comparison I’ve made:
http://slrgear.com/reviews/zpr…
http://slrgear.com/reviews/zpr…
(They don’t have a recent full frame test of the Canon 70-200, trying to stick to apples v apples here. Comparing the Sony to the Nikon, you see at long focal length, the Nikon has much better corners)
Sony 24-70/4:
http://slrgear.com/reviews/zpr…
Canon 24-70/4:http://slrgear.com/reviews/zpr…
(No comparison)
Want to use reviews done by the same methodology… so switching to photozone:
Sony FE 35/2.8: http://www.photozone.de/sonyal…
Very weak corners… The closest Canon is probably their 40/2.8:
http://www.photozone.de/canon_…
Again, much more consistent across the frame.
Now, we could go through every test of every lens ever made for every camera. And we would find examples where CaNikon lenses have soft corners (especially if looking at older lenses). You will find examples where Sony lenses are absolutely exemplary across the board — as with the 55/1.8, the 35/1.4.
But in many cases — as with the kit lenses, the 16-50, the 24-70, and others, Sony has shown a willingness to compromise corner performance likely as a size trade off.
You want to be a fanboy and yell that every single Sony lens ever made is superior to every single CaNikon lens ever made… go for it. I’ve posted a ton of actual evidence. And I suggest anybody who wants the actual truth, simply look at the actual evidence. You will find cases where Sony lenses are superior, you will find cases where they are equivalent, and find cases where they are inferior. And in some of the cases where they are inferior, the weaknesses may not even matter to the shooter.
If you put as much dedication into your reading as you did into your cherry picking of test results, you’d be able make a bit more sense.
I didn’t say Photozone tested the 50mm f/1.8, I said the SLRGear test contradicts your claim that the corners are bad, compared to the 50mm f/1.8 lenses from Nikon or Canon. It’s rather “surprising” that you didn’t mention that, since you like to reference them so much. They showed corner results as good to much better, which was confirmed by DXOmark (testing multiple lenses, not one like Photozone). The last year released upgraded Nikkor kitzoom does appear to be better, but that’s to be expected, 4 years to improve on the Sony design.
But you’re still ignoring the glaring contradicting evidence (especially more relevant when using multiple copies and comparing on the same sensor) for the lenses mentioned (including the 24-70 and 16-35). Or you attempt to dismiss them (like the DXOmark results) based on a lack of understanding of their measurements (they have MTF graphs too, to evaluate relative performance over the whole frame, the perceptual MP number isn’t that relevant here). Making some generalizations based on cherry picked results, won’t convince anyone here. Neither will cute fallacies such as below:
“You want to be a fanboy and yell that every single Sony lens ever made is superior to every single CaNikon lens ever made… go for it.”
You won’t be able to back that up with actual quotes of mine, since I didn’t claim any such thing and already stated I’m not that interested in owning E mount cameras. I said your generalizations don’t hold much water, which is backed up by many other tests, some of which more relevant too.
Seems like your emotions are in control, rather than reasoning based on what I actually wrote and referenced.
And just one note…. Be careful with dxomark testing, as they are showing the results of sensor+lens, and their methodology makes side by side testing difficult. The higher resolution bodies always end up giving the lenses much higher scores.
Uh, all the tests you have linked to are a result of sensor+ lens too.
At least DXOmark shows us a more detailed results over the whole frame at the respective apertures and focal lengths and tends to use several copies, averaged. Where Photozone does not. Roger Cicala also averages results over several copies and tested them on the same sensor.
Yes, but slr gear does. Problem with dxo, they are testing combined lens+sensor, and then creating a low threshold to score “excellent,”– therefore, if the body has high resolution, the lens always scores highly. photozone, slr gear, lens tip, consider how the lens does relative to the body. Dxo mark scores any perceptual resolution over 12mp as excellent, whether the testing body is 50mp or 12mp. Once they start testing with 5ds bodies, miraculously Canon lenses will all get amazing scores. While right now, Nikon and Sony lenses get the best scores — because they are testing on 36mp bodies.
I never look at the perceptual MP scores, which do not tell us much if anything about corner performance (the subject here) to begin with. I was obviously referring to their (measured) MTF charts. And those don’t magically show improved corner performance (better sharpness uniformness) when testing on a higher MP body. See the D810 vs D600 results.
And based on everything that I have read (and their results regarding corner performance), SLRgear does not average their results over several copies of the same lens. One lens, several shots averaged.
You mean the pancake kit 16-50? I hope you are not serious with that.
As to the “horrid” corners, I have some difficulty understanding how you measure that, but if you are going with your idea (as above) of measuring percentages between center and extreme corners and then judging it based on that, then you really have no idea what you are talking about.
Sure but that’s part of the compromise between size and performance.
