Obsessing

[AdamJ]

I'm worried that lens designers are being driven by lens reviewers. Lens reviewers seem to be obsessed with things like corner sharpness. I don't care much about corner sharpness because the subject of my image is invariably in the centre. I'd rather pay $500 for a lens with good centre sharpness and average corner sharpness, as opposed to $2,000 for a lens with good centre sharpness and good corner sharpness.

A good photograph rarely depends on lens sharpness.

 

[LuCoOc]

A good photograph rarely depends on lens sharpness.

A good photograph rarely depends on lens sharpness as long as you don't need large prints. And yes - corner sharpness isn't neccessary for every image... it depends on the photographers style and what s/he is taking a picture of. I really respect comments like yours but I don't like them to be generalized.

 

[NormanBates]

I agree that you don't always need corner sharpness, but if you need it and it's not there you'll just miss the shot

I'd like to see more bokeh tests
http://www.rickdenney.com/bokeh_test.htm

 

[Loswr]

A good photograph rarely depends on lens sharpness.

For you. For a photographer who shoots a lot of architecture, for example, the opposite is true - buildings are hard, and soft corners just don't cut it.

I'd like to see more bokeh tests

Check out the photozone.de reviews. Most of the Canon lens reviews have a section on bokeh with a standardized target setup (particularly the FF reviews, only a few of the APS-C tests show bokeh results).

 

[awinphoto]

It really depends on what you are doing... Architecture, like someone said usually want 100% focus with the exception of detail shots. Portraiture, yes tends to be center of the frame for most family and amateur shots, however once you move the subject to the thirds, or even if you're being creative and shooting the subject AT the corner of the frame for affect/white space/drama/whatever then at that time then It becomes a very big problem. Lastly, if you shoot product shots, automotive, aviation, commercial, etc... you get what you pay for.

 

[Flake]

It's worse with a high MP count FF camera too. Take a lens like the 17 - 40mm f/4 L which wide open at 17mm has no measureable resolution at the corners, the details of bricks simply dissappears while the centre stays acceptable, when the images are downsized for web use or print the problem doesn't get any better. On low MP count FF cameras its less of a problem because it's not as noticable.

Corner & border resolution are more of an issue for wide angle & ultra wide angle lenses, applications such as architecture, landscape, & street scenes all need good full scene resolution, Bokeh on the other hand is an unknown word !

 

[branden]

I'm worried that lens designers are being driven by lens reviewers.

Where did you get this idea? Lens designers are driven by a mix of sales and technology.

 

[AdamJ]

I've been unwell for a while so haven't been able to reply until now.

I was indeed generalising and I'm sorry about that. I absolutely accept that architecture, landscape and street scenes are subjects for which corner sharpness is sometimes a practical advantage.

Corner sharpness was only an example to illustrate my point but I'll stick with it. My concern is this. Lens test websites often give the impression that a lens is not worth considering if, for example, it doesn't deliver good corner sharpness. Flake mentioned the 17-40mm and said it has 'no measurable resolution at the corners.' I'm not even sure what 'no measurable resolution' means but anyhow, this is a good example of a lens getting a bad reputation. I have a 17-40mm and I like it, not least because it was affordable and is well built (and it definitely does have some resolution at the corners!). A sharply resolved brick in the corner has not once been an important feature in any photograph I've ever taken. I'm not saying a sharp brick in the corner is never important to anyone, just that corner softness is one example of an issue that is conferred such significance in lens reviews that some potential buyers might be put off the lens altogether.

The 17-40mm may one day get replaced and when it does, I wouldn't mind betting that the replacement will have much better corner sharpness and a 50% higher street price. Look at the the price-hike on the 70-200mm f2.8L IS II. The newer one is measurably better than the older one - photographing a brick wall to ensure a flat plane of focus proves it. The thing is, how often does even an architectural photographer take an exactly perpendicular photo of a brick wall?

I just feel that this pursuit of optical perfection is not so much about any practical or creative benefit but rather about profiting from those who are increasingly being led to believe that only the best is good enough. If it continues to push prices higher and higher, I feel sure it will dissuade many in the future from taking up SLR photography as a hobby.

 

[awinphoto]

I just feel that this pursuit of optical perfection is not so much about any practical or creative benefit but rather about profiting from those who are increasingly being led to believe that only the best is good enough. If it continues to push prices higher and higher, I feel sure it will dissuade many in the future from taking up SLR photography as a hobby.

