[Class A]

Here are some reasons why designed for film lenses may not work to their full potential on digital sensors...

This would more adequately read as "Here are some reasons why some designed for film lenses may not work to their full potential on some digital sensors".

The sensors in the Sony NEX-7 and NEX-5 are good examples for sensors that are more (NEX-7) or less (NEX-5) susceptible to this problem. I understand that it just depends on the design of the microlenses.

If you look at the NEX-7 vs NEX-5 discussions, you'll furthermore notice the only older wide-angle lenses cause grief on the NEX-7.

I understand that the incident angles created by "normal"/"standard" or higher focal length lenses are not a challenge for modern sensors.

There is another difference between film and digital sensors which is the depth of the imager. It is conceivable that lenses designed for film had their CAs tweaked in such a way that they matched the order of colour sensitive layers in film. A digital sensor using a Bayer-array would then not benefit from such an optimisation (whereas a Foveon-type sensor could). However, if I remember the respective quantitative analysis correctly, the effect of this potential issue would be noticeable but not be a show stopper.

[pinholecam]

How come Klaus/Photozone always gets lemons from his Pentax lens donors? (are they really Canikon saboteurs instead )


The emphasis on numbers (esp. MTF) on lens review sites has resulted in a new generation of enthusiasts who only want/know sharpness.
The sad effect of this is that this will push lens design towards meeting these numbers regardless of more aesthetic aspects.
If there is any chance to show off other aspects of a lens in a lens review other than the numbers, I'd think a series of shots that would include :

1. Studio setup (45deg light and/or side light); Bkgnd light
2. Small object; near and further focus
3. Larger object (close to a human head and shoulders); Near and further focus
4. Object close to half body portrait; Near and further focus
All at apertures from wide open to probably f8.

Probably too tedious to do.

Don't forget, Klaus works hard on his reviews, but many Pentax users have much more experience with these lenses than he does. Characteristics like image sharpness near the outside of the frame may matter for some photos, but usually goes unnoticed. At the same time, the characteristics mentioned above will be noticed, even if unconsciously by some viewers (which is usually a good thing). And I dare say that in most situations where edge sharpness is critical one would stop down anyway (such as when using a copy stand or taking most landscape pictures).


.

Well said.

[selar]

That'd be interesting, lenses designed by computers, assembled by robots, and tested and ranked on computers.Or is that already reality?

[lytrytyr]

The emphasis on numbers (esp. MTF) on lens review sites has resulted in a new generation of enthusiasts who only want/know sharpness.

Pixel-peeping may be more to blame.
If anything, the MTF data given in lens review sites
is less detailed than the old-format MTF curves.

[Asha]

I thought purple fringing was due to sensor blooming, ie, when a "pixel" is oversaturated, some charge spills over into the next "pixel". That can occur anywhere from center to edge of frame.

Chromatic aberration, on the other hand, is definitely something that happens around the edge of the frame. I don't know about the sensor contribution, however, dispersion in glass and plastic most certainly will cause CA from the optic itself. CA can be minimized by optimizing glass selection within the lens design.

ETA: I should have more correctly stated that CA is most emphasized at the edge of the frame, not that it ONLY occurs there.

[selar]

I wonder why digital sensors are not made concave, would get rid of a lot of problems!

[Rondec]

How come Klaus/Photozone always gets lemons from his Pentax lens donors? (are they really Canikon saboteurs instead )


The emphasis on numbers (esp. MTF) on lens review sites has resulted in a new generation of enthusiasts who only want/know sharpness.
The sad effect of this is that this will push lens design towards meeting these numbers regardless of more aesthetic aspects.
If there is any chance to show off other aspects of a lens in a lens review other than the numbers, I'd think a series of shots that would include :

1. Studio setup (45deg light and/or side light); Bkgnd light
2. Small object; near and further focus
3. Larger object (close to a human head and shoulders); Near and further focus
4. Object close to half body portrait; Near and further focus
All at apertures from wide open to probably f8.

Probably too tedious to do.
.

Klaus "borrows" lenses from actual users. I know, for instance, that his 60-250 test was taken from a lens owned by Falk Lumo who posts here under the name Falconeye. Clearly these people are happy with their lenses, or they would have returned them/had them repaired. I guess it goes to show that these sort of flaws, even if present, probably don't effect every day shooting that much.

