[BarryE]

Accepting sharpness is not necessarily the most important characteristic of a lens, it does seem to evoke much comment when lens quality is judged.

One lens that often comes out on top is the d-fa 100mm. My question: is this lens and others that are ff compatible, felt to have better perceived sharpness because in cropped sensor cameras the natural edge fall off is not seen ?

If so might these sharp, on cropped, ff compatible lenses be shown to have weak edge sharpness on ff ? I guess this means that until the ff is released AND each lens is tested we will not have any idea how they will perform on ff. With the suggested delay of the ff release pushed back, any notion of re-jigging my lens collection gets delayed further - probably for the best ...

[fwcetus]

Accepting sharpness is not necessarily the most important characteristic of a lens, it does seem to evoke much comment when lens quality is judged.
One lens that often comes out on top is the d-fa 100mm. My question: is this lens and others that are ff compatible, felt to have better perceived sharpness because in cropped sensor cameras the natural edge fall off is not seen ?

Well, there certainly are a lot of K, M, F, and FA (and other) pre-DA, FF lenses out there, and many, many, many of 'em have been used on cropped sensor Pentax DSLRs for a number of years now. The assumption seems to be that APS-C bodies utilize just the "sweet spot" of the lens, and therefore the lenses all have "better perceived sharpness" on APS-C sensor DSLRs.

However, I cannot think of any specific FF versus APS-C tests of FF lenses offhand myself. Can anyone else come up with some specific tests (either personal and/or published)? "Inquiring Minds Want to Know."

[Adam]

Even in the extremes of the frame it's rare for modern first-party prime lenses to perform poorly. You'll see more vignetting and possibly softer edges, but it shouldn't be excessive. While we will have to re-test a lot of lenses, plenty of users have already tried them out on Sony FF mirrorless bodies and on film cameras, so we would have heard of any drastic issues by now.

[enoeske]

Accepting sharpness is not necessarily the most important characteristic of a lens, it does seem to evoke much comment when lens quality is judged.

One lens that often comes out on top is the d-fa 100mm. My question: is this lens and others that are ff compatible, felt to have better perceived sharpness because in cropped sensor cameras the natural edge fall off is not seen ?

If so might these sharp, on cropped, ff compatible lenses be shown to have weak edge sharpness on ff ? I guess this means that until the ff is released AND each lens is tested we will not have any idea how they will perform on ff. With the suggested delay of the ff release pushed back, any notion of re-jigging my lens collection gets delayed further - probably for the best ...

You can use pentax lenses on other FF cameras, you know. We have an idea of how they will work.

[Na Horuk]

One lens that often comes out on top is the d-fa 100mm. My question: is this lens and others that are ff compatible, felt to have better perceived sharpness because in cropped sensor cameras the natural edge fall off is not seen ?

There are two schools of thought.
1) APSC and FF doesn't matter, what matters is the resolution of the sensor. 24MP APSC has higher resolution than 24MP FF, and therefore the APSC is more demanding of lens resolution. However, the FF uses different parts of the lens' image circle - parts that the APSC sensor does not cover. And usually lenses have their weak spot in the image edges, so it is possible that on the FF, the extra FoV, the edges, will appear to be softer than the edges of the APSC sensor. But if you take that FF photo and crop it digitally to the same FoV as an APSC sensor, then the edges would look the same. The sensor does not affect the image quality, but it can record more or less of the image circle, and it can have higher or lower resolution.
2) This school of thought uses odd concepts like "equivalences" and comes to the conclusion that FF sensors improve the resolution of any given lens (Im simplifying here). The same lens will suddenly appear to have higher measured resolution on FF. But again, the corners are another matter, because the APSC sensor simply does not record that area. Here the idea is that the sensor does somehow affect image quality, as far as I understand. (I fit into the former school of thought, don't agree with this one)

Tl;dr: Your DFA 100mm will be super sharp on APSC, and just as sharp on FF. The lens will not be the bottleneck, will not be the weak spot, even if you use 42MP sensor, APSC or FF. But can the extreme edges on FF be worse than edges on APSC? Especially if you use sensor shift? Yes. Edges are sometimes almost as good as centre, but never better than it. But that does not mean they are terrible. And with FF you can always crop to APSC (or APSH) to cut off the "poor" edges.


---------- Post added 23rd Sep 2015 at 23:26 ----------

I guess this means that until the ff is released AND each lens is tested we will not have any idea how they will perform on ff.

The DFA and FA lenses should perform at least adequately in the edges of FF. The DA don't have this guarantee. There is a thread called DA lenses on full frame (or something like that, feel free to search for it), where pretty much all of the lineup has been tested on film bodies or adapted to Canon, Sony FF cameras. Now, sure, these are not laboratory sharpness tests, but they do give an indication. Beyond that, yes, we will have to wait for testing companies to buy FF camera and then do tests on it. This will probably take a long while, as those companies generally don't care much about Pentax cameras or lenses.