I wouldn't mind some amateurs in the future to be dissuaded from SLR photography, They have to be the 100% competitor for most modern day pro photographers haha. With that said, I have the 17-40 and use that for my set up and shoot architecture... Then again on my 7D, to get the most sharpness and detail I usually shoot at F8-11 (lower the aperture (16), the worse the degree of the lens softening and the higher (4), the shallower the DOF) and really any architectural photographer would know better than shoot at that if they want corner sharpness. At stopped down apertures corner sharpness is ok... Plus dont forget a sturdy tripod.

Lastly, regarding sharpness... should the new 17-40 be announced and is not going to cost me a second mortgage, or third, I would be in line to trade mine and pick up a new one. Sharpness is one of those things where if you dont want all the sharpness, then it's ok, you dont have to have it, however it's always better to have the extra sharpness and soften in post as needed then not to have it and need it. Even the 135mm soft focus has the option to have the soft focus on or off... If i want an artsy fartsy effect I would rather have full control in post rather than have no control what-so-ever.

 

[Flake]

Corner & border sharpness is more important on some lenses than others. Wide angle lenses really do need it, because in the main they're used for architecture, landscape, or internal shots. Standard zooms pretty much fall into this catagory too, because of the wide end probably right up to about 70mm. After that it becomes less of an issue as the subject tends to be in the centre of the frame and depth of field on longer lenses lessens.

On EF lenses the FF performance is more important, with current sensors at 21MP and set to go higher the issue is the very noticeable difference there is between the centre resolution / sharpness and the gradual fall off from the centre to the corner. On a 12MP body this effect isn't quite so noticeable, but imagine the effect on a 40MP body?

I don't think anyone on these forums envisages MP count to reduce or stay static, so lens designers need to think of their products and how they will perform with future bodies not yet even envisaged, and that is the driver. Imagine a body released with really showed the limitations of a range of lenses, and the clamour for redesigns? To some extent this happened with the 5D MkII and it's nice to know Canons new lenses will perform well on at lest the next generation of FF bodies.

 

[motorhead]

The statement "no measurable resolution" in a lens test to me means "its the worst lens we have ever seen but the company would sue us if we said that".

I understand why some might choose to ignore the more scientific gobbledegook of lab tests, but I take them very seriously. Not for me the fluffy "review" where its always described in glowing terms, I'd rather know upfront exactly what I'm getting, the good and the bad with no pulling of punches.

 

[Admin US West]

Both measured tests, and actual photography tests provide valuable information. Resolution can be measured, but Lo CA is not measurable, neither is Bokeh
As for color, digital cameras see black and white, and their software guesses at colors. The colors can be set to anything I like. Microcontrast is another made-up quality. MTF is a measure for contrast (sharpness).

A reasonable lens test is one that can be repeated by other photographers or testers. When someone takes a photo of a group of trees, for example, we usually do not know where the camera actually focused, and lens flaws are hid in the detail. The same for distortions. You will not be sure if the out of focus or blurry edges are due to depth of field, curvature of field, or other lens flaws, so it provides information, but only that its possible to get either sharp edges or poor ones, depending on the image.

 

[Flake]

Boka ?????

 

[Admin US West]

Boka ?????

Typo

 

[Edwin Herdman]

Resolution can be measured, but Lo CA is not measurable, neither is Bokeh

Actually, this raises an interesting point. Most of these features of lenses can be described in a completely mathematical way, and thus can be rated on a scale. Perhaps these is simply an unmet demand for more types of metrics for lens reviews. Personally, knowing a lens's CA characteristics (of the different types) is very useful to me.

I think it can be measured - lateral CA anyway. lateral CA can be measured in terms of percentage of a frame width of fringing, though people usually talk about how many pixels out from a like of marked contrast that fringes exist. It can be fixed in a standard, programmatic way (not longitudinal CAs so much though). For axial (longitudinal) chromatic aberrations I figure that the measure would be how far out of sync the two color channels are, comparing the sharpness of images at different wavelengths. It would certainly be harder than the traditional MTF test but seems possible regardless. Boke would be even trickier since there is more of a subjective quality, and trying to test for adherence to a certain desired kind of defocus quality might be a point of contention, with some people liking their boke looking different.

Still, review sites like photozone.de, despite not standardizing a test to translate boke into a graph still feature it, and it is clear enough that there is a comparison to be made in this quality across different lenses, and more credit given to lenses with better defocus area quality.

Photozone does feature a graph of purple-green fringing lateral CAs, though sometimes I differ with their results. They claim that purple-green fringing is "not field relevant" on the TS-E 90mm, but on my first day of testing I discovered quite a bit of it, in various circumstances, that is visible even from the confines of a fit-to-screen monitor view (about 1920x1200). Either my (secondhand but essentially new) copy is somehow defective from the formula (it's not a mechanical issue) or their review methodology is flawed.