[BruceBanner]

I recently acquired my first zoom, the HD DA 20-40. I haven't really had the chance to fully play with it yet but thought I might have to do some homework to discover when using what focal length what might be the optimum aperture to use at those ranges. I then realised ephotozine does stuff like this already, and I think PF but not in a graph manner? (Maybe I'm wrong about that).

20mm


30mm


40mm


Summary

If I am to understand their findings correctly it looks like at each range going past f11 results in significant drop in IQ. F5.6 is the sweet spot for 30 and 40mm, however f8 is better at 20mm (but not by much, so for simple memory recall you can pretty much say f5.6 gives the best overall IQ for uniformed centred and edge sharpness).

f2.8 should not be shy'd away from using at 20mm as long as it's centre sharpness you only care about.

At 40mm you will get drastically better results if you can stop down to f5.6.


Would you all agree with those findings?


And now I'm wanting to know, is this not what PF review found themselves? What about our longtime HD DA 20-40 users out there, are their findings in line with ephotozine tests?


TIA!

Bruce

[Rondec]

I think it isn't surprising. On APS-C you are going to start to get into diffraction at f11 and so of course resolution will drop. At the same time, most lenses benefit from being stopped down a couple of stops from wide open. The DFA *50 is probably the only Pentax lens that is an exception to that rule.

[PJ1]

The old caveat that 'MTF is not everything' always applies. But I use MTF charts the same way you do: look for the sweet spot and remember it. It is always a good place to start. I have found that the sweet spot tends to be at smaller apertures at longer focal lengths on zoom lenses. I also like ephotozine. Their tests are easy to understand. The 20-40 certainly comes up well in their test. It should give nice bokeh for people shots wide open at 30mm.

[Paul the Sunman]

I give more thought to desired depth-of-field rather than peak resolution. Nevertheless, those results for the DA 20-40 Limited are very impressive, with some centres going right off the charts!

Thinking back, if I recall correctly, the DA 20-40 f/2.8-4 came out the same week as the Sigma 18-35 f/1.8, and was completely swamped by it in online coverage. Sigma's fast zoom was the big photo news for months; the poor little slow variable Pentax was so meh. However, now that things have settled, the DA Limited is more than holding its own. The PF user rating is 9.24 (42 reviews) compared to 8.77 (32 reviews) for the Sigma, and it doesn't have the AF problems of the Sigma.

Sometimes slow and steady wins the race.

Enjoy the lens Bruce. It looks like fun.

---

Last edited by Paul the Sunman; 05-16-2019 at 04:27 AM.

[BruceBanner]

I think it isn't surprising. On APS-C you are going to start to get into diffraction at f11 and so of course resolution will drop. At the same time, most lenses benefit from being stopped down a couple of stops from wide open. The DFA *50 is probably the only Pentax lens that is an exception to that rule.

What does the DFA* 50 get then, why is it the exception? At more stopped down apertures it still performs rad?

And yeah, if a lot of lenses do get worse after f11... why then is it even really 'offered'. Like this lens goes up to f22, but when would you really ever use that? A lot of landscape stuff where you want a good deal of near and far objects in focus can be had at f8-11, I'm not sure I ever understood surpassing that aperture ever...

The old caveat that 'MTF is not everything' always applies. But I use MTF charts the same way you do: look for the sweet spot and remember it. It is always a good place to start. I have found that the sweet spot tends to be at smaller apertures at longer focal lengths on zoom lenses. I also like ephotozine. Their tests are easy to understand. The 20-40 certainly comes up well in their test. It should give nice bokeh for people shots wide open at 30mm.

Yes, or even 20mm wide open should still do quite well. What's good to remember is at 40mm you probably want to stop down a tad.

I was using my FA43 again this week as well, I noticed an enormous change from shooting f1.9>f2, the mushy centre starts to become a lot more sharper!

I give more thought to desired depth-of-field rather than peak resolution. Nevertheless, those results for the DA 20-40 Limited are very impressive, with some centres going right off the charts!

Thinking back, if I recall correctly, the DA 20-40 f/2.8-4 came out the same week as the Sigma 18-35 f/1.8, and was completely swamped by it in online coverage. Sigma's fast zoom was the big photo news for months; the poor little slow variable Pentax was so meh. However, now that things have settled, the DA Limited is more than holding its own. The PF user rating is 9.24 (42 reviews) compared to 8.77 (32 reviews) for the Sigma, and it doesn't have the AF problems of the Sigma.

Sometimes slow and steady wins the race.

Enjoy the lens Bruce. It looks like fun.

Thanks, I'm testing it quite a lot of FF as well, seems around 25mm is workable.

[bdery]

I think PF but not in a graph manner

Yes, I do that, with at least three focal lengths for zooms. I do not use a quantitative method, but rather a comparative method. I think for most readers looking at examples is more telling than looking at a graph.

