[flyer]

Sigma doesn't have it.

[laissezfaire]

When the D90 first came out, I was choosing between it and the K20D. I spent a considerable amount of time at B&H comparing the two. This poster did a great job summarizing, so I will just add a few comments.

I will agree with you about two things: in-body image stabilization, and the value/cost ratio. These are two of the biggest reasons to buy Pentax.

Also, the weather sealing did it. Keep in mind that for just a bit more than cost of the VR, PLASTIC mount lenses like that 18-105 that you have, you can have a DA STAR lens that has an in lens supersonic motor (SDM), is sealed again dust and simply shoots phenomenal pics. You will pay 50-75% more for comparable Nikon glass with in lens VR and motor, but it will also be full frame glass whereas the Pentax DA series is not.

There are some other things about Pentax cameras that are truly wonderful. The build quality of the K20D is fabulous. Similarly priced Canon and Nikon cameras feel like toys by comparison. But the Nikon D90 doesn't feel like a toy to me.

I couldn't disagree more. The D90 feels like a toy to me. The K20D feels like an instrument.

I wish you were in North Texas. We could trade cameras for a week and perhaps both of us would feel better afterwards. :-)

Why don't you guys just swap cameras?

[jbinpg]

One correction in the above: it is only the shorter focal length DA lenses that will vignette on a full frame sensor. Longer focal lengths are fine on either.

Jack

[Eigengrau]

You simply can't argue the physics.

Really? I haven't seen you show any "physics" to argue against.

The biggest argument I've seen for in-body comes in the form of the fast primes. Canon and Nikon have few stabilized primes, and nothing like a cheap fast 50 with lens stabilization. Most of their stabilized offerings, in fact, are no faster than 2.8.

So, having a range of f/2.0 and faster lenses, plus the in body stabilization provided, makes up for any 1 stop advantage that stabilized lenses have. Plus, these fast primes are sharper and smaller than stabilized 2.8 zooms.

Overall, this means a couple of things -
1. In the cases where Nikon and Canon offer stabilized lenses, Pentax can usually come within 1 stop of their performance with a similar lens and their in-body stabilization.
2. A Pentax photographer could use primes in this instance, and gain back that stop or even more.
3. If you are shooting a Canon or Nikon and using a non-stabilized prime or zoom vs a similar Pentax, Pentax will win hands down by 2-3 stops vs the non-stabilized cameras.

Pentax has a compelling system. There are advantages and disadvantages, but I find that for the dollar conscious Pentax is a hard one to beat. Even for those with surplus funds, Pentax stays competitive.

Sorry to rant- but you can't just throw "physics" out there and call it a day. I'm currently majoring in physics, and when you try to do things like that to win arguments, it makes science sad.

[Marc Sabatella]

Really? I haven't seen you show any "physics" to argue against.

Well, in one post he did at least reference what he is talking about - the *amount* by which one has to shift a lens element versus the amount by which one needs to move the sensor. Especially at long focal lengths, it does seem quite plausible that the lens elements wouldn't need to be moved as much as sensor.

But on the other hand, we're now talking about having to move a much larger mass, and that's going to be rather more difficult to control precisely. I also wonder a bit at the possibility for the shift introducing CA or focus issues.

Bottom line, I think actually working through all this to see which approach physics says should be more effective in theory is going to far more complex than anyone on this forum is probably able to work through. And in any case, which could be more effective in theory and which *is* more effective in practice may not be the same. So even if someone has done enough post-graduate work in physics at a level where you are able to work through all the math of all these variables to arrive at an answer, I'm not sure that's goin to be more meaningful than simple empirical tests.

The ones I've seen have been all over the map. Some appear to show a full stop advantage to lens-based systems in general; some appear to show virtually no difference. Depends a lot on your methodology. For example, it may be that sensor stabilization is very effective at reducing significant amounts of shake at very slow shutter speeds compared to lens SR, but lens might be remove more shake at intermediate shutter speeds. Or vice versa. It could also be that body stabilization works better for folks whose hands are steadier than average and lens stabilization for folks whose hands are shakier than average - or vice versa. You can't know until you perform each specific type of test.

Conventional wisdom says lens does a bit better especially at long focal lengths, and I have no evidence to the contrary, but to imagine that it's as simple as saying lens is better by 1-2 stops in all cases because that's what the specs say or because that's wha *one* aspect of the physics might is *grossly* oversimplifying what is a very complex topic.

[pentaxmz]

Well, in one post he did at least reference what he is talking about - the *amount* by which one has to shift a lens element versus the amount by which one needs to move the sensor. Especially at long focal lengths, it does seem quite plausible that the lens elements wouldn't need to be moved as much as sensor.

But on the other hand, we're now talking about having to move a much larger mass, and that's going to be rather more difficult to control precisely. I also wonder a bit at the possibility for the shift introducing CA or focus issues.

Bottom line, I think actually working through all this to see which approach physics says should be more effective in theory is going to far more complex than anyone on this forum is probably able to work through. And in any case, which could be more effective in theory and which *is* more effective in practice may not be the same. So even if someone has done enough post-graduate work in physics at a level where you are able to work through all the math of all these variables to arrive at an answer, I'm not sure that's goin to be more meaningful than simple empirical tests.

