do image stabilisers decrease image quality?

[neuroanatomist]

A single lens that produces an image with little or no flaw is possible. The human eye has only one lens. Glass and Crystal lenses abilities are flawed in comparison.

It has one structure called a lens, true, but actually the eye can be considered to have at least 4 elements, perhaps better thought of as 4 groups, 3 groups with one element each and one group (the lens itself) comprising multiple elements. The cornea actually does most of the refracting, but both the aqueous humour and vitreous humour have refractive indices that differ from the cornea and lens. The lens itself has a refractive index that varies through the structure, and the action of the ciliary muscles changes the shape of the lens, which alters not only the shape of the lens surfaces, but also the differential thickness of the regions of different internal refractive indices.

So if we assume the eye is producing an image with little or no flaws, understand that it's far more complex than a single-element lens.

 

[PackLight]

A single lens that produces an image with little or no flaw is possible. The human eye has only one lens. Glass and Crystal lenses abilities are flawed in comparison.

It has one structure called a lens, true, but actually the eye can be considered to have at least 4 elements, perhaps better thought of as 4 groups, 3 groups with one element each and one group (the lens itself) comprising multiple elements. The cornea actually does most of the refracting, but both the aqueous humour and vitreous humour have refractive indices that differ from the cornea and lens. The lens itself has a refractive index that varies through the structure, and the action of the ciliary muscles changes the shape of the lens, which alters not only the shape of the lens surfaces, but also the differential thickness of the regions of different internal refractive indices.

So if we assume the eye is producing an image with little or no flaws, understand that it's far more complex than a single-element lens.

I would say it is more complex. Compared to Nikon its sensor has about 3x the dynamic range. I wonder how that would rate with DxO.

But I do have to admit it does have flaws, when I turned 50 I had to start using reading glasses. So I guess mine has a 5th element that I use for macro work.

 

[mb66energy]

A single lens that produces an image with little or no flaw is possible. The human eye has only one lens. Glass and Crystal lenses abilities are flawed in comparison.

It has one structure called a lens, true, but actually the eye can be considered to have at least 4 elements, perhaps better thought of as 4 groups, 3 groups with one element each and one group (the lens itself) comprising multiple elements. The cornea actually does most of the refracting, but both the aqueous humour and vitreous humour have refractive indices that differ from the cornea and lens. The lens itself has a refractive index that varies through the structure, and the action of the ciliary muscles changes the shape of the lens, which alters not only the shape of the lens surfaces, but also the differential thickness of the regions of different internal refractive indices.

So if we assume the eye is producing an image with little or no flaws, understand that it's far more complex than a single-element lens.

That's what I thought as I read the comment, that the human eye is a single lens construction. I would prefer to see it as a 4 element - 1 group system but that might be a matter of taste.

There are two further reasons why the eye doesn't compare to photographic lenses:

• The eye has optimum quality only near the image center where the spatial resolution is maximal and spherical aberration or CAs arent a problem. The image borders are optimized for faster "frame rates" and only B&W ...
• The brain corrects the raw data from the retina and adds fragments of detailed information from the image center (while scanning our surroundings) to give us the perception of a detailed larger picture with corrected perspective.
• EDIT: Image stabilized by coupling the eye to the vestibular system to keep the direction during slower motions + soft tissue between eyeball and bones to damp higher frequency motion.

 

[Mr Bean]

Regarding the title of the post:
I think people are getting a little caught up on the number of lens elements in the optical path, which is really a separate issue to how IS works (in Canon lenses at least). Yes, IQ is affected by the number of lenses in the path, but bear in mind that multiple lenses are used to correct multiple optical issues. If you didn't have those lenses, then spherical or CA (or other issues) would degrade the quality. Its a balance.

IS on the other hand, is done by moving an element or elements on and off axis. This is going to create aberrations such as astigmatism. The effect would be much greater in lenses that are fast, as the latitude for astigmatism would be much smaller.

As an example, for those who wear glasses, hold them an inch or 2 in front of your eyes and move them around. As you will see, the image moves accordingly. This is, in effect, the same approach as IS in a Canon lens. Pushing it further, while holding the glasses in front, move the glasses so you are looking through the outer edge, not the center. In some cases, the IQ will drop, and the image blurs. This is the extreme example of an off axis lens.

