David Brooks' comments to a reader in the latest SB echoes your feelings, Bogie or should I say, David?
BTW,
"The world thrives on BS, just don't believe your own."
quoted by me so many times, I thnk I own the phrase.
About lenses or anything else.

You did not say whether you have ever actually looked at what an unsharpened raw dSLR image looks like. It's not a matter of BS if you can actually test whether lenses or camera sharpening makes the difference.

Bogie,

It's hard to imagine anybody NOT looking at an unsharpened image, since that's how you always start, at least when you don't have in-camera sharpening. I never apply sharpening in-camera. Lens sharpness becomes more important as you print larger sizes. A lack of lens sharpness may be unnoticeable in small prints, but becomes very evident in larger ones and post-processing sharpening can only do so much.

We deal with another kind of unsharpness all the time, caused by parts of the subject being out of the depth of view range. That lack of sharpness is very noticeable and is probably worse than the underlying lack of sharpness of most lenses, but it cannot magically be eliminated by post-processing sharpening.

So no, I cannot agree with you that the issue of lens sharpness for dSLR's is hype.

It is not how most people start if they just choose Raw output. There are a lot of different menu choice configurations available with different cameras, and I would assume most photographers do not deliberately choose a picture quality other than the standard default for Raw, which does include sharpening. With some cameras to get no sharpening there is a sharpening slider which you would have to deliberately adjust to full negative or zero to eliminate the application of post exposure sharpening, and with some you would have to choose an output format that is specifically not enhanced by any post exposure processing. In other words, even saving in Raw format, all dSLR's apply sharpening after exposure unless specifically and deliberately set to not apply any sharpening.

What I referred to as hype is that with 6,8, and even 10 MPX dSLR's the nature and size of sensor pixel sites is such that no appreciable advantage in image sharpness is going to be gained between using a standard camera brand lens including what is referred to as a kit lens and a very much more expensive pro-quality lens. And that is due to the fact digital capture by a CCD/CMOS sensor is inherently soft until it is sharpened post exposure. In other words the sensor chip cannot actually record the subtle differences in sharpness that would show up on film between good quality consumer optics and high-priced pro quality optics.

Bogie,

Most photographers beyond the bare beginners level should be aware of default settings and modify those to their own preferences. It is not hard to do at all and understanding it is a first requirement to take control of the total image creation process.

Differences in lens sharpness used in 6 to 10MP dSLR's will show up in larger prints whether you like it or not. It is an issue of what you want to accomplish at what quality level and at what price-point, not hype.

Bogie, I have been misconstrued, the BS may be from the manufacturers. Every one has some manure to spread. Eg., in another post here in this forum, one poster was told by a salesman that digital cameras cause vignetting and that using a filter on the lens will solve the problem. Utter BS, if anything, the filter itself might cause vignetting.
Consider also the mega pixel wars, the more the merrrier, supposedly. My understanding, and if other members will correct me, is that digital resolution does not equate to film resolution when it comes to recording fine detail, but the manus are falling all over each other to produce cameras with more and more sensors on the same sized sensor.
At the same time, we read in the photographic press that the larger the individual pixel, the better it is able to resolve fine detail.
In a recent article in SB, it is stated that the pixel wars are a moot issue with the introduction of the new GF program.
I won't get involved here with full frame 35mm sized sensors or full 645 sensors for MF cameras.
What is the consumer to believe?

Obviously you have not made the simple test I described in my original post, or you won't admit it if you have since.

Believe what you like, only you fortunately have to live with your beliefs.

Bogie,

Don't make assumptions about what I have and haven't done. I've done this test when I first got my Nikon D70a and it proves that sharpening in-camera makes images sharper. It doesn't demonstrate the difference is sharpness for different lenses and how that translates into the quality of the final image.

Yes BS comes from all sides. That's why I described the simple test in my opening post. Anyone who has a prosumer or dSLR camera that offers the Raw save option and also has menu controls to adjust quality factors like saturation, contrast and particularly sharpness, can turn off or minimize post exposure sharpening by the camera. Then by opening a raw unsharpened exposure (of detail on a tripod), and the same shot made with normal sharpening on-screen with your computer, and zoom in to 400% so you can begin to see the individual pixels, you'll have visual evidence of 1. how soft a raw digital capture is, and how much the sharpness of a normal default capture depends on the camera's post exposure sharpening. It is much the same thing as in the old days of Pop Photography when they would use extreme blow-ups of a small section of 35mm film to show comparisons of lens sharpness on film.

