Thanks, Michael. Do you think you could post those ZM 35/2 crops a little larger? I'm having a tough time seeing the differences. I would expect the ZM 35/2 to be a little better on the GXR, for sure. Granted, I believe that it has the longest exit pupil of all of the wide ZM lenses, so it should be the best of the ZM bunch on NEX, too.
wfrank wrote:
Is there a reason why the NEX images are softer? I assume it's RAW similar treatment or?
The GXR has no AA filter, so NEX cameras require more sharpening to get as much detail. Generally, proper sharpening routines can level the playing field between aa and aa-less cameras quite a bit. However, the GXR's lack of an AA filter also creates less astigmatism with these rangefinder lenses and their high angle of incidence, so loosing the AA filter gives the GXR and advantage in a couple of ways. The disadvantage is potential moire, which may or may not be an issue, depending on what you shoot.
I've owned cameras with and without AA filters, and I'd often get frustrated when I was shooting fashion with AA-less cameras, because of moire in clothing patterns. However, as pixel counts