 

[Mt Spokane Photography]

Yes, something is lost, and it is light.
But ... does IQ get worse with each additional element, as he said? If so, why not a 1 or a 2 element lens? IQ should be much better than those 22 element lenses - less CA, sharper, less distortion, etc - Really ?? Where are those one or two element lenses with superior IQ, or for that matter, 5 element lenses?
Its a nice theory, but the best lenses with the best IQ have many elements, and they do lose light, but not all that much. The 9 element 85mm f/1.8 has a Tstop of 2, so it loses 0.2 stops in the glass. The 20 element 70-200mm f/4L IS has a Tstop of 4.6, so it loses 0.6 stops. 20 elements, and it has suburb IQ and sharpness!
So, how many elements until you lose a stop? Maybe 30 some elements? Your worry that light won't make it thru a lens due to the number of elements seems a bit far fetched.

It is more than just light lost, it is the quality of the light itself, how the light breaks down going through the lens.

It does get worse with each element, because each progressive element corrects one problem and adds another, or not depending. If the lens has 20 elements the last element corrects flaws made through one or more of the previous 19. It will never be perfect and each element takes away, even if it is slightly. There is more lost than just light and loosing a stop, that is why we look for other flaws in our IQ other than just sharpness or quantity of light. No doubt the L lenses deliver great IQ with multiple elements, but it doesn't change that fact that a bit of quality is lost with each element.

A single lens that produces an image with little or no flaw is possible. The human eye has only one lens. Glass and Crystal lenses abilities are flawed in comparison.

I think your theory says it all. A 20 element lens has lower IQ than a 10 or 7 or 3 or one element lens.
The only problem cones in the FACT that actual measurements disprove it.

 

[PackLight]

Yes, something is lost, and it is light.
But ... does IQ get worse with each additional element, as he said? If so, why not a 1 or a 2 element lens? IQ should be much better than those 22 element lenses - less CA, sharper, less distortion, etc - Really ?? Where are those one or two element lenses with superior IQ, or for that matter, 5 element lenses?
Its a nice theory, but the best lenses with the best IQ have many elements, and they do lose light, but not all that much. The 9 element 85mm f/1.8 has a Tstop of 2, so it loses 0.2 stops in the glass. The 20 element 70-200mm f/4L IS has a Tstop of 4.6, so it loses 0.6 stops. 20 elements, and it has suburb IQ and sharpness!
So, how many elements until you lose a stop? Maybe 30 some elements? Your worry that light won't make it thru a lens due to the number of elements seems a bit far fetched.

It is more than just light lost, it is the quality of the light itself, how the light breaks down going through the lens.

It does get worse with each element, because each progressive element corrects one problem and adds another, or not depending. If the lens has 20 elements the last element corrects flaws made through one or more of the previous 19. It will never be perfect and each element takes away, even if it is slightly. There is more lost than just light and loosing a stop, that is why we look for other flaws in our IQ other than just sharpness or quantity of light. No doubt the L lenses deliver great IQ with multiple elements, but it doesn't change that fact that a bit of quality is lost with each element.

A single lens that produces an image with little or no flaw is possible. The human eye has only one lens. Glass and Crystal lenses abilities are flawed in comparison.

I think your theory says it all. A 20 element lens has lower IQ than a 10 or 7 or 3 or one element lens.
The only problem cones in the FACT that actual measurements disprove it.

Theory? I didn't say that, but here since you want to add that Theory I will add it.

How does your theory test out and the FACT of actual measurments stack up. 70-200mm with 20 elements. 200mm f/2.8 with 17 and 300mm f/2.8 with 16. But wait, the 300mm is Canon's sharpest lens and it has 4 less elements.

By your reasoing shouldn't the 70-200mm be better, it has more elements?

 

[neuroanatomist]

Theory? I didn't say that, but here since you want to add that Theory I will add it.

How does your theory test out and the FACT of actual measurments stack up. 70-200mm with 20 elements. 200mm f/2.8 with 17 and 300mm f/2.8 with 16. But wait, the 300mm is Canon's sharpest lens and it has 4 less elements.

By your reasoing shouldn't the 70-200mm be better, it has more elements?

Primes vs. a zoom - yep, that's a logical comparison. So PackLight's Postulate of Fewer Elements says the 35mm f/2, with only 7 elements, should have far better IQ than any of the above lenses. Does it?

 

[DBCdp]

This is a very good question. It would seem that it's a matter of perspective. If one handholds the camera then IS helps make the obvious difference in your own body movement. If you handhold a lens without it, then the 1:1 shutter speed/focal length rules really come into application and if you don't observe this religiously your hit rate will suffer accordingly.

My 70-200 2.8 IS was a phenomenally sharp lens 5 1/2 years ago. Now I notice it's flaws. I think the variables within ourselves far outweigh the IQ differences in IS/non IS lenses and render it a mute point.