And there’re good performing lenses too: For APS-C the 24/1.8Z, 35/1.8 and 50/1.8, even the 10-18/4 or Zeiss 12/2.8 for their aperture. On FF the 28/2, 35/1.4Z, 55/1.8Z and all Zeiss lenses … But you better avoid zoom lenses if you want best IQ.
Adding the adapter to the Canon makes it about 2cm longer! AND it’s considerably wider. That makes it noticeably less compact on the Sony than the Sony G
Since the 90/2.8 was tested on a much higher resolution sensor, the difference is exaggerated. I wish he’d used the A7 instead of A7r to “level the playing field” in that regard. However, the numbers are far enough apart that I think the 90/2.8 would’ve still come out ahead in that case.
Astute.
It is a bit funny that in their second test the Sony 90mm G corners (1380 lp/ih) out-perform the Canon 100/2.8 L average score for the whole frame (1290 lp/ih)
🙂
If you look at the linear resolution difference, you expect the Sony lens to provide 1.27 times the resolution of the Canon. Then there is an additional factor accounting for the lack of AA filter. The Nikon D800E delivers about 1.15 times higher lp/ph than the D800 according to tests by lensrentals. So one can expect a resolution higher in the order of a factor of 1.47. According to this, the center resolution behaves accordingly almost perfectly, the corner resolution should be a bit better on the A7r, however. Therefore, optically my guess is that both lenses should be pretty much on par in terms of sharpness, with a slight advantage in the corners for the Canon.
I applaud your math skills, but you are giving the Canon equipment too huge benefit of the doubt. You are assuming that, if the 5DIII had a higher resolution sensor, it would match the Sony combo in the center and outresolve it in the corners, but there are absolutely no facts supporting this statement, just positive statement.
The fact is A7R + 90/2.8 G out-resolved the Canon combo in all areas in this test today, and if I am looking for one such macro set-up today I should be buying the Sony combo.
Of course, this is just a fast estimate of what I would expect. It would be weird , if the AA-less A7r wouldn’t outresolve the Canon combo. My calculation just shows that it is within expectation based on simple available data we have. Personally, the 90mm looks to be a very interesting lens for me, although I would have opted for a larger focal length (like 150mm Sigma). But this would result in a large and heavy lens. I own the Tamron 90mm VC, which is a very good lens. But manual focussing is not as nice and I expect the Sony to be better here.
Except that Canon tends to use very light AA filters to begin with these days, where as the D800 (non E) does not (relatively strong).
Uh, not really. The AA filter in the 5D3 is fairly strong, the original 5D was the last FF Canon body to have a weak AA filter.
The relatively high levels of moiré (in RAW) clearly disagree with you. As do resolution measurements, which come close to the theoretical maximum with the best lenses, despite the AA filter.
The AA filter is actually quite strong, see Philip Bloom’s before and after with his ‘HotRod’ 5D3.
I have seen his comparisons and find them unconvincing, both in terms of visuals as reasoning.
I don’t see a big difference in sharpness, despite the much more contrasty light in his “after” screen grabs. And that shouldn’t be much of a surprise either since the sampling in video is at a much lower resolution to begin with. 95% of the softness is due to low resolution read out in video and low quality (limited by processor mostly) blurring to hide moiré and strairstepping (due to low res read out/ line skipping).
Just for reference, there’s no line skipping in the 5D3 video. When you looking at the RAW video output before the DIGIC processor gets to it (through MagicLantern), it’s actually extremely sharp. Canon just decided to rub vaseline on the 5D3 video, most likely to protect the Cinema cameras.
In any case, even just looking at DPReview samples comparing the 5D3/D800/D800E/D600 it’s plainly obvious that there’s much less aliasing/moire in the 5D3 than the D800/D600 which would suggest that the 5D3 at least has a stronger AA filter than either of those bodies.
That is one big mama of a 90mm lens! 😀
It was mentioned above by Lee that the comparison was made on two very different sensors, one with and one without AA filter, and one bearing 0,59 x the resolution of the other (5DIII vs A7r). The mean difference in benchmarks between them mostly chops up to the differences in sensor. The Canon lens should be tested on the A7r next to the Sony. That would make it a fair comparison.
True. I factored in the resolution difference and an estimated effect of the AA-filter (D800/D800E) in a comment above, resulting in comparable sharpness, however.
Misleading test. Why not testing the 90mm lens on a A7II? Or the Canon 100mm on the a7R?
I think everyone on this forum should know already that lens has higher scores when testing a7R or D800E
the resolution difference between two body in one linear is 1.27X（not the total pixel, but length）but the score for each lens has an average difference of 1.36X，or at center 1,45X.
and for macro photo, people more likely care about center. meaning the 1.45x difference is more practical
so give up on this canon fans, 1.36 is larger than 1.27, no matter what calculation or what body they use for the test, Sony still wins.