[Rondec]

What does the DFA* 50 get then, why is it the exception? At more stopped down apertures it still performs rad?

And yeah, if a lot of lenses do get worse after f11... why then is it even really 'offered'. Like this lens goes up to f22, but when would you really ever use that? A lot of landscape stuff where you want a good deal of near and far objects in focus can be had at f8-11, I'm not sure I ever understood surpassing that aperture ever...



.

The DFA *50 is excellent wide open. That's all I mean. If you want edge sharpness at f1.4, there aren't any other lenses in the catalog that probably match it.

As far as why lenses go to f22, I don't think there is any negative to have f stops you don't use very often, but occasionally, you want to drag your shutter and don't have an ND filter with you and you just stop down, even though you lose sharpness. I know that I've done that before and even though there is some softness, depending on the image it isn't the end of the world.

[Luppenko]

I recently bought the DA 20-40, but I haven't done any testcharts. It's my first modern lens (apart from the 18-55), before I only used the old primes M28/2,8 and M50/1,4. The best thing, next to the awesome colours and clarity of the DA 20-40, is the limited zoom-range. In terms of how I take pictures, it feels like having three primes on my camera.

[normhead]

I also like looking at graphs. What the charts that get published by manufacturers mean, I have no idea. Not only that, most manufacturers charts are derived from the design specs, not from actual shooting. Actual shooting brings sample variation into play. Whether that's here or there is another whole debate.

The one thing I'd mention....

The better the lens, the lower the ƒ-stop at which it achieves maximum resolution.

Looking at the DFA 50 1.4
you can see it's sharpest from ƒ2 to ƒ4. That's a hallmark of an exceptional lens.

Most consumer lenses are sharpest at ƒ5.6. But ƒ8 is still quite good if you need more depth of field. Less so on APS-c than on FF, the the diffraction limit is between ƒ5.6 and ƒ8 for both but the larger the format the lower the diffraction limit, which of course ends with Ansel Adams shooting ƒ64 of 8x10 film. So you might want to be a little quicker to use ƒ8 on your K-1 than you were on your APS-c sensor.

To me the whole argument for the DFA 50 1.4 is sharp edge to edge at wide apertures. Because landscapes need wide depth of filed, I'm not sure that even matters to me.

DA* 55 chart


For me, sharp edge to edge from ƒ2 to ƒ4 but falling off on the edges from 5.6 on would be of limited use. AT 5.6 to ƒ8. the DA could be the better lens, if you're looking edge to edge. But not necessarily. You still have to do some image comparisons to be sure.

I use the graphs to tell me what I should be looking for when I test.

Why would I need further testing? From the graphs you know what the numbers are, but you don't know how that correlates to real world images. I don't think any of us are so smart we know how images will compare based on a bunch of numbers. IN the above examples I know the DA* is more even edge to edge at ƒ8, but because they are on different formats, I really know nothing. I don't know if the DA*s edges fall off considerably after the APS-c circle. Where the DFA 50 looks weaker on the edge, for all I know it's still stronger on the edge than the DA* is in the middle.

The first thing I'm looking at comparing these two would be how many images do I take that would take advantage of the DFA's strength. Not including wildlife, of my 900 keepers this year, 15 would qualify as f2-ƒ4 images. For me the DFA would be a waste of money. It would be rare I could actually see a difference. So what I would be looking at on the charts would be ƒ5.6 and ƒ8, and those are the places I would look at comparison images.

To me the charts are a good way to narrow down your search and focus on those situations you are actually buying the lens for.

Looking at the DA 20-40 chart, its a traditional Pentax design. Soft edges where the depth of field is narrow, meaning large areas of the image will be out of focus and soft edges won't be as critical. Good edge to edge from ƒ5.6 to ƒ11 for landscape. A very functional lens for the way I shoot.

The DA 20-40 is definitely in the spirit of the FA 31 ltd


When I see a lens like the DFA* 50 1.4 I wonder "who shoots like that?" Who shoots from ƒ2 to ƒ4 and needs edge to edge sharpness? Do people shoot walls in dim light (and why?)

But everyone is not me. Having better edge to edge would effect 1.6% of my images. But I can imagine that other people might shoot more in that range. That's what you should be buying a DFA 50 for. If not, and you're like me, you're buying capacity that you won't use.

Of course people can buy lenses just because they can. No other reason is needed. I do wonder if anyone actually makes use of ƒ2-ƒ4 on their DFA 50 1.4, or if it's just a prestige item. Personally I look at the DFA and DA* images and decided the DFA* wouldn't help many of my images.

And I wonder, who are the folks who benefit from the DFA and what are they shooting that makes use of it's qualities.