I also remember that little test with all the filters stacked on a lens. Given, light passed through. But the IQ was horribly mutilated as I recall.

I used to try and use my old 100mm 2.8 macro for portraits and was almost always disappointed. My own fault for not obeying the 1:1 rule. My new 100Macro with IS is a far better tool for this, as it helps me where I need the help. So, like all questions concerning lens/camera combinations...what are you using it for? What is your style? The person operating it is the weak link in most cases and the knowledge/abilities we have are almost always the limiting factor. Can you shoot a football game with a 300 2.8? Of course. Will you get more keepers with a 300 2.8 IS? Definately! Will either one of these great lenses save you from certain disaster when the play comes your way and 5 250lb+ guys land on top of you? Sorry, not gonna happen!

So the bottom line is that there are always trade-offs and it's up to us to know what applies and when then put the best key elements into play that are available to us to make our vision a reality. So, what am I missing?

 

[DBCdp]

As an aside, let me ask you to point a laser at an object while standing, say...a doorknob across the room. Can you keep the dot exactly in the center? Sitting? Laying down? Now extend the distance to what you'd shoot with a 300mm or 400mm lens. How do you like the way that dot moves now?

So basically, if we're touching the camera, we're disturbing the shot. Learn what foods and drinks skew your aim and how to control that bullseye accuracy necessary and watch your keeper rate go up! Funny how we spend thousands on the newest high tech equipment to help us in our quest, even while many of us let our own bodies go to pot (in varying degrees, intentional or not, as aging is unavoidable).

Good luck!

 

[dafrank]

No lens elements, while leaving the light unchanged and, therefore, "perfect" obviously can't be called a lens. Yes, if one were able to correct all abberations and still define a desired focal length with only one physical lens element, it would necessarily be better than one which corrected the light equally with two, if only because more light would be able to be transmitted through an equally transmissive single element than through two. In this same manner, because, practically, if not theoretically, IS lenses seem to always use more lens elements for the same generation of focal length and aperture lens made by the same companies, usually the non IS versions are and should be better, other factors not withstanding. I'm not technically savy enough to explain this with certainty, but it does seem so in reality. Furthermore, one must take into consideration how IS works. Some element (usually a "group" of lens elements) must move within the lens to compensate for the optically equal and opposite movement of the camera/lens combination in order to cancel out that movement, thereby minimizing the blurring effects of said movement during exposure. It will always be harder, with a correspondingly greater degree of potential errors, to keep such a lens configuration with laterally moveable elements true to its ideal optical path as well as to make aberration corrections with the least number of lens elements and most effective design of those elements possible, as the elements that need to be moved must be designed and grouped for IS functioning, rather than for their ideal placement within the lightpath for abberration correction only. In other words, in IS lenses, compromises with pure optical performance must be made, even though they may sometimes be very small and sometimes even irrelevant when, for instance, an IS lens is able to use a generation newer and better lens element design and fabrication process over a non IS lens using older lens design technology. Well, that's my take.

Regards,
David

 

[PackLight]

Theory? I didn't say that, but here since you want to add that Theory I will add it.

How does your theory test out and the FACT of actual measurments stack up. 70-200mm with 20 elements. 200mm f/2.8 with 17 and 300mm f/2.8 with 16. But wait, the 300mm is Canon's sharpest lens and it has 4 less elements.

By your reasoing shouldn't the 70-200mm be better, it has more elements?

Primes vs. a zoom - yep, that's a logical comparison. So PackLight's Postulate of Fewer Elements says the 35mm f/2, with only 7 elements, should have far better IQ than any of the above lenses. Does it?

It is as logical as the argument that by adding additional elements in of itself improves image quality. My belief would be the quantity of elements does not matter, what matters is having the appropriate number of elements to get the desired IQ. So the comparison was an unequal argument to a position with little merit.


But the point I was making to start with is that the majority of the elements are providing "correction", the elements are not improving the image quality that the Raw light would have already held in itself. It is modifying and manipulating it, the best lenses would of course loose the least amount of available IQ in the process. Obviously with the technology available today it takes a certain amount of elements, but that does not mean it will be so in the future.

 

[risc32]

so the lens elements, most of them you say, are "correcting" the image not improving it... hmmm. where i come from correcting problems IS improving the image. Maybe you are trying to say that all those silly elements are fouling things up, and that leads to more elements to fix these new problems, and so on, and so on. hmmm again, but i don't think so. i guess a billion dollar company with the better part of 100yrs making lenses can learn some stuff from us here.
note to Canon HQ- use much fewer lens elements from now on cause companies like leica make do with only a handful.
I've watched a few Canon lens production videos on youtube, namely the 500mmf4, and to me at least it looks like they have things well under control.
as am typing and rereading your last post i've decided i don't even know what you're trying to say.

