Testing123 said:
Because the "pixels" on the D3s are much larger than the D90 and D7000 and therefore resolve less lines per mm and therefore need a less perfect lens.

The D3s has the same resolution as the D90 per picture height but lower per mm. One must understand the distinction between the two to understand lens sharpness.

Thanks for your reply, Testing! Yes, I reasoned as much. But I'm still having a bit of trouble getting my brain to accept that explanation.

hey super i truly love a great shot be it a picture or a pun!

if a body is soft focusing it is out of alignment. The front lens bayonet needs to be perfectly aligned with the sensor plane where the image is recorded. there is a tolerence which the lens bayonet needs to be within usually a few fractions of millemeters which is necessary play in the mount so that the lens can be mounted onto the body. if the body or lens is dropped or banged around this is usually the area that takes the most stress/abuse. regarding softness of the lens, having a metal mounting/bayonet plate is preferable to some of the plastic mounts found on some lenses. the overall sharpness of a lens is primarally due to the design of the element groups themselves and the quality of the glass and multicoating used. If a lens gets dropped or banged around it can affect the focus as the lens goups or elements themselves can shift and/or the lens barrel can become distorted causing a shift of the lens groups which will result in softness due to misallignment.

:) There are only two? I mean these arguments all sound the same. They just keep going on and on.

So which 2 esteemed members of our beloved forum are you refering to??

http://www.youtube.com/watch?v=KRWJybzkwN4

We are not there yet Niko! But give us a few more days.... :)

Testing123 said:
Nope. That is not.... what I said.

OH God! Here we go again :^(

Thanks for the explanation of that, makes things a lot clearer.

PB PM said:
Actually, it is the exact opposite of what you said. You said that a lens should appear sharper on FX than DX.

Nope. That is not a safe conclusion to draw from what I said.

I said that low-density sensors (as currently found on FX bodies) are more tolerant of lens flaws.

I also said in the follow-up that high-density sensors (as currently found on DX bodies) have a weaker AA filter and are therefore capable of greater resolving power when measured per mm.

The distinction is important.

Which "class" of sensor produces a sharper image with any given lens (of course taking into account the variables TTJ mentioned) is dependent on the particular lens and spatial density of the lines one is using to judge sharpness.

Studio460 asked "...is this possible" and my reply (in a nutshell) was "yes, it is possible and this is why:".

TTJ mentioned many factors which are more likely to be the cause, but I can only assume that studio460 accounted for as many variables as possible and was asking an honest and informed question. He has shown no recklessness in these matters in the past. It would be rude of me to assume otherwise this time. Therefore I gave a factually sound answer with as simplified of explanation as I found possible. If I led you astray through the simplification I apologize.

TaoTeJared said:
I think it is Bjrichus that said he has a tack sharp 28mm F2.8! He is the first person I have ever heard of with one of those! Variances exist. We shouldn't make conclusions based on just one sample or even a couple. Usually see variances as being bad but sometimes they work out great!

Indeed. I am kind of/half way/almost ready to buy a couple of other focal length lenses and do some *REALLY* detailed side by side testing. Ah but for the time (and money) to actually DO it. :-)

For me, the message in this thread and the article we are all referring to is that ALL of our photo gear is made to be within a set of tolerances - and the trick for "good" results is to get them as compatible as possible - and to understand that a new body/lens may produce worse results, even if it has outstanding specs when used with the old body/lens... and why...

Testing123 said:
Sure it does. Reread what I said.

Think also about the way in which a sensor's AA filter works and the implications this creates for spatial resolution as it relates to photosite size.

Actually, it is the exact opposite of what you said. You said that a lens should appear sharper on FX than DX.

If you read the article they go into much detail what various testers have found.

The conclusion I drew from it, is that at the end of the day, there are variances in bodies, mounts, lenses, sensors, pixel density, light gathering power, micro lenses, analog converters, chip based software, bit depth, contrast, and about 10,000 other details that camera makers look at to try to "get it right." To oversimplify and state it is just MPs, or density or any singular item is not being fair to the 1,000s of engineers that work at camera companies who do this for a living every day. I think it is amazing how sharp every lens is on any camera considering that there are 100s of lens combinations the designers have to consider will be used.

What some testers found was that the variance in many items was smaller than what is allowable with medical devices. I have read articles where some labs use lasers to test lenses - scientific (research on my 60mm and 105mm macros) and the end conclusion was that there is no image capture technology that exists today or in the near future that can capture what lenses are capable of.

With all of that said, it is well known that lenses do have sweet spot or have their focus set at the factory at a particular F-stop and it can be changed. The ability to also adjust focus in the camera is a huge plus. I have found my 50mm AFD was front focused at F1.4 and was able to adjust it and save it. Every time I put the lens on is auto-magically sets it. Doing this is requires a lot of testing using something like the product from Lensalign, and much reading and full understanding on what you are doing. For instance, for my lens above, it now focused at F1.4 but was slightly off at F2 and disappeared above that due to DOF. This is acceptable to me since I would rather shoot at F1.4 than F2.

For those who think DXO actually does something creditable with lenses, I ask, when is the last time you heard them getting a lens calibrated to their cameras or have calibrated it themselves? They test sensors and maybe can make generalizations about 1 lens on a body with the Serial number XXXXXXXXX.

I think it is Bjrichus that said he has a tack sharp 28mm F2.8! He is the first person I have ever heard of with one of those! Variances exist. We shouldn't make conclusions based on just one sample or even a couple. Usually see variances as being bad but sometimes they work out great!

