Lens softness - is this the best I can expect?

Pete_W

Senior Member
Messages
4,186
Solutions
7
Reaction score
5,496
After seeing consistent softness on the right hand side of photos taken with my PL 12/1.4, I decided to take it back to the store on Friday and get it swapped for another copy.

I took the new copy out this weekend for some test shots to see if it was better. Unfortunately there is still softness on the right hand edge.

I've pasted some examples below. The first two were taken with wider apertures while the third was stopped down to f/5.6 for comparison and this obviously helps sharpen up the edges.

But that kind of defeats the purpose of having an f/1.4 lens.

Am I being too picky? Is this what I should expect for a $1,500 lens?

0c7e5f8ce8e24c28bcd6a13bbb8f0af7.jpg



0735020791f24c6e8e3608bee68622aa.jpg







3f72fea5b8454473901249f49764f2ac.jpg







10369ddf4ffe4bb3b467094e5a267128.jpg

--
Pete
 
Solution
Thanks to everyone who contributed with helpful suggestions and comments, it's really very much appreciated.

I spent some time last evening reviewing everything that I've shot with this lens and also looking at the sample images online from reviews (e.g. LensTip which matches my experience) so I made the decision to return it. I simply didn't want to spend even more time testing and reviewing, and felt that for the price of this lens it just wasn't good enough.

The camera store were happy to exchange it so I've swapped it for a Panasonic 7-14mm, a lens I've been interested in for some time. They didn't have the PL8-18 in stock yet but having seen some comparison reviews, I think the older less-expensive lens is more suited to my...
This f/1.4 Panasonic for $1300 is one of those m43 lenses that don't make any sense.

For only $450 you can buy the Sony 28mm FE f/2.0 that has gotten really great reviews for it's sharpness, even wide open, and it's equivalent to f/1.0 on an m43 camera. Plus it weighs less than the Panasonic monstrosity.

So for a little more than the price of just the Panasonic lens, you can buy a FULL FRAME Sony A7 II camera.
The Sony lens is very good, in fact, I think it's the best lens in their line-up for quality/price ratio. But unfortunately even I, who shoot with very few primes, can't limit myself to one lens.

A7II is a dog, like the A7 before it. They shoot like cameras from 2001, they are slow and clucky in operation and autofocus. I had an A7 (just for the fun of it) and it was the worst ILC I ever put my hands on. I see people comparing the top m/43 cameras with the low end FF cameras, and I find those comparisons meaningless.
I totally agree about that.

Before I went fully to mirrorless, I went to try so I was ready to go to Sony. So I bought the A7 and so on. Only to sell it off after couple months as it didn't perform at all next to Canon as should have.
You're right when you say the QC is at an all-time worse. All manufacturers seem to have serious issues with both lenses and cameras. Yes, it's outrageous that happen on a $1300 lens, but check the other forums to see how often it happens on $2500 lenses. The answer is: VERY OFTEN. And that is a shame.
Yes. And DSLR rear/front focusing is very often happening problem as well, even way more often than manufacturing quality control problems.
I'm surprise people here keep blaming the wide aperture, field curvature, or how wrong it is to shoot landscapes wide open (it isn't), but what I see is the right side looks like it was photographed through a coke bottle bottom.
Very clear decentered as both sides are soft and center is sharp.
The only think I'm not sure if it is the lens, or the IBIS system causing this.
If it would be IBIS, then it would be with every lens.
Let's wait for the OP tests without IBIS
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?

Andrew
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew
 
I would have done the same thing. I like prime to use prime lenses, but I figure that I'm getting more value out of ones that are both faster so that I can shoot in lower light and give me the ability to shoot really shallow depth of field. With a wide lens like, a fast aperture give you the ability to shoot in lower light but wides aren't really so great for subject isolation so with all of the extra cost and weight you still don't get much of that option. Besides, if you're shooting landscape type shots the speed probably isn't going to be such a big factor as in that case you can use a tripod when the light is too low and you be stopping the lens down anyway.

I think that zooms are more useful for the ultra-wide stuff and for the really long lenses, where you'd likely be "zooming in" (where the distances might be too long to cover by "zooming with your feet). It's for this reason that I feel that more mid-length lenses are best as primes; medium wide to short tele lengths...
Thanks Aaron, you've summed it up well. The value I will get from the 7-14 will far exceed that of the PL12 prime. It's certainly been a learning experience!
 
Wide apertures lead to less sharpness in shots like those you have provided. It is not the fault of the lens...the user needs an update.
I haven't had the same problem using the PL15 at all, and the samples from the PL12 on the Lenstip website matched what I was seeing. If this lens really must be stopped down to f/5.6 or more to get sharpness across the frame, then it's not worth the money IMhO.

--
Pete
 
Last edited:
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew

--
Infinite are the arguments of mages. Truth is a jewel with many facets. Ursula K LeGuin
The way Oly labels it can be counter intuitive, but the sensor definitely doesn't tilt... It moves in two axis and also rolls/rotates. If it did tilt it'd be moving out of the focal plane (and out of alignment with the lens), thus defeating IBIS' purpose, there's no reason to do that. The five axes refer to five different types of camera shake that are counteracted, not five different types of sensor movement. This comes up ever so often:

https://www.dpreview.com/forums/post/40558437

https://www.dpreview.com/forums/thread/3824769

IBIS itself can very well cause a soft picture, but probably not unevenly across the frame like lens decentering and alignment flaws can... Unless the whole mount is off, which has been known to happen, or the whole sensor/IBIS assembly (which can also happen).
 