"At the same time, we read in the photographic press that the larger the individual pixel, the better it is able to resolve fine detail."

I have not read that larger sensor pixel cells are better able to resolve fine detail myself. Its a pretty stupid lie simply because anyone who understands digital capture should realize all that each individual pixel records is how much light is striking it. An individual sensor pixel site cannot record image detail, just an exposure value, that is then reproduced as a pixel on screen or in a print with a single value. Detail can only be recorded by a number of pixels each recording different exposure values, so the more pixels the more and finer detail that can be recorded digitally.

But if as you suggest cram more and more pixels sites into the same physical dSLR sensor chip, the individual sensor cells have to be smaller of course. The main disadvantage to that is that as the pixel cell size is reduced, it gathers less light to produce a lower electrical signal that records exposure. That smaller individual pixel sensor cell size can contribute to noise in shadows or at high ISO settings.

Software programs that interpolate image files to produce big enlargements do a very good job. However, they cannot change the fact they begin with just so much image information, and software cannot add anything to that information, as if you had actually taken the photo with a higher megapixel camera - it is still just a blow-up of a small original. An interpolated 16x20 print made from a 6 megapixel digital camera even if it looks good will still only have half the image information compared to a 16x20 made from a 12 megapixel camera.

I don't like being critical here, and I am all for doing one's own experiments. I do it myself fairly often: for example, is there a noticeable difference in resolution between shooting raw or at the highest setting jpg? (Answer: yes). However, the initial post advocates a flawed experimental design. True, it calls for only one treatment at a time, but it is not the treatment that would answer the question. If you want to know if "better" (e.g., more expensive, better reviewed, more resolution - however it's defined) lenses generate more effective detail, you need to compare the lenses, not other treatments with the same lens.

Based not on my experiments with this question but published reviews, it appears that kit lenses are pretty close optically to the more expensive variants, just slower and flimsier. Whether digitally specific lenses outperform legacy lenses is an interesting question, however.

>>If you want to know if "better" (e.g., more expensive, better reviewed, more resolution - however it's defined) lenses generate more effective detail, you need to compare the lenses, not other treatments with the same lens.<<

I've done that between a kit lens and a near $1,000 pro zoom lens covering just about the same focal lengths on a 6mpx dSLR. At maximum blow-up the prints were not noticeably different in detail and sharpness.

The "experiment" I described is not to compare lenses, but to realize visually on screen what an unsharpened raw digital sensor capture looks like zoomed in so you can see the individual pixels, compared to another shot that has received the standard default camera post-exposure sharpening processing in Raw format.

My contention is based on what is actually recorded by a dSLR sensor, minor differences in lens sharpness will not make a significant difference in print results as far as detail and sharpness are concerned. If you want sharper images than what your getting, try kicking up the sharpening level a notch in the camera's menu selections.

>>Whether digitally specific lenses outperform legacy lenses is an interesting question, however.<<

Digitally specific in a dSLR lens brand/model usually only means the focal length range of a zoom is configured to accommodate an APS-C or thereabouts area of image field. In other dimensions of optical design and performance lenses for film and digital don't perform differently to any advantage.

I have to very much disagree with you on your last sentence. There is a huge difference between 35mm and APS-C lenses besides their focal length. Digital-only lenses are on average smaller, lighter and lower cost and the good ones have additional anti-reflective coatings to reduce flare from the light reflected off of the very reflective sensor.

Of course a lens with a shorter range of focal lengths would be physically smaller and lighter, but there is nothing in the specs that indicates the optical properties are significantly different from a lens designed for film. As to light reflecting from the surface of a sensor it is not significantly more reflective than many films. As to lower cost, have you checked out the very wide angle "digital" zooms made by Nikon and Canon, not cheap at all.