[stevebrot]

I guess this means that until the ff is released AND each lens is tested we will not have any idea how they will perform on ff.

No, not really. As mentioned above, K-mount FF glass is pretty popular among users of Sony FF cameras and if there are complaints, they must not be very loud. In addition, with the exception of the newest D FA models, all previous series of Asahi and Pentax glass have a history of one to six decades use on 35mm film. Without launching into the film vs. digital debate I will merely suggest that a usable negative is just as hard to get as a usable high resolution digital capture and that the usable print sizes from each (without upsampling) are similar.

Here are a few bullet points repeated from the dozen or so similar threads in the last month or so with maybe one or two extra thrown in:

• A given lens will perform as well or better on FF digital as it does on 35mm film with very few exceptions*
• In general, a given lens will perform as well or better on FF digital SLR as it does on APS-C digital (assuming same FOV/magnification, same pixel count, same flange registration distance)
• The prominent exception to the point immediately above is corner sharpness for many (most?) vintage wide angle lenses (28mm and wider). The reasons are related to how those lenses are used on the 35mm format and the fact that most surviving wide angle glass was made for the consumer market. A true flat field in a 28mm or 24mm lens was fairly rare in 1980 and the same was true in the year 2000. The corners were soft on those lens when they were new and used with film and remain soft today.
  
  Another prominent exception are lenses whose original design goals did not include edge-to-edge sharpness. A good example would be the Zeiss Biotar 58/2 and derivative lenses in the Helios-44 series. The Biotar was designed for excellent center sharpness at all apertures, but at the expense of the corners. Both the Biotar and Helios lenses perform quite nicely corner to corner on APS-C, but the soft corners are very evident on FF cameras.
• Incidence angles are not a huge issue with SLRs and the lenses designed for them...the mirror box is our friend. Assuming adequate image circle, vignette for FF digital should be essentially similar to 35mm film.
• For a given un-cropped composition, a pixel is a pixel is a pixel regardless of format

I probably missed something A good rule of thumb might be to consider the experiences of Nikon users where film era lenses on FF digital has been a fact of life for several years.


Steve

* Things like purple fringing (essentially a sensor artifact) and rear element sensor reflections (possible on film, but much less so).

[KDD]

DA 35mm f2.4 is full frame, too.

[Paul MaudDib]

The issues with older non-telecentric lenses fringing and smearing in the corners on digital is the main issue. It's highly likely to be a Sony sensor, so once we know the resolution we can figure out which A7 camera it was in previously and then you can figure out how your lens performs on it. It's rumored to be the 42 MP A7R-ii sensor, which is supposed to have finally solved the issue.

There was a gigantic leap forward in optical design during the 90s which occurred due to 2 factors. First, computer-aided optical design became commonplace. Second, Pentax's patents on SMC expired and were immediately copied by everyone and their brother. The improvement in contrast and flare allowed much more complex optical designs to take off. Today these are considered solid but uninspiring lenses, like the -F and -FA type series. There has been another revolution within the past 5 years as we see even more complex designs that use multiple fluorite, UD, and aspheric elements. Sigma, in particular, is kicking butt and taking names, with superfast primes and superfast zooms.

Vintage lenses are what they are, just like modern lenses. Some lenses are sharp, some have a gentle rendering, and so on. A lot are mediocre nowadays or just bad. Good equipment has always been extremely expensive (with limited exceptions), and optical design has improved drastically over the years. Typically the truly top-shelf glass was so expensive that very few units were sold and they command high prices even today. These are often not worth it compared to their modern counterparts. The K28/2 is a great lens but it's not $800 great. You want to look for lenses with character or value. For example the K35/3.5 is a fantastic lens even today, and relatively affordable. But the various K/M/A 35/2 are not worth it compared to their modern counterparts. They are equally or more expensive than current gear, they don't have any distinguishing optical signature, and they're not that sharp by modern standards. Anything wider than 28mm - same boat, they are really hard to make sharp without exotic design and the modern stuff spanks them even at the same or lower price. And bargain-bin stuff? Well, Joe Sixpack never pixel-peeped his 4x6s and doesn't know what he's missing.

Sigma's Art series are all uniformly fantastic. The same is true of Samyang - they are admittedly manual-focus only, but if you were thinking of doing vintage glass you should be buying a split-prism focus screen anyway. The 85/1.4 is the only one that falls into the "good" range in their lineup, everything else is excellent and above. If you're jonesing for a 24mm, 35mm, 85mm, 135mm and are considering glass, give their 24/1.4, 35/1.4, 85/1.4, and 135/2 some serious thought. They are dirt cheap and rock solid optical performers with modern designs. They usually beat the first-party lenses at 1/3 to 1/2 the price or less. Value is fine, character is fine, but don't buy vintage just because it's vintage.

