[normhead]

2. Imatest: If the results of the ZEISS Milvus were impressive (50.1 lp/mm), then the HD PENTAX-D FA ★ 50mm F1.4 SDM AW they are generally stunning today (2018). The maximum result on the open aperture is 69.4 lp / mm!

OK, how are you seeing the Milvus is better, according to these numbers the DFA out resolves it by about 40%.

Especially since the DFA is better in the red channel. The red channel is definitely the most difficult to resolve, The top end for blue channel is around 100 MP for APS-c. Yellow green is middle of the pack. IN the red spectrum, if memory serves me well, red spectrum can only theoretically produce around 40mp on FF or 16 MP on APS_c.. So basically, you can cheat to make the lens have higher test chart resolution by making it more sensitive to blue and fudging the red. which is what from this analysis what the Milvus has done. That however does not produce a more a pleasing image. It just works better on a test chart.

If memory serves me well, theoretically,the red channel is maxed out on a 16 MP APS_c or 36 MP FF camera, (Which explains in part the need for an FF sensor to get to the 51 MP and the disadvantage to going to 100 MP on FF. 100 MP on FF if the theory is correct is only capturing at most 60 MP resolution in the red channel, while may capture 100 in the blue and possibly the yellow/green. You can go a long way because most things are not uniform red, there may be enough green or blue detail in them to provide the illusion of sharpness in the red channel, and that works until it doesn't.

In short, a focus on lens clarity in the red channel is probably the most important thing you can do to create the impression of a sharp lens, in real images where as a lens that emphasizes the blue channel will produce better test results, by reducing the impact of those muddy reds with the really wide wave length that are much more likely to cross pixel borders than those skinny little blue ones.

This also relates back to "lenses for the way people take pictures, not for the test charts" Made for the test charts could mean the lenses were made blue spectrum heavy to increase test scores, where as a more natural looking lens optimized for reds, might score slightly lower. So to me, just the fact it has both a higher lw/ph plus better reds, makes it even more of an accomplishment. Higher test scores and better reds.. how do you beat that? IN terms of "the way people take pictures" better reds is already an improvement, even without the higher test scores.

It also explains my feelings about all ZeIs lenses. They look sharp, but the colours don't look right to me and there's something un-natural about the look of the images. I'm guessing they all sacrifice sharpness in the red channel to increase lw/ph in the blue channel, for better test scores.

It makes me wonder coupled with kenspo's comment in another thread if Pentax is the only company that has noticed the way to increase real world resolution (not test chart resolution) is to have the best results in the red spectrum.

The best test chart resolution would be produced by blocking yellow, blue , green with the colour cast of your lens and just exposing for blue, but it would be a crappy picture in terms of realistic appearance.

---

Last edited by normhead; 11-02-2018 at 08:46 AM.

[awscreo]

Precision has a history of communications issues. Their work is typically better than their talking. Sigh. I wish they understood how it negatively affect customer satisfaction.

I think they low budget is really really evident with the customer support. When I called, the guys said they have 2 (!!!) people there answering calls. Plus I believe they closed their support centre as well? Feels like they're just sitting in a room somewhere.

[clickclick]

Ha ha, I just couldn't stay away from the first truly high-end Pentax lens in a long while.

Yeah, this lens is a "when" not an "if" for me. It really looks to be beutiful from all angles from construction to performance. Can't wait for the 85 and others that are in the works. We have some really nice things to look forward to.

[Darkover]

Yeah, this lens is a "when" not an "if" for me.

Yes, I thought this lens it to big, heavy, to expensive and only a 50mm. And then tested it at photokina for 3min, sigh. And now....

http://www.criseis.ruhr.de/bilder/Blatt.jpg 1:1 section from this evening...

Now I will take picture with 50mm until xmas. :-)

---------- Post added 03-11-18 at 18:27 ----------

Oh..and this is a first indoor picture with the DFA50mm 1.4:

http://www.criseis.ruhr.de/bilder/BuchUndTee.jpg

Olaf

[Medex]

OK, how are you seeing the Milvus is better, according to these numbers the DFA out resolves it by about 40%.