PB PM said:
That doesn't explain why my 50mm F1.8D got sharper results on the DX bodies I tested it with (D90/D300), while it looks soft on FX (D700). I compared the center of the frame, so it isn't an issue of not looking in the "sweet spot." I did the test via liveview auto focus, so phase detect AF errors cannot be the issue. I also did a manual focus test, just to make sure AF wasn't the issue.

Sure it does. Reread what I said.

Think also about the way in which a sensor's AA filter works and the implications this creates for spatial resolution as it relates to photosite size.

heartyfisher said:
LOL! So would anyone here be buying a set of +5 and +4 cameras and lenses!

If they were 50% less than list price, then maybe.... at least it would be a (more) informed purchase, which is better than now!

That doesn't explain why my 50mm F1.8D got sharper results on the DX bodies I tested it with (D90/D300), while it looks soft on FX (D700). I compared the center of the frame, so it isn't an issue of not looking in the "sweet spot." I did the test via liveview auto focus, so phase detect AF errors cannot be the issue. I also did a manual focus test, just to make sure AF wasn't the issue.

studio460 said:
Soft on DX; Sharp on FX?

I just got home with my new (refurbished) D3s, and the first thing I wanted to do was to throw my AF Nikkor 18mm f/2.8D FX lens on it to see how sharp it looked. This lens always seemed a bit soft on my D90/D7000. Well, some very unscientific test shots show the 18mm f/2.8 looking near tack-sharp on the D3s. Any reason these observations may actually be valid?

Because the "pixels" on the D3s are much larger than the D90 and D7000 and therefore resolve less lines per mm and therefore need a less perfect lens.

The D3s has the same resolution as the D90 per picture height but lower per mm. One must understand the distinction between the two to understand lens sharpness.

I asked a similar question a month or so back and it opened a can of worms. There were lots of explanations from knowledgeable members, but none of them really answered my question. Both the D90 and the D7000 (especially this one) are saturated with pixels versus the D3S, which has much fewer pixels per square mm than either. The D7000 works out to 36 megapixels if the same per mm density was spread over a full size sensor. I noticed that it was tack sharp (in production images) with the 35 1,8DX lens, but with every professional zoom, it just seemed 'fuzzy' for lack of better term.

I expect to get flamed, but I don't really care. We don't have FX cameras that are anywhere near the per mm resolution of the D7000 not to mention the 7D. When we do have them, we can have a proper argument, or we can finally see what the truth is. Pundits will say that it is all the light going to the sensor, or the AA filter. But until we have FX sensors of 36 megapixels or higher, we can't prove that such a highly packed sensor is or isn't outresolving the lens.

My D200 is 100% sharp with all of my lenses, but then again, it is just about as packed per mm as the D3X, and that is a long call from the density of the D7000.

studio460 said:
Soft on DX; Sharp on FX?

I just got home with my new (refurbished) D3s, and the first thing I wanted to do was to throw my AF Nikkor 18mm f/2.8D FX lens on it to see how sharp it looked. This lens always seemed a bit soft on my D90/D7000. Well, some very unscientific test shots show the 18mm f/2.8 looking near tack-sharp on the D3s. Any reason these observations may actually be valid?

NikoDoby said:
And by the why the article was updated in March of 2010 from 2008 and not recently.

Looks like my internal clock is still in 2010 :D

Soft on DX; Sharp on FX?

I just got home with my new (refurbished) D3s, and the first thing I wanted to do was to throw my AF Nikkor 18mm f/2.8D FX lens on it to see how sharp it looked. This lens always seemed a bit soft on my D90/D7000. Well, some very unscientific test shots show the 18mm f/2.8 looking near tack-sharp on the D3s. Any reason these observations may actually be valid?

bjrichus said:
And if the likes of Nikon were to quantify products (as per the article, -1, 0, +1 as an example), we wouldn't need to pay $400 to tweak ...

LOL! So would anyone here be buying a set of +5 and +4 cameras and lenses!

heartyfisher said:
If you are really picky about having perfect focus then you simply need to get your whole set of lenses and cameras matched. It will cost but it is done by a few pros. even the d90 can be tweeked by the good service centres.

And if the likes of Nikon were to quantify products (as per the article, -1, 0, +1 as an example), we wouldn't need to pay $400 to tweak ...

If you are really picky about having perfect focus then you simply need to get your whole set of lenses and cameras matched. It will cost but it is done by a few pros. even the d90 can be tweeked by the good service centres.

The original article (updated) does raise a good question especially in regard to my recent 'test' of a couple of old FX lenses and color as I also found less of a difference between one pair of lenses than the other pair as well as each other. Ok, so one of the FX lenses was clearly better than the DX but the other one (the 70-200) was much closer to the DX lens than I imagined it would be.

It appears to me that us buyers have NO WAY to know where a lens is on the QA scale used to describe if a lens is acceptable for sale.

I am not saying that a maker should not use any particular %age + or - of theoretical 'perfect' but that they should TELL US what that %age is.

This would allow them to charge more for lenses closer to 0 and less for those further away from 0 .....

Oh wait! By not telling, they can make more that are further away from 0 than those closer and charge us more for them all ANYWAY!

Cynical? Me? :)

Doesn't all this underline the importance of finding a lens that works well on a particular body and not expecting these lenses to work well on a new body... ???

Ok... I'll shut up now and let everyone else argue this one!

Sorry Monty, but the forum rules don't allow direct links to commercial sites anymore. The rule has to apply to everyone including mods. Borrowlenses is also now a sponsor.

Google "this lens is soft and other facts" to see the article. And by the why the article was updated in March of 2010 from 2008 and not recently.