Incidence angles are not a huge issue with SLRs and the lenses designed for them...the mirror box is our friend. Assuming adequate image circle, vignette for FF digital should be essentially similar to 35mm film.

One note here - the mirror box doesn't change anything, because the register distance is the same. Instead of a mirrorbox, a MILC has an empty tube providing the same spacing and angle of incidence. If a lens has corner problems on Sony FF, barring sensor improvements it won't be any better on a Pentax DSLR.

[stevebrot]

One note here - the mirror box doesn't change anything, because the register distance is the same. Instead of a mirrorbox, a MILC has an empty tube providing the same spacing and angle of incidence. If a lens has corner problems on Sony FF, barring sensor improvements it won't be any better on a Pentax DSLR.

Historically, it was the mirror box that forced the long registration distance. You are very correct that if there are problems when adapted on the A7, those same problems will likely be present on the Pentax FF. Thank you for being explicit. Do you know of any K-mount lenses that cause problems on the A7 series cameras? I don't remember any reports of vignette or color artifact, but that does not mean it has never happened.

I am curious about the telecentric designs you refer to. I am familiar with the advantage on short registration distance cameras and am aware that they are not uncommon for mirrorless system lenses. Are there any K-mount lenses that are telecentric or at least close to being so?


Steve

[Digitalis]

If a lens has corner problems on Sony FF, barring sensor improvements it won't be any better on a Pentax DSLR.

Incorrect. The sensor stack - AA rectifier, low pass filters have different thickness depending on the maker of the camera. Sony tends to have a thicker sensor filter and thus astigmatism is increasingly problematic, especially with lenses with wide angle lenses with high incidence angles at the sensor plane. Thinner sensor filters like that used in the Leica M8 reduce this effect, though at the cost of colour accuracy due to increased UV/IR contamination.

[Na Horuk]

Yeah, Nikon and Pentax have sometimes squeezed more out of a Sony sensor than Sony did in their cameras.. So the Pentax camera, if it uses the Sony sensor, it will be at least as good, but possibly slightly improved.

[Abbazz]

Incorrect. The sensor stack - AA rectifier, low pass filters have different thickness depending on the maker of the camera. Sony tends to have a thicker sensor filter and thus astigmatism is increasingly problematic, especially with lenses with wide angle lenses with high incidence angles at the sensor plane. Thinner sensor filters like that used in the Leica M8 reduce this effect, though at the cost of colour accuracy due to increased UV/IR contamination.

I'm a bit confused on how you intend to adapt high incidence angle lenses on the future Pentax full-frame DSLR...

Cheers!

Abbazz

[Digitalis]

I'm a bit confused on how you intend to adapt high incidence angle lenses on the future Pentax full-frame DSLR...

The Long flange of the K mount takes care of that by reducing the angle if incidence. It is with short flanges where the high angle if incidence becomes problematic.

[Madaboutpix]

Sigma's Art series are all uniformly fantastic. The same is true of Samyang

No offence, but IMHO the adulation showered on Sigma's Art series of lenses sometimes borders on the grotesque. They are daring and probably pioneering optical designs, and they excel in laboratory tests that tend to overemphasize sharpness - admittedly the OP's main concern - over other factors that are at least as important to perceived image quality, regardless of whether we are talking APS-C or FF use. Colour rendition and saturation, for instance, or flare resistance, or the ability to produce images that seem to have that prized "3-D look". In order to achieve their mind-boggling specs, Sigma engineers (have to?) use an excessive amount of glass - yes, they are not alone in following this trend - and one thing gets increasingly clear: the image look you get from such designs is certainly different from the output that slower, lower-element, excellent-coating designs have given us. They have other reported issues as well, some of them, it seems, related to QC, others to the simple fact that they are back-engineered. More than just a few users have reported iffy PD AF accuracy, sometimes beyond fixing with their USB Dock.

I'm really not trying to say they're mediocre lenses, just that their design philosophy has its trade-offs, and that we should try to separate the substance from the hype induced by their smart and high-profile marketing.

Similarly, Samyang's offerings have their unique attractions, too, including sharpness (provided you get a good copy), but how long will the sharpness hold up in actual use if they still have to catch up with truly durable construction? I guess LensRental's Roger Cicala would have a thing or two to say to this. Again, I know, we have to consider the price point we're talking, and intended intensity of use ... I'll stop here, not wanting to hijack this thread for ranting.

[stevebrot]

The Long flange of the K mount takes care of that by reducing the angle if incidence. It is with short flanges, that where this becomes problematic.

Yep...while digital sensor edge responsiveness may be increased with lenses that project more orthogonal to the sensor, the improvement is incremental due to the long registration (flange) distance.


Steve