I usually look to see if the number back up my opinions. If they do I post them. :D
If they don't, I try and figure out why.
I don't think anyone thinks the DA*16-50 is a bad lens... the question is, can I get away with a DA 18-135 for the types of images I shot, and the answer is, for the most part I can. Part of that is having the 60-250, which covers it's weak end, and part of that is having primes in it's focal length.
It's good points are listed above in. My consideration in most of these discussions is the idea that IQ favours the DA* lenses in this case. I'm not convinced that's true. There are areas where the DA*s are on their own, at ƒ2.8 and in the long end at ƒ4 as well. But in real world comparisons, I don't see an advantage. A guy who can take good images with the DA*s can take good images with the DA 18-135. The IQ differences are relatively minor.
When I test all my lenses cable of achieving 90mm, in many ways even the cheapest lenses, like the Sigma 70-300, the images all came out pretty good. But at that time the three that really stood out were the Tamron 90 and DA*60-250 with the 70-300 in close third. It quickly became obvious that resolution wasn't the ticket. When i looked up the difference between the Tamron 90 and the 60-250, the only thing that stood out was were the CA values. To me, there is a correlation between CA values and micro contrast. The sigma, had the worst CA by far. It even showed some fringing on my test images. The DA*60-250, probably one of the most corrected zooms in existence, and stacked up well, but the Tamron 90 beats it in control of CA. The published resolution numbers for the two lenses are close enough to be a saw off, so the CA was the last factor un-accounted for. We were able to agree, myself looking at the images, my wife in a blind test, in that she didn't know which image was which, and we both agreed the Tamron 90 image while practically identical, look sharper, because it had better micro-contrast. There was no significant difference in detail, even with the Sigma 70-300. SO my understanding would be, that resolution while important is over-rated. In earlier discussion with jsherman and the like I actually laid out mathematically why that's true. (Calculate the difference in a 20 inch deep picture the difference between, 2000 lw/ph and 2300 lw/ph, and tell me you can discern that difference looking at an image from 8 inches).
But, even CA under a pixel in width affects micro-contrast, and the appearance of sharpness. My rule of thumb has become to not buy lenses with their highest CA value ever .7 pixels for the camera I'm buying and I prefer it from .2 to .4 pixels.
Of course I have many lenses that don't meet the criteria, at the time I did the tests. Of the 5 lens I tested, the fourth one being the 18-135, which as it was a landscape with all the detail at infinity finished very poorly because of edge detail, and an old 35-80 I threw in just as a sort of control, to make sure I knew what bad really looked like.
But after all that pixel peeping, the end result was, any of those top 4 images was good enough to print. The differences between the top 3 were only noticeable pixel peeping on a 27 inch monitor. And I have yet to determine what the output would be, that would make that level of excellence or would be able to even take advantage of that level of excellence.
Following this particular line of thought, I've also done tests that show a 36 MP image is indistinguishable from a 16 Mp image by about 7 or 8 Mp when you reduce the size. So long story short, on my monitor which is 27 MP across, shooting uncorked images, you won't be able to tell the difference. And I know of no test that establishes how big you have to print to be able to tell whether you shot 36 MP or 16 Mp.
All in all, what my rather un-scientific and cheaply executed investigation have led me to believe is that being comfortable with what you shoot and taking full advantage of the strengths of whatever you have, is more important than the actual hardware. People want to believe they can buy an advantage. If you look at work like MikeSF his work was spectacular with a K-5 and it's spectacular with a 645D.. but is it more spectacular with the 645D? That's a really tough call. People need to understand that when comparing things like a 16-50 and 50-135 combo to a 18-135, there are times when both are going to give you the better image. And sometimes you're going to get the image with the 18-135 because if you stop to change lenses you're going to lose it. And other times, like shooting at ƒ5.6 or ƒ8 you may get a better image with the 18-135 because it controls CA better than the 16-50 or the 50-135 at some FL and apertures.. Same at 135mm compared to the DA* 50-135 if the edges are OOF.
This blanket assumption that when you buy a better lens you get the better image is seriously flawed, at least that's what my testing would show. If you buy the better lens and can use it to it's strength, you get the better image. But you can't always use a lens to it's strength. It may well be that you buy a more expensive lens and the image you get in particular circumstance is one you could have matched or surpassed with a cheeper lens, just because in the case of zooms, the cheaper zoom just happened to be sharper with less CA at that focal length. At that point it doesn't matter that the more expensive lens might have come out superior in 90% of the images you shoot. If the best looking image is your goal, you should have been shooting the cheaper lens.
I tend to think, unless of course you're spending thousands on your lenses, buying better glass gets you better odds of getting better IQ, but not necessarily better IQ on every image you take with it. IN the case of the DA*16-50 and DA*50-135, the CA values leave me shaking my head. But I also know, if it's good otherwise, the likelihood of having an output device that is accurate enough to show the differences are pretty remote. If it's acceptably sharp, has reasonable control of CA you're probably splitting hairs. Which is what we had to do to see any difference at all between our top 3, one of which can be had for $200.
So at this point my opinion is, take picture with what you have, it's probably good enough. You can sell images taken with what you own, and when you do, you can buy something better. But to me, the reason for buying better glass is, your skills have advanced to the point where it will make a difference. And people who have reached that point, don't need someone else to tell them what they need. So my answer to "should I buy better glass?" is always no. If you don't know why you need it, you don't need it.