Would it not have been a simple matter to mount the Canon on the A7r and thus give us exactly
comparable results? Different cameras with different sensors with different AA
strength and so on introduce way too many variables. I wonder how we would have
moaned if they had tested the Canon on their new 50mp model ?
Wow, so the day when Canon fans are saying ‘if you want to get maximum performance from this Canon lens slap it on a Sony body’ have finally arrived!
Explaining how a Canon lens got a lower score because it is being tested on a Canon body 🙂 or if you want to make the lens test fair you should use a poorer Sony body to give a chance to Canon 🙂
Nothing to do with poorer body. Just different resolution. Is the a7s a poorer body than the a7? Lenses will score much higher resolution on the a7 than a7s… Guess the a7s is poorer?
Canon produces the best Macro lens? Whose been smoking the wacky baccy? The best 1-1 macro lens is the Printing nikkor 105mm 2,8. The canon is a blurred contrast less piece of bottle glass in comparison.
I haven’t used the Canon macro. But from experience, I can say the Minolta/Sony 100 macro is better than the Nikon 105. (Which I use, and it’s not bad).
I wasn’t referring to the standard Nikon 105mm macro…I repeat the best 1-1 macro lens is the “Printing nikkor 105mm F 2.8” It has zero distortion and no CA at 1-1 and it operates best at F2.8 no other lens can match it in that department I repeat I not referring to the standard Nikon lens. The lens i am talking about is produced by Nikon Rayfact
I’d probably give the award of best macro lens to the Coastal Optics 60mm f/4 macro lens.
Yep thats another lens that beats the canon…but My understanding of the MTF charts the Printing Nikkor may have the edge
It´s nice indeed that Sony build such a great lens. But 1200€ for a 90mm macro lens? This is the generel thing with Sony FE. The bodies are cheap but the lenses are expensive. Sony builds best lenses but for what Money? e.g. 1300 € for wide angle (16-35 f4) , 1900€ for 35 1.4, 1200€ for a 90mm macro. I think for pretty much People (at leaset for me) it´s like this lenses wouldn´t exist, because they are simply not affordable. This Limits the number of affordable lenses for FE. Why not a f4 wide angle for around 800-900€, a 35mm f1.8 for 700-800€ or a macro lens for 600-700€? Nevertehless great work Sony! They keep the market interesting with there inovative products!
True but why you need FF if you want cheap/budget lenses? Wouldn’t APS-C be more suitable for your budget? There’re lenses <1000 €. Also very good ones. If I'd buy a body, I also include follow-up lens costs.
On the other hand, A mount is a very matured system that grew over decades (from Minolta). Sony can't deliver on all fronts (FF and APS-C, budget and expensive). And they're still alone (no Tamron, no Sigma – or very few) with some Zeiss support.
Macro lenses belong to the most uniformly rendering lenses of their respective systems. The Sony combo doesn’t seem to live up. Center resolution is great though.
You’re forgetting the effect of the AA filter. The difference between the two in the center is consistent with the results Roger posted when he compared the Zeiss 35 f/2 on both the 5D2 and A7R. The new lens is most likely better, but it’s simply too early to be able to say anything other than that the FE lens is not worse than the Canon in the center.
These numbers are using different test charts than Otus 55 tests. Absolute comparisons between those numbers are meaningless.
Point is, you didnt have to be nostradamus to see that, but you really had to be extremely stupid to constantly criticize a year old (which consisted of several lenses, among them one best in its class) system for the lack of glass.
You opinion seems to be a bit contradictory. On one hand you don’t like the big difference, but on the other hand you want the center to be as sharp as possible.
If the center was reduced to 1600 lp/ih, it would appear more “uniform” as you wanted, but that wouldn’t make the lens better. It would make the lens worse.
Hello Boiz, true if I but a body I consider lens costs. That´s what I meant and that´s why I unfortunately didn´t buy a Sony FF (but a Canon). I don´t think a Nikon or a Canon 16-35 f4 is a Low Budget lens which I get for around 900€ and I still would have also cheaper possibilities (e.g. 18-35 f3,5-4,5 Nikon and 17-40 f4 Canon). Not to speak about the 35 1,8 G from Nikon which is a quite good and alos affordble FF lens. Ok it´s currently difficult for Sony, right. It´s a quite new system (E-Mount). But from a customer point of view I can only consider the currend market situation. 1200€ for a 90mm Makro? And 90mm on FF it´s quite limited for Makro photography.
If you pull in Nikon or Canon, you need to compare them to Sony A mount. There you have all kind of lenses at all kind of quality and price levels. Usually, all prices are very similar. Nikon also provides the AF-S 35mm 1.4G for 1500 €. Their 85mm 1.4D is also above 1000 € and their AF-S 85mm 1.4G is 1400 €. It depends on what quality you’re asking for. The problem with E mount is, you don’t have many options. And Sony opts for high quality (almost) only.
Time will tell …