Especially since the DFA is better in the red channel. The red channel is definitely the most difficult to resolve, The top end for blue channel is around 100 MP for APS-c. Yellow green is middle of the pack. IN the red spectrum, if memory serves me well, red spectrum can only theoretically produce around 40mp on FF or 16 MP on APS_c.. So basically, you can cheat to make the lens have higher test chart resolution by making it more sensitive to blue and fudging the red. which is what from this analysis what the Milvus has done. That however does not produce a more a pleasing image. It just works better on a test chart.

If memory serves me well, theoretically,the red channel is maxed out on a 16 MP APS_c or 36 MP FF camera, (Which explains in part the need for an FF sensor to get to the 51 MP and the disadvantage to going to 100 MP on FF. 100 MP on FF if the theory is correct is only capturing at most 60 MP resolution in the red channel, while may capture 100 in the blue and possibly the yellow/green. You can go a long way because most things are not uniform red, there may be enough green or blue detail in them to provide the illusion of sharpness in the red channel, and that works until it doesn't.

In short, a focus on lens clarity in the red channel is probably the most important thing you can do to create the impression of a sharp lens, in real images where as a lens that emphasizes the blue channel will produce better test results, by reducing the impact of those muddy reds with the really wide wave length that are much more likely to cross pixel borders than those skinny little blue ones.

This also relates back to "lenses for the way people take pictures, not for the test charts" Made for the test charts could mean the lenses were made blue spectrum heavy to increase test scores, where as a more natural looking lens optimized for reds, might score slightly lower. So to me, just the fact it has both a higher lw/ph plus better reds, makes it even more of an accomplishment. Higher test scores and better reds.. how do you beat that? IN terms of "the way people take pictures" better reds is already an improvement, even without the higher test scores.

It also explains my feelings about all ZeIs lenses. They look sharp, but the colours don't look right to me and there's something un-natural about the look of the images. I'm guessing they all sacrifice sharpness in the red channel to increase lw/ph in the blue channel, for better test scores.

It makes me wonder coupled with kenspo's comment in another thread if Pentax is the only company that has noticed the way to increase real world resolution (not test chart resolution) is to have the best results in the red spectrum.

The best test chart resolution would be produced by blocking yellow, blue , green with the colour cast of your lens and just exposing for blue, but it would be a crappy picture in terms of realistic appearance.

I just translated author's main conclusions.
Your thoughts are very interesting... about red color...

[normhead]

I just translated author's main conclusions.
Your thoughts are very interesting... about red color...

Thanks, it brings together years of observations, but is still complete speculation, in terms of not having anything definitive to suggest what I suspect is happening is what is really happening. Just a collection of "might be related" data.

[suncomp]

For those who are interested to look at the photos, Pentax is always on the right in the field shots. It is on the left in the studio shots. You can move the slider, so it's kind of neat. Some things he says:
1. Comparing Pentax with Zeiss - Pentax has fewer chromatic aberrations (Hello Photozone, or what was that widely-discussed test that showed green fence as presence of chromatic aberrations?)
2. Based on the field test he concludes that there is no significant difference in resolution between Pentax and Zeiss. To me though it looks like Pentax is both sharper and has better contrast...
3. Based on studio shots, he concludes that Zeiss resolution is "just under" 80 lp/mm, Pentax between 56 and 80, saying that his target does not allow for a more precise estimate
4. DFA outresolves the sensor.
5. Imatest allows one to evaluate the resolution of the lens if it exceeds resolution of the sensor
6. "If Zeiss resolution is impressive (50 lp/mm), that of Pentax (69.4 lp/mm) is truly incredible for today's technology!" - by the way, that's my translation, could be different from Google's
7. As Medex said, he concludes that in field tests resolution looks about the same, although "an experienced photographer" may notice that Zeiss has slightly higher resolution. Studio tests support that conclusion, but additional mathematical analysis suggests that Pentax could exceed Zeiss if it was used with a sensor with higher resolution.
8. of all tested lenses the closest competitor is Sigma art, but it is soft at the edges. Pentax performs better at the edges.
9. At the end of this discussion he talks about Star tracer... He must really like that function (he does say that only Pentax has it and if you were to use a different camera you'd have to use a wide angle lens and a high ISO).