I have done some testing for the purposes of a camera club talk, to establish whether or not RAW capture was of benefit to the A3 prints we regularly produce for competitions, and indeed to compare different lenses. At 6.1MP, I can tell you that the SMC Pentax-DA 16-45mm f4 has a significant improvement in sharpness over the kit lens SMC Pentax 18-55mm f3.5-5.6. Both will produce excellent A3 prints which to the untrained eye will appear equally good, and indeed at the proper viewing distance will be equally good, but on critical examination there is an advantage to the more expensive lens. Now that we are moving to 10.2MP no doubt this will become a more marked difference.
AS regards RAW capture, this has been debated on other threads, but I concluded that it could be an advantage in some respects, but not necessarily so for all images. For example, the texture of a top quality JPEG can actually look better than a RAW capture.
I don't think it is hype, I think there are real differences in lenses and even at 6.1MP the difference can be seen to the critical eye.

I also believe in having the best lenses you can afford.

As for JPEG vs. RAW, the JPEG should look better than the RAW file in that the JPEG is post processed in camera and the RAW file requires post processing afterwards in your editing program. So there is more potential for a better quality image shooting RAW. Depending on the subject and whether you use proper in camera settings you'll do just fine shooting the highest quality JPEG. I do for over 95% of my work.

>>I don't think it is hype, I think there are real differences in lenses and even at 6.1MP the difference can be seen to the critical eye. <<

That is an unwitting or self-inflicted HOAX. The reason is evaluating the product of a dSLR camera for image sharpness what you are seeing is the result of an application of a sharpening algorithm. Even in Raw format camera post exposure sharpening is applied, so how do you separate lens performance from what a Sharpening Algorithm does? Software sharpening whether in a camera, at a particular setting, or used in a computer will produce varying results depending on differences in image subject content.

Unless you bench test the lenses optically, or use a very fine-grain film to expose under controlled conditions, you'll not get an objective result as to what the sharpness of a particular lens is, period. If you look at the raw and completely unsharpened image from a digital camera you will find they are all incredibly soft.

Sorry folks, any evaluation of image sharpness of one lens compared to another by means of dSLR output files is hype simply because there is not any way to separate the performance of the lens from the action of the sharpening filter affects applied.

It's obvious that if the camera is the same and therefore the algorithms are the same, and of course the subject and lighting are the same, then the only variable can easlily be the lens. And one lens can happily provide a sharper image than another. I don't really think there's much of an issue with that, so I wouldn't agree that it's hype. It is certainly not a hoax, which is a deliberate attempt to mislead.

A sharpening algorithm is not really a constant because it functions relative to the information put through it. That's why with Photoshop for instance Smart Sharpen has a fairly complex set of sliders that allow adjusting the function of the algorithms to the attributes of the image.

The hoax is to me what people, especially those with analog film experience try to assume about optical functions with dSLR camera's conflating the performance effects of lenses with film to having the same advantages/disadvantages with digital. Few people I have corresponded with have ever set their camera on Raw and then turn down or turn off sharpening and exposed a test subject, and then without any sharpening in conversion opened the image and viewed it a 100% magnification on screen. Do that and from what you see, if you are the least bit logical, you have to wonder just how much optical performance can make very much difference.

I've had Canon SLR's for years and have several lens adapters to a Canon EOS body so I can use a really odd collection of lenses I have collected over the last 40 years with my current Canon 5D. It is really quite amazing what you can do with software sharpening by setting parameters in the camera and Photoshop adjustments to make up for optical deficiencies in sharpness, contrast and detail definition.

I have nothing to gain by bringing this perspective to your attention, other than maybe you will think twice about such and such a super-duper digital lens high cost, and will it really buy you anything more than an ego boost and bragging rights.