At the same time, the Imatest charts that Klaus uses are pretty standard and should give consistent results for the same sensor. To me, the most useful thing about the tests is seeing where the lens "peaks." So, the FA 31 peaks at f2.8, although it remains strong to f8. With this copy of the FA 43, the center is actually pretty strong throughout, but the borders don't seem to get sharp till f4. Klaus usually does do a walk around with a new lens as well and posts photos from that.

[lytrytyr]

I wonder why digital sensors are not made concave, would get rid of a lot of problems!

Because they're made on standard chips, which are flat.

However, the Kepler Space Observatory sensor
http://upload.wikimedia.org/wikipedia/en/thumb/2/2a/Keplerspacecraft-FocalPl...cutout.svg.png
is made up of an array of small flat sensors
arranged in a large-scale concave pattern.

Back in the 1950s,
Kodak "medium format" box cameras with simple meniscus lenses
used to curve the film track to match the Petzval field curvature of the lens,
at least in the horizontal direction.

[Blue]

How come Klaus/Photozone always gets lemons from his Pentax lens donors? (are they really Canikon saboteurs instead )


The emphasis on numbers (esp. MTF) on lens review sites has resulted in a new generation of enthusiasts who only want/know sharpness.
The sad effect of this is that this will push lens design towards meeting these numbers regardless of more aesthetic aspects.
If there is any chance to show off other aspects of a lens in a lens review other than the numbers, I'd think a series of shots that would include :

1. Studio setup (45deg light and/or side light); Bkgnd light
2. Small object; near and further focus
3. Larger object (close to a human head and shoulders); Near and further focus
4. Object close to half body portrait; Near and further focus
All at apertures from wide open to probably f8.

Probably too tedious to do.

. . .

As posted earlier, Hirakawa Jun didn't make the MTF target the primary concern although he made it an important one. It may be interesting to find out how he tested lenses of these types.

I thought purple fringing was due to sensor blooming, ie, when a "pixel" is oversaturated, some charge spills over into the next "pixel". That can occur anywhere from center to edge of frame.

Chromatic aberration, on the other hand, is definitely something that happens around the edge of the frame. I don't know about the sensor contribution, however, dispersion in glass and plastic most certainly will cause CA from the optic itself. CA can be minimized by optimizing glass selection within the lens design.

ETA: I should have more correctly stated that CA is most emphasized at the edge of the frame, not that it ONLY occurs there.

Although not as bad, purple fringing can occur on film.

[ogl]

I'm from Germany. I do have a silver version of the FA43.
Klaus already got back to me. Seems like he has already found a lens to redo the test.
Let's see whether the results will be different this time.

Good news.

[bossa]

Because they're made on standard chips, which are flat.

However, the Kepler Space Observatory sensor
http://upload.wikimedia.org/wikipedia/en/thumb/2/2a/Keplerspacecraft-FocalPl...cutout.svg.png
is made up of an array of small flat sensors
arranged in a large-scale concave pattern.

Back in the 1950s,
Kodak "medium format" box cameras with simple meniscus lenses
used to curve the film track to match the Petzval field curvature of the lens,
at least in the horizontal direction.

What a great idea!

[selar]

Harikawa Jun

Previously in this thread, Kirakawi Jun, then Kirakawa Jun

Poor Mr. Hirakawa Jun, in addition to being laid off, seeing his lens designs criticized by Photozone, must suffer having his name misspelt several times in this thread.

[Blue]

Previously in this thread, Kirakawi Jun, then Kirakawa Jun

Poor Mr. Hirakawa Jun, in addition to being laid off, seeing his lens designs criticized by Photozone, must suffer having his name misspelt several times in this thread.

Wow. A couple of transposed letters is the best you can do. Actually, I am not sure if it is Mr. or Dr. However, I doubt you knew who he was before this thread. How soon before you start calling people Rip Van Winkle again?

[selar]

I am pretty sure you dont like people misspelling your name, I have already provided the correct spelling earlier in the thread, why do you persist in misspelling the man's name?

[Blue]

I am pretty sure you dont like people misspelling your name, I have already provided the correct spelling earlier in the thread, why do you persist in misspelling the man's name?

I said I transposed some letters.


平川6月