Last edited:
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew
 
Tommi K1 said:
Sorry, I could have done better with a 150€ body and 75€ lens. And I even mean the Olympus 9mm f/8 BCL on GM1. No need for FF at all!
[Edit]

Anyway, as you seem to keep changing the goal posts, here's something for you to aspire to. Not shot with FF either. [Edit]













[Edit]
 
Last edited by a moderator:
[..] tilting it wouldn't accomplish anything. Think about it, seriously. A lot of marketing diagrams on this are just as misleading as the name is to begin with btw.
Fascinating, so this video is wrong about tilt compensation?

 
[..] tilting it wouldn't accomplish anything. Think about it, seriously. A lot of marketing diagrams on this are just as misleading as the name is to begin with btw.
Fascinating, so this video is wrong about tilt compensation?

Fascinating, you found a poor CGI representation (it's not the first one either btw) made by a marketing department that appears to show a sensor tilting... On the other hand if you listen it says nothing about the sensor actually tilting, it says it *compensates* for the camera tilting, and that's done by moving it along the same axis as when the camera shifts (but based on different movement sensor data).

Some of ya are in such a hurry to claim that it tilts that you're not even considering what would happen if it actually did... Again, think about it. This isn't any kind of mystery, there's plenty of videos of the actual physical sensors in motion (I think Fredrick who posts here often has some on his blog, using two cameras hooked up to each other etc). I can guarantee you that you'll never find one where the sensor is tilting around.
 
Last edited:
[..] tilting it wouldn't accomplish anything. Think about it, seriously. A lot of marketing diagrams on this are just as misleading as the name is to begin with btw.
Fascinating, so this video is wrong about tilt compensation?

Here's another silly marketing video, the first half shows exactly what you wanna believe (why exactly I wouldn't know), however during the second half of it the more technical visualization of the coil motors that enable IBIS show zero tilt:


Research the subject a little, compensating for camera pitch and yaw by moving the sensor along the same axis as when compensating for other shifts (WHILE keeping the sensor aligned with the lens) makes a lot more sense than tilting it around and basically throwing everything off.


BTW, why do you think it's always alluded to as a 'sensor-shift' stabilization system? Because shifting it in two axis (and rolling it) is what's actually happening, and that's enough to correct for 5 axis of camera movent. Tilting it wouldn't compensate for anything.
 
Last edited:
Some of ya are in such a hurry to claim that it tilts that you're not even considering what would happen if it actually did... Again, think about it.
It would be same thing as tilting the rear standard on the large format camera. And how many knows what it really does?

So many has problems to understand it by even seeing when it is done.
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew
 
Your explanation is not clear to me. How can you compensate for the effect of pointing tge camera downward by moving the camera only in the x-y plane?
By shifting the sensor. This is reason why the 4/3 standard creator Olympus defined the image circle diameter, not the sensor size! So the sensor can be shifted around so much that the framing can be kept stationary while the handshake the people get (Olympus really has lots of research and studies done for everything from hand shakes, hand sizes, wanted DOF, sharpness, colors all of it!) that it is compensated without vignetting the sensor by clipping image circle by shifting sensor.
You can compensate for changes in the framing but the sensor and therefore the focal plane was tilted by your movement.
The IBIS corrects only very small motions, not huge ones that would be required so the focal plane (DOF) would change so much that the subject would become out of focus.

If you use center focus point and you do "focus + recompose" method you can easily throw subject out of focus like in portraiture move the depth of field so much that instead eyes being in focus now their ears are in focus. But that requires far more camera tilting than IBIS can compensate in first place.

The IBIS is not like a rear standard in large format camera. You are not going to do a 30% shifts movements for composition or framing reasons. It is there just to minimize the camera shake and to do so it is controlled by the algorithms that are patented by Olympus that what shifting controls are done for the sensor when specific kind 5-axis motions are on camera body.

Notice that Olympus can't do 6-axis correction, as the 6th axis is the Z-axis movement (backward and forward) and that would allow to change the focus by doing very careful sensor movement forward or backward, and that would be possible if the sensor would be tilted.

The sensor is floating in magnetic field, locked in specific flange distance but still freely floating when not suspended in the field. It can as well be locked to specific position with magnets so it doesn't move at all in normal operation if you don't want to use IBIS.

So if Olympus would make a 6-axis IBIS, it would change again the market a lot as you could now do a movements for macro photography on sensor instead lens AF.

Pentax has the algorithms for the sensor shifting so that you can do small framing corrections by moving sensor up/down or left/right without starting to move the camera itself or needing a macro slide.
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew
 
<snip>
You mean the stabilization would cause it? The sensor isn't tilted, only shifted.
So what do you understand by 5-axis IBIS?
That what the Olympus material says and what the videos of the IBIS operational etc says.

It doesn't tilt, only shift in X and Y axis and roll at Z axis. Using correct mathematical formula you can compensate all 5 axis movements.

This is why Olympus can't make a sensor tilt-feature that would make every lens as tilt-lens as sensor doesn't tilt.

But use your eyes, look those samples, and if you know what a decentered lens is, that is exactly how it looks like!
I agree it's a decentred lens with field curvature.

I'm afraid you have described three axis motion. A plane has six degrees of freedom about its centre. Three are linear translations and three are rotations. All motions about the centre are a combination of these.

Andrew
 
I need to think about this and draw out some geometric diagrams. Since I have a T-S lens, I can check that the effects priduced are consistent with the analysis.
This may be of interest: https://en.wikipedia.org/wiki/Scheimpflug_principle
Daniel

Thank you for the thought. I'm familiar with that. You are correct that theory to design experiment and predict observations comes first followed by experiment to test against predictions.

Andrew
 

Keyboard shortcuts

Back
Top