The ones I've seen have been all over the map. Some appear to show a full stop advantage to lens-based systems in general; some appear to show virtually no difference. Depends a lot on your methodology. For example, it may be that sensor stabilization is very effective at reducing significant amounts of shake at very slow shutter speeds compared to lens SR, but lens might be remove more shake at intermediate shutter speeds. Or vice versa. It could also be that body stabilization works better for folks whose hands are steadier than average and lens stabilization for folks whose hands are shakier than average - or vice versa. You can't know until you perform each specific type of test.

Conventional wisdom says lens does a bit better especially at long focal lengths, and I have no evidence to the contrary, but to imagine that it's as simple as saying lens is better by 1-2 stops in all cases because that's what the specs say or because that's wha *one* aspect of the physics might is *grossly* oversimplifying what is a very complex topic.

Thanks for the assist Marc.

It's too late and I'm too tired (been travelling) to provide the other responder with the 'physics', but if he wants some, I will be glad to contribute tomorrow (or later today).

As you restated, it is simply a matter of fact that a much smaller movement of a lens element is required in order to stabilize the image on the sensor.

In comparison to in-lens stabilization, it isn't possible to physically change the position of the in-camera image sensor, as fast, or in distance.

BTW Mark, AFAIK there is a special moving lens employed to enable stabilization and this controls CA, focusing issues, barrel appearance, etc. Remember that this is Canon and Nikon... one can't imagine that they would engineer an expensive IS lens with serious faults like that.

However, one could also argue that one more lens element, for light to pass through, is not usually a good thing.

Anyhow, I was not advocating the in-lens stabilization system, and I fully acknowledged that Pentax got this one right.

[Marc Sabatella]

As you restated, it is simply a matter of fact that a much smaller movement of a lens element is required in order to stabilize the image on the sensor.

And as I said, I have no trouble believing that. But it also seems quite obvious that this only part of the story. Moving large amounts of glass with precision is bound to be more difficult than moving a *much* more lightweight sensor.

In comparison to in-lens stabilization, it isn't possible to physically change the position of the in-camera image sensor, as fast, or in distance.

Why on earth not? If you can move a lens element at a given speed, what makes yo think you couldn't move a sensor at the same speed? You might need to move the sensor *faster*, which could be harder, but again, given that it is a tiny fraction of the weight, it is by no means obvious that it's actually going to work out to be harder. Do you have an advaned degree in physics and reference to some published papers on the subject to prove your point?

BTW Mark, AFAIK there is a special moving lens employed to enable stabilization and this controls CA, focusing issues, barrel appearance, etc. Remember that this is Canon and Nikon... one can't imagine that they would engineer an expensive IS lens with serious faults like that.

Oh, I'm not saying that it's doomed to failure. Just that those strike me as hard problems to solve. My guess is the designs employed by Canon & Nikon are quite advanced and do a very good job, but it's doubtful that it would be perfect. After all, *no* lens is perfect in these respects even when just sitting still; it seems highly unlikely that adding a special moving lens element wouldn't come at *some* small penalty above and beyond the imperfections that *already* would have existed. Again, not to say this makes the whole idea useless - obviously, that is not true. My point is just that it's got to be a giant oversimplifcation to say "physics says lens stabilization is better" based on one piece of data (differenc ein distance that needs to be moved) and an implication that Canon & Nikon could have somehow managed to make a *moving* lens free of CA & distortion issues when no one in the history of optics has ever made a *fixed* lens that is free of those issues.

Anyhow, I was not advocating the in-lens stabilization system, and I fully acknowledged that Pentax got this one right.

FWIW, I'm not really interested in calling one "right" (and hence, the other "wrong") in absolute terms. I'm quite sure - based on empirical evidence, there are *some* advantages to lens stabilization systems. Just how big an advantage is difficult to say for the reasons I've alluded to, and also how *important* those differences are compared to the basic "obvious" advantage of in-body stabilization we all agree on (works with all lenses) - that's kind of a subjective thing. I'm with you here - I think that assuming body stabilization comes "close" to lens stabilization, Pentax's approach makes more sense for more people. But it's still worth trying to quantify just how "close" it really comes, and my perception is that despite the various reviews out there that attempt to put numbers on this, they are all rather oversimplistic in their methodology. I wouldn't mind seeing some more thorough tests. Not because it would change my own mind about what's right for me, but just to satisfy my own intellectual curiosity.

[tuicat]

I shoot with a k10D, and k20D and I spend a lot of time at work with a Nikon D200 and D300 (however I havn't used it much)

I bought a Nikon D60 for my sister, nice camera for the price.

I have shot with the D70, and played with the D80, and held the D90 in the store.

I personally prefer my k10D over the D200... the D200 is faster, but I really like the feel of my K10.... to be honest I have a brand new 20D... that has barely been used because I always grab the K10. (I really hate the live view feature, I can't stand it.. it seems cheep)

I almost gave up on Pentax a couple of years ago, I really didn't think the ist was up to the Pentax name, and it didn't stand up to the Nikons, or even the low end Can Rebel. I was about to go with Nikon, when the k10 came out.. went with the k10, lucky I waited.