And I seriously like graphs , so I can quickly look for numerical differences that highlight what exactly I should look at when comparing the images taken by the lenses. But you still have to compare images to see exactly what the numbers mean in terms of how each lens design was executed. I look at the graphs as the key to intelligent evaluation, letting you focus your attention at lens differences, and not going through the areas where they are pretty much the same. Every tool used to filter information helps.

Especially for those of us with short attention spans.
We have to get to the important part quickly or we might forget what we're doing.

[UncleVanya]

@normhead you are a man after my own heart! The trend towards the heavy ultra corrected edge to edge sharp fast lenses befuddles me.

My 20-40 is a fine lens and I'm happy to own it.

[normhead]

@normhead you are a man after my own heart! The trend towards the heavy ultra corrected edge to edge sharp fast lenses befuddles me.

My 20-40 is a fine lens and I'm happy to own it.

It must be how old we are or something.

This guy makes excellent use of sub ƒ4 images and does really good work.
Dustin Abbott | Flickr

It would be interesting to see the page of someone shooting the DFA * 50 1.4 doing similar work.

[microlight]

My 20-40 is also a very fine lens; suits my style well.

[c.a.m]

I purchased a Pentax DA 20-40mm in April 2019, and have used it several times in the field. So far, I am very pleased with the lens.

Prior to deciding on this lens, I did a fairly extensive research of online comments and reviews. Here is a copy of my summary notes.

Several sites had reviewed the lens and provided quantitative details of its sharpness/resolution (here, I use the two terms interchangeably). In the information given below, the qualitative terms for the ratings are those used in each respective lens review, except where noted.


1. ePhotoZine

ePhotoZine uses four ratings for sharpness: poor, fair, good, and excellent. In my analysis of their charts, I introduced a fifth rating of 'very good' which I placed between 'good' and 'excellent'.

Center: excellent sharpness 20-40mm at f/2.8 to f/11 at all focal lengths (FL).

Edge:
Good f/2.8-f/4, all FL
Very good or better f/5.6 to f/16, all FL

For very good sharpness across the frame, stop down to at least f/5.6 at all FL.


2. erPhotoreview

Center: very good to excellent sharpness f/2.8 to f/16, all FL
Edge: Very good to excellent, f/4 to f/11, all FL

Wide open: 'good' at all FL


3. Lenstip

Lenstip criteria: 'Decent' sharpness when MTF50 greater than 35 lpmm; 'Good' quality when MTF50 >= 40

Center:
20mm. Softer at f/2.8 to f/4 and f/16 to f/22, but 'good' f/2.8 to ~f/13
30mm. Good f/3.5 to ~f/13
40mm. Good f/4 to ~ f/13.

Edge. 20-40mm: stop down to at least f/5.6. Soft at f/2.8-f/4 and beyond ~f/13.


4. Imaging Resource

• Summary: "Sharp lens"
• Wide open at 20mm: noticeable corner softness, but sharp center
• Stop down to f/5.6-f/11 for "excellent" sharpness
• 24mm: the 'weak spot'. More softness wide open at f/3.5
• 30mm: some corner softness wide open. Stop down to f/5.6.
• 34-40mm. Sharp images with ever-lessening corner softness when wide open. Sharp at f/4-f/16 across the frame.

Best across-frame sharpness:
20mm: f/8 (less than 1.5 blur units across the frame)
24mm: f/8 (less than 2 blur units)
34mm: f/5.6 (1-1.5 blur units)

For excellent sharpness, stop down to f/5.6 20-30mm and f/4 30-40mm. At 40mm, f/4 is wide open.

I also compared the DA 20-40mm Limited to the Pentax DA 18-135mm. In summary, the 20-40mm Limited has a 1-stop sharpness advantage across the common focal length range, except at ~24mm where the 18-135mm has its peak overall sharpness.

- Craig

[RGlasel]

f2.8 should not be shy'd away from using at 20mm as long as it's centre sharpness you only care about.

I'm still trying to understand how centre resolution at 20mm and 30mm actually decreases as aperture narrows from wide open. Stopping down to f4 and f5.6 actually decreases resolution?

[mikesbike]

Test charts while useful, I trust my own eyes more than charts. Even more than the evaluations given by the testers. Looking at the test photo by Imaging Resource both at wide open aperture and at mid-aperture f/8, examining the image at is central area and also at the near-edge areas will reveal enough to draw your own conclusions. When you click on the image you get a huge blowup you can navigate around to take a look at different parts of the scene.

Also keep in mind you are examining a blowup larger than ordinarily seen in actual circumstances.

My conclusion in practice verifies the test scene and then some. That the DA 20-40mm Limited is an exceptionally fine zoom lens. Not the only option, but for a compact (emphasis), well-built, faster-than-average, well-performing zoom lens, it is right up there in terms of that combination.