35/1.8 challenges 35/2

Revision said:

The flash specular reflection in the cat's eye shows obvious
chromatic aberration for both lenses.
While the f/2 has transversal (lateral) aberration, the f 1.8 has the
much harder to correct longitudinal aberration.

Click to expand...

It may be harder to remove, but it is also far less severe. Look at the 100% f/2 crops in the above post and in the post below dealing with bokeh. I would like to see someone post process the 35/2 shots to match the 35/1.8 shots, but I'm pretty certain it can't be done.

Pat Joh said:

So with the results of your tests do you think that the 35mm F1.8 is
a better lens than the 35mm F2 in all aspects (on a D300/D200 body)?
From what I read the only thing that the 35mm F2 does better is the
less barrel distortion is that about it?

Click to expand...

My tests lead me to conclude no reason to choose the 35/2 over the 35/1.8. That said, every difference I can detect save two are only relevant at or near 100% enlarged on my monitor. One is the 35/1.8 being better at f/2 than the 35/2; and the other is barrel distortion, and I'm not even sure that the barrel distortion doesn't add an intangible charm to the 35/1.8 images.

In addition to its lower price, the 35/1.8 has AFS which allows manual focusing and quick override of the AF without having to fiddle with the AF mode switch on the front of the camera. Based on those two considerations alone I would choose the 35/1.8 even if the 35/2 could match it at wider apertures -- which the 35/2 simply can't.

no text here

I'm glad you're measuring for T-Stop Values I think it is severly under reported in reviews.

I think the lens that has the largest T-Stop inconsistency is the 16-85mm VR, I had one for a while and it always seemed to under expose by 1/2 ev to 3/4 ev with heavy vignetting compared to other lenses at the same FL and F/stop , the fact that the lens is extremely slow as it is ( F/3.5 - F/5.6) makes the VRII much les effective, I almost feel that it should be labeled " F/4.5 - F/8 shoot in sunny daylight only"

Also how did you fnd the viewfinder and focusing accuracy under low light? My experience with lenses that have a lower T-Stop is that the viewfinder is not as brightand that they hunt in low light.

photonut2008 said:

In addition to the 35/2 having less barrel
distortion, another reason would be that it is about a third of
T-stop faster than the 35/1.8. To illustrate how this can have an
effect on exposure, I have included green channel histograms from
four exposures I made with the two lenses (two each).



As you can see, the difference sometimes changes metering. I forget
what aperture I used for those shots, but given the shutter speed it
was obviously a pretty fast one; anyway, the difference occurs
throughout the aperture range. Nonetheless, later on I will show why
I believe you are still ahead with the 35/1.8 lens at wider apertures.

Click to expand...

Jorge Barrantes said:

...how did you fnd the viewfinder and focusing accuracy under low
light? My experience with lenses that have a lower T-Stop is that the
viewfinder is not as brightand that they hunt in low light.

Click to expand...

I could not detect a difference in either viewfinder brightness or AF accuracy between the two lenses.

RAL said:

Does your copy of the 1.8 have more distortion in the the horizontal plane than in the vertical (with the camera in horizontal orientation)? It just sounds like something is possibly wrong if changing to vertical orientation would show less distortion. I realize that on the wide direction you get further from the center than on the short side, but that should just be dependent on how far away from center something was, not the camera orientation.

Click to expand...

I did quite a bit of testing of this the last couple of days and got consistent 1% differences along the vertical or horizontal axis from any given point.

photonut2008 said:

Seems to me that the lenses would be circular with the optical center being in the center, not rectangular with differing distortions depending on orientation. They would have to get the lens rotated just right in the barrel when they manufactured the lens in that case. I guess I'm not too old to learn new things everyday.

Click to expand...

The result could just be my copy, but it is a consistent result (which means it is symmetrical). This isn't unprecedented though, some lenses have wavy distortion patterns. I'm not saying this is a big deal, I actually don't think it is as far as just re-orientating the camera when the subject is centered is concerned, but if I compound the differences in distortion with the differences between the edge and the center then the issue becomes very field relevant:



Excuse the pole sticking out of my head; I used that pole, a tripod, re-orientated the camera using its L bracket, and rotated the ballhead (the one I have has a rotating base) to minimize changes in perspective and assure precise focus distance between the two shots -- all of which seems to have worked reasonably well since the EXIF data indicates that both shots were focused at precisely one meter. My quick measurement of these is that I have managed to distort myself by about 3%, which ends up being visually significant.