I still think I made the right choice. I had a bag full lens, and the K10 feels a lot like the D200 (which was much more $). The D200 is still a better camera than D90, and my k20 is nicer then both.

To be honest, Pentax lets me down because they have not introduced a pro camera. My k20 cannot sit on the same self as the d300.. let alone the D700. And the D700 isn't even there PRO camera.. it is still considered by many Nikon fans to be semi pro.

CAN and Nik have pro level cameras, Pentax does not. If you are going to stay semi-pro stick with the Pentax.. if some day you want to go pro.. your going to have to change companies. Unless Pentax really surprises me. I sometimes wish I went Nikon, because I keep buying lens, and I am ready for a Pro camera. Even if I am not ready, I want a Pro camera.

But I cannot afford to buy all my glass again (there is a lot.. )

So K20 is better then the D90.. but the D90 may give you more options down the road.

[asdf]

A wedding kit to consider: K20D+Pentax SMC-A 35-105 f/3.5 (constant). The latter lens is highly praised, although it is manual focus.

Manual Focus Forum / PENTAX SMC-A 35-105 F/3.5

[Riktar]

Well the wedding is this coming Saturday and I am as set as I am going to be. I'll be shooting as an assistant and here is what I gots:

K20D - Well DUHHHH,,,,,,
Battery grip - Oh man does the camera feel better with this on!
Sigma 24-70mm f/2.8 macro
7500EDF flash
50mm f2 - Maybe some detail shots?

Just a quick note on focus speed: The Sigma f2.8 does make a noticeable difference in the K20D's ability to lock on. I tried some lower light scenarios and it might get a bit tight (time wise) getting a lock for the shot. Then again I am just a second banana and have the freedom to just roam at the reception. It should not be a problem with static shots.

Does anyone have an opinion on what is the fastest focusing lens they have used on the K20D?

I think I will be looking into grabbing a focus screen for the camera.

[Riktar]

While driving with the K20D I spotted these 2 and had about 4 seconds to stop, grab the camera (Thankfully it was on) and rip 2 shots off through the passenger window. Both images are cropped since I did not have the time to zoom much before they bolted.

Both shots were PP is CS3 with auto levels and light Smart Sharpening.

Edit: Both pictures have been moved to the Post Your Photos area. I assume that would be better than posting here.

https://www.pentaxforums.com/forums/post-your-photos/53476-couple-deer.html#post516623

[asdf]

Does anyone have an opinion on what is the fastest focusing lens they have used on the K20D?

I don't own it--although I'm thinking of buying one for my wife--but DA 40 f/2.8 seems to be exceptionally fast to focus, which makes sense since the motor has to drive something very light (EDIT: Also, make sure your batteries are fully charged http://joshs-photos.blogspot.com/2008/11/focusing-speed-comparison-pentax-k200d.html):

YouTube - Pentax K10D + DA 40 Lim AF speed

Although, a brighter lens may have better luck in low light with the AF electronics. Hopefully someone with both a FA 50 f/1.4 and a DA 40 f/2.8 can tell us.

As for focusing screens: K20D's viewfinder at f/1.7 seems to be bright enough so that I can manually focus easily in fairly dim conditions in-doors, without a split prism. When using a split prism focusing screen, you'll still probably watch the focus-confirming hexagon anyway, while the focusing screen may be off unless you install a different shim. Install a split prism focusing screen, if you don't mind getting some dust specs on it or are equipped to clean it.

Good luck.

[kenyee]

Although, a brighter lens may have better luck in low light with the AF electronics. Hopefully someone with both a FA 50 f/1.4 and a DA 40 f/2.8 can tell us.

It not only has to be fast, but also have contrast wide open. The 50/1.4 doesn't have enough contrast wide open and didn't work all that well for bride coming down aisle shots w/ my old K10D...

[WMBP]

A wedding kit to consider: K20D+Pentax SMC-A 35-105 f/3.5 (constant). The latter lens is highly praised, although it is manual focus.

Manual Focus Forum / PENTAX SMC-A 35-105 F/3.5

I shoot weddings with a K20D and a K10D, so I agree with you about the camera. But I personally would not take the lens you mention to a wedding, for a few reasons. First, I don't want to rely on manual focus. Second, f/3.5 might work if there was better than average light, but I shoot a lot during the ceremony at f/2.8 or faster (I have a couple faster primes that I use if I need to). Finally, the 35-105 range works well for a 35mm film SLR (full frame) but for my wedding work, and given the crop factor on Pentax cameras, a 35-105 lens is not wide enough at the wide end and not long enough at the telephoto end. I tend to go into the ceremony with a 17-50 f/2.8 on the K10D and a 50-135 f/2.8 on the K20D.

If I had to use just one lens to shoot a wedding it would be the Tamron 28-75 f/2.8. But I can't imagine using just 1 lens now.

Will

[rfortson]

(...snip) My k20 cannot sit on the same self as the d300.. (snip...)

Yep, with lower resolution, no weathersealing, no stabilization, and at a price 50% higher, it's really not fair to compare the D300 to the K20D, is it?