 

[weixing]

Hi,
WOW! Still discussing on this... we must be very bore while waiting for new Rumors... ha ha ha ;D I'm also very bore while waiting for "news" on 70D or/and 7D2...

Anyway, the OP questions about IS consist of 2 parts:
1) Does the IS action affect IQ?
I agree with what Mr Bean mention below.

My feeling is that IS would affect IQ as the lens or lenses in the IS unit are moved off center, to negate movement of the photographer. This process of moving the lenses off axis in the optical path would have been "allowed" in the design, but it'll be a case of "....the lesser of 2 evils".

2) Does the extra elements in IS affect IQ?
This questions also consist of 2 parts:
a) Does IS increase the elements in a lens?
This is a very difficult question to answer since I'm not an optics designer, but base on the trend in the lens specification:
Lens with IS had more elements than similar lens without IS. For example:
- EF 70-200mm f/2.8L IS USM has 23 elements, but EF 70-200mm f/2.8L USM only had 18 elements.
- EF 70-200mm f/4L IS USM has 20 elements, but EF 70-200mm f/4L USM only had 16 elements.

b) If IS does require more element, does the extra element affect IQ?
From the other thread, "Can a UV filter affect IQ (sharpness) on a lens?", most of poster agree that a filter does decrease IQ to a certain degree, but the decrease in IQ is negligible if the filter is a very good one. So if adding a filter does decrease the IQ, so does every element in the lens since a filter is also consider an optical element (even air in lens is also consider as an optical element).

By the way, IMHO, using modern coating and manufacturing technology, the decrease in IQ cause by a few extra elements is very minimum, but you gain 3 to 4 stop of advantage with IS, so it's great to have IS as long as it's does not increase the price tag by a lot... ha ha ha ;D

Have a nice day.

PS: Where my 70D or/and 7D2 rumors!!

 

[PackLight]

The 500mm video is really cool, if you are talking about the one at the assembly plant. The new 500mm II has 1 less element than the old 500mm. I guess by the more elements is better reasoning the old 500mm has better IQ.

Improving the image from what point though? Before it entered the first lens it didn't in improvement, it only need to be focused. Correcting other flaws we see in our images takes place at or after the first element.

I doubt Canon R&D will learn anything from reading this thread, since they already understand the principles they are working with. Canon marketing may take a lesson and start putting additional elements in the lenses so that the element counters among us can feel good.

 

[marekjoz]

Happily nad hopefuly liquid lenses would not require IS at all And guess who's got the patent for it.

 

[SwissBear]

I observe (sometimes) with my 24-105 that if the IS kicks in heavily, the bokeh SUFFERS.
The focussed parts are still quite sharp, but the bokeh gets wild and rough.

This makes actually some sense, as the theory says that "good" bokeh comes from well aligned elements, we all agree that most primes have a "better" bokeh than zooms.

 

[eddiemrg]

In my town this problem is called:"SEGA MENTALE".

 

[Ewinter]

I'll just throw this out there for those who think that more optical elements affect IQ and stop the light being "pure".
That may well be true, but sometimes it's necessary.

After all, try ripping out your cornea and see how "pure and sharp" the image looks then.

As for the OP's question- depends on the situation. If it allows the image to be sharp handheld, then no, it helps the iq. If it messes up an action pan or causes vibration on a tripod then yes- just turn it off.

 

[PackLight]

After all, try ripping out your cornea and see how "pure and sharp" the image looks then.

"How far can an SLR lens actually succeed in duplicating the characteristics of the human eye? The natural colors of the subject seen by a clear eye, precise expression that can focus directly on what it wants to look at, a swift angle of vision that losses no time in catching even the fastest moving objects. Canon's pursuit of the techniques and technologies that will allow photographic lenses to approach the purity, expression, and dynamism of the human eye will never be tarnished by compromise."

"lens elements invariably have properties and imperfections which prevent them from accurately converging light rays into a single point and which tend to disperse light near the edges. These properties which prevent a group of light rays from a single subject point from re converging at the ideal image point or cause dispersion when light rays pass through the lens, are called aberrations."

I wonder where those quotes came from....

 

[tron]

Yes they do decrease IQ. If you don't believe me start shooting with yours turned off when you are doing hand held shooting and see how much your pictures improve.

;D