This was a controlled comparison, shooting on the fly from above or below and distortions will be further exacerbated (although issues caused merely by changes in perspective are equally problematic for both lenses I tested in this review). In comparing the two lenses I would say that if I had taken the time to do this test on the 35/2 it would have shown less distortion; based on the distortion numbers at Photozone it might have stretched me vertically or horizontally by 1% instead of 3%.

I'm kinda confused... please clarify.

Umm the bottom photo on the left the only thing I see that might be slightly noticable would be that your face/body seem to "come out of" the image a bit more. Would that be distorion that you are reffering to?

So left say if I were to take a photo of a bridge would it be very noticable; the 3% barrel distortion?

Your test case gives an unfair advantage to the vertical shot because the subject is in the center. In the horizontal orientation you placed the subject towards the corner where barrel distortion is more pronounced!

If you place the subject in the center in either orientation you should see no difference in distortion.

Cheers

Mike

Pat Joh said:

I'm kinda confused... please clarify.
Umm the bottom photo on the left the only thing I see that might be
slightly noticeable would be that your face/body seem to "come out of"
the image a bit more. Would that be distortion that you are referring
to?

Click to expand...

No that would not be the primary effect of distortion, what you are seeing is more a matter of composition. In some ways the barrel distortion increases this effect, but I doubt that if I had taken the same shot with the 35/2 that you would notice much difference between the two for that shot.

The distortion I'm referring to is twofold. First there is the issue of orientation of the camera, where there appears to be a 1% between a subject that is shot in one orientation and the other -- with the vertical orientation being the thinner one. Second, I'm illustrating where the existing barrel distortion in conjunction with the differences in orientation can be combined to further distort my face (as you can see when looking at me up close with the same magnification and the same amount of resizing).

Pat Joh said:

So left say if I were to take a photo of a bridge would it be very
noticeable; the 3% barrel distortion?

Click to expand...

1% would be noticeable to me; I even notice the less than 1% distortion from the 35/2. I don't think either of theses lenses would be a good choice for architecture; but the 35/1.8 would be slightly worse.

Mike Neary said:

Your test case gives an unfair advantage to the vertical shot because
the subject is in the center. In the horizontal orientation you
placed the subject towards the corner where barrel distortion is more
pronounced!

If you place the subject in the center in either orientation you
should see no difference in distortion.

Click to expand...

You would see a 1% difference; at least I do with my copy. Also, this illustrates a situation where the 35/1.8 would do worse than the 35/2, primarily because the 35/2 has less barrel distortion to begin with.

My guess (since I've now done more than enough testing on this issue) is that the 35/2 would show close to no difference if I had tested it too. One reason I didn't even bother though was that the first thing that would jump out at anyone looking at the two shot side by side in with that set up would be that the 35/2 is actually longer than the 35/1.8. Now I've covered the issue of slightly different focal lengths between these two lenses earlier in this thread, and the appearance of that factor here would just confuse the entire issue I was getting at, which is that you want to be careful in how you frame your subject with the 35/1.8 and that portrait orientation is better than landscape orientation with that lens.

photonut2008 said:

Mike Neary said:

Your test case gives an unfair advantage to the vertical shot because
the subject is in the center. In the horizontal orientation you
placed the subject towards the corner where barrel distortion is more
pronounced!

If you place the subject in the center in either orientation you
should see no difference in distortion.

Click to expand...

You would see a 1% difference; at least I do with my copy. Also,
this illustrates a situation where the 35/1.8 would do worse than the
35/2, primarily because the 35/2 has less barrel distortion to begin
with.

My guess (since I've now done more than enough testing on this issue)
is that the 35/2 would show close to no difference if I had tested it
too. One reason I didn't even bother though was that the first thing
that would jump out at anyone looking at the two shot side by side in
with that set up would be that the 35/2 is actually longer than the
35/1.8. Now I've covered the issue of slightly different focal
lengths between these two lenses earlier in this thread, and the
appearance of that factor here would just confuse the entire issue I
was getting at, which is that you want to be careful in how you frame
your subject with the 35/1.8 and that portrait orientation is better
than landscape orientation with that lens.