Я - русский . Вы все правильно поняли !!! I'm Russian. All of you correctly understood!!!

[Sandy Hancock]

Are you sure it isn't field curvature? Check in with Sandy Hancock to see if he can shed light on this. I know he loves his 15-30.

I'm no expert - just a cashed-up gear freak My copy needed a trip to Japan to sort out some weirdness in the upper right hand quadrant, and yes, I'm now very happy with it. I can't wait to get it to Argentina/Antarctica - less than three weeks now....

After much discussion with Sigma Japan, they directly sent me a hand written letter along with the 4th copy of the lens, along with charts certifying its performance*.

* Which I'll admit to testing the lens on my own bench, and their test results were accurate.

^ This is the guy you need to PM.

[monochrome]

@normhead As I r call, Hoya ‘retired’* Jun Hirakawa because he refused to optimize new lens designs to score best on test charts. Rather he designed lenses to produce the best image. IOW he designed lenses for the real world, the way people photograph.

He went to Tamron, where (according to PM discussions with a European member) he eventually managed the Lens Design Department. Apparently he has announced his upcoming retirement from Tamron.

It seems his philosophy remains the guiding principle of the current Pentax team, at least as far as * lenses are concerned.

[awscreo]

I'm no expert - just a cashed-up gear freak My copy needed a trip to Japan to sort out some weirdness in the upper right hand quadrant, and yes, I'm now very happy with it. I can't wait to get it to Argentina/Antarctica - less than three weeks now....

My lens has been in Japan twice now I wonder if it's my mount, seems unlikely that I had 3 bad copies in a row, and it couldn't be fixed in Japan at the factory twice..

[Sandy Hancock]

Sample images?

[beholder3]

7. As Medex said, he concludes that in field tests resolution looks about the same, although "an experienced photographer" may notice that Zeiss has slightly higher resolution. Studio tests support that conclusion, but additional mathematical analysis suggests that Pentax could exceed Zeiss if it was used with a sensor with higher resolution.

So the summary there is that the Pentax DFA* 50 on 36 Mpx without pixelshift already resolves pretty much as good as the Zeiss Milvus 50 (which lacks any autofocus) on 50 Mpx.
Makes the Zeiss Milvus look a pretty bad choice actually in direct comparison.

[monochrome]

So the summary there is that the Pentax DFA* 50 on 36 Mpx without pixelshift already resolves pretty much as good as the Zeiss Milvus 50 (which lacks any autofocus) on 50 Mpx.
Makes the Zeiss Milvus look a pretty bad choice actually in direct comparison.

Regarding Point #7, which you quoted, didn’t the lead designer write on the lens Special Site they are designing the * lenses for future higher resolution sensors?

[IgorZ]

So the summary there is that the Pentax DFA* 50 on 36 Mpx without pixelshift already resolves pretty much as good as the Zeiss Milvus 50 (which lacks any autofocus) on 50 Mpx.
Makes the Zeiss Milvus look a pretty bad choice actually in direct comparison.

I think that's a good way to look at it

[amstel78]

The Pentax-D FA 50 is one of my favorite lenses. I use it more than my Pentax 24-70. It has a level of sharpness that I've just never seen before in a 50mm lens designed for an SLR or DSLR. The only problem I've ever had with it, and it's more me and the camera rather than the lens, is that it can be sometimes difficult achieving critical focus at wide open apertures. For instance, focusing on a model's eye, if the camera or the model moves even minutely, can result in the eyelashes being tack sharp with the pupils being slightly less so. It becomes evident on a 4k monitor looking at 100% pixel resolution, but for casual prints, not so much.

What surprises me even more however is the lens' resolving power at the edges wide open. It's just silly sharp at the top, bottom, left, and right edges of the frame. I'd love to see a 24mm or 35mm prime from Pentax with the same design ethos.