With respect, I think you are assuming a few things that may not be correct. Firstly, yes I did switch off the sharpening in RAW. That's not the place to apply sharpening IMHO. Secondly, this has nothing to do with ego boosts or bragging rights, which are if I may say so a strange concept. What has that got to do with the assesment of lenses? The plain facts are what I personally sought, and I can assure you that the better Pentax lenses at least perform better than the cheaper ones, hardly an amazing conclusion. What is interesting is that the difference when stopped down is subtle and may well not be noticed by the casual eye, but the difference wide open is more marked. This means that the cheaper lenses can actually be quite useful for portraiture, with a nice softness that can be exploited. They are not to be dismissed for critical work either, and the pentax kit lens is actually well reported as probably being the best kit lens from any manufacturer.
I make no attempt to comment on Canon lens performance, as I have not used Canon cameras or lenses. The Canon kit lens is not highly regarded but I have no personal experience of that.
To answer a specific point you raise, my tests were made both on screen and then as A3 prints as well, both unmodified and then modified as I would normally process any picture (to give the optimum result from each file) so I do think as far as logic goes the methodology was sound.
I accept that you cannot be aware of any of this and your comments are generalising and it is true that there is a lot of dubious opinion on the web, and even sometimes it creeps into print in magazines.

Hummmm.
Why buy a good lens??? Maybe purchasing a better lens is looking toward the future. In about 5 or 10 years digital sensors will be so far ahead of film that all of you will be throwing away your cheap glass and I will still be using my L series stuff.

I have a tokena f2.8 28-80 pro and have compared it to the 24-70 2.8 L canon and you can tell a definite difference even on a 6mp rebel

David,

If a dSLR only has a 6MPX sensor lens quality is not going to show up as being critical any more than it would shooting 800-1600 ISO speed 35mm film, the coarse grain largely precludes much image definition. As the megapixel count goes up the sharpness potential also rises proportionally, also like film if you used slower fine-grained film you get more image detail definition recorded. It's all about what the image recording potential is in terms of how much information a particular media can capture and reproduce.

Most lenses made by any sizable and reputable company in recent times has the advantage of both computer optical design and computer controlled manufacturing which both reduces optical design errors as well as tightens manufacturing tolerances. I have not really seen any bad lenses since the early days of zooms, and that was a long time ago, the early 70's. Some longer range zooms those that cover both wide angle and telephoto however do suffer some because they have so many elements and refracting surfaces cause image internal contrast to be lowered. This would show up in a digital SLR capture more obviously than a slight difference in optical resolution.

I noticed this testing the Epson RD1 rangefinder digital camera using a couple of new Rollei/Zeiss lenses with the camera, getting punchier images with better separation of tones than what I am used to shooting with SLR zooms that have instead of 5 or six element 10 or 12 and in a more complex retrofocus optical design.

Most of the differences like what you are reporting are comparisons from everyday shooting of different subjects under varying conditions. They are not the result of optical tests done under highly controlled conditions that eliminate all of the variables that may affect image resolution and apparent sharpness except differences in lens performance exclusively.

Its like when the Canon digital came out with the 18-55mm kit lens. There was an immediate bunch of carping about this lens in the web forums. So I took one and a Rebel and my personal 20-35mm f/2.8L that new cost almost 10X as much, and a heavy Gitzo plus a 50lb sandbag to weight the tripod down solid, and set up a shot of a very detailed interior of a mission chapel. Without changing any camera setting I took a series of exposures with exactly the same image framing with each lens the Rebel Kit lens and the Canon L. Back in my lab I made 12x18 inch image prints on 13x19 paper from exposure made by each lens. You could not see any difference in detail and definition in either print, and I showed them also to a quite astute local photographer friend and neither could he see any difference.

This could go on endlessly and isn't really a useful thing as we can't see what we are talking about, we can only describe it. I don't see any difference in your Canon lens test to other tests that use the same situation to make comparisons.
The Epson RD-1 isn't very exciting results-wise from the accounts I have read - as is often the case with expensive digital kit that delivers little compared to less expensive marques. I think there's more point to that than the claim that there is no difference between lenses, which is demonstrably not what we have observed. However, if that's what you have observed then we will just have to agree to disagree!

Fine, live with it!

David
I too used a tripod and a test pattern, equal settings on both lenses and made identical shots, the Tokena was the looser, it was visable at 6mp, I even sent the results and pics to tokena with the lens. they just shrugged and said it was normal.

David,

There are some dogs out there in lenses, but as I said elsewhere about current model lenses, any of the more popular brands use computer design and computer controlled manufacturing, eliminating most of why some lenses in the past were dogs. But when some people buy a price rather than a product, its really too low to obtain anything but a dog.

And of course if I haven't done the testing myself everything else is just too salty to be trusted<S>.