Click to expand...

Excuse me, but this simply makes no sense at all. The lens is round; the aperture is round; distortion will be the same at same distance from the center. That's just physics.

Look at the distortion chart at photozone:

http://www.photozone.de/nikon--nikkor-aps-c-lens-tests/422-nikkor_35_18g?start=1

Now pick out a square area around the center of the image. You will notice that the degree of distortion is the same no matter if you're looking at the vertical or horizontal line.

Cheers

Mike

Mike Neary said:

Mike Neary said:

I wrote:
You would see a 1% difference; at least I do with my copy. Also,
this illustrates a situation where the 35/1.8 would do worse than the
35/2, primarily because the 35/2 has less barrel distortion to begin
with.

Click to expand...

Click to expand...

Mike Neary said:

You wrote:
Excuse me, but this simply makes no sense at all. The lens is round;
the aperture is round; distortion will be the same at same distance
from the center. That's just physics.

Look at the distortion chart at photozone:

http://www.photozone.de/nikon--nikkor-aps-c-lens-tests/422-nikkor_35_18g?start=1

Now pick out a square area around the center of the image. You will
notice that the degree of distortion is the same no matter if you're
looking at the vertical or horizontal line.

Click to expand...

It appears to be mild "wavy" or "mustache" distortion; as I've already written, it's not a big deal but it is there. Just because Photozone doesn't report it doesn't mean it's not there; after all, it wouldn't be the first time they've missed this issue:

http://www.naturfotograf.com/lens_zoom_02.html#AFS18-200VR

http://www.photozone.de/nikon--nikkor-aps-c-lens-tests/242-nikkor-af-s-18-200mm-f35-56-g-if-ed-vr-ii-dx-review--test-report?start=1

Mike Neary said:

Your test case gives an unfair advantage to the vertical shot because
the subject is in the center. In the horizontal orientation you
placed the subject towards the corner where barrel distortion is more
pronounced!

If you place the subject in the center in either orientation you
should see no difference in distortion.

Click to expand...

Mike Neary is right about placing the subject (myself in this instance) in the center of the image circle to avoid any noticeable differences; however any off-centering of the subject begins to cause differential distortion, and at the borders of the image circle the differences become more evident.

In retrospect, I should have given that last example a different caption; and highlighted the overall importance of framing with orientation as a subset of that. It turns out that 1/3 of the effect came from the center versus edge framing of myself in the frame; but 2/3 of the difference still came from the orientation of the camera.

Mike Neary said:

Excuse me, but this simply makes no sense at all. The lens is round;
the aperture is round; distortion will be the same at same distance
from the center. That's just physics.

Click to expand...

This is all true except that the area recorded by the sensor is not square. MikeinIndy got to the point pretty well earlier in this thread about the location of the recorded area relative to the image circle, and below is an illustration of the what is physically happening:



In my previous reply to Mike I wrote that I thought the issue was the complexity of the distortion, but having looked closer at this issue it is obviously a characteristic in all barrel distorted lens. The issue is more significant with the 35/1.8 than with the 35/2 -- it is twice as significant and that is consistent with the actual barrel distortion numbers for these lenses.

Below are the differences between the long axis and the short axis taken from the center and from the edges of the 35/1.8:



I would say the difference in orientation at the center of the image circle is insignificant and within the margin of error, but at the borders it is large enough to be a factor in some limited number of compositions -- i.e., you should avoid shooting people in landscape orientation framed at the border of the image circle with the 35/1.8 lens. On the positive side, this effect can be used to make someone look a little taller if you frame them at the borders of the image circle with the camera in portrait orientation.

n/t

P Hartung said:

Thanks for your work here.

Click to expand...

You're welcome.

P Hartung said:

I hope it encourages folks to get the newer lens if they shoot DX.

Click to expand...

I doubt my contribution will be significant. While I think people would be remiss in not considering this lens for DX format, what they choose doesn't effect me.

P Hartung said:

My copy is brilliant in its bokeh rendering and overall contrast.

Click to expand...

So is mine.

P Hartung said:

Such a deal...

Click to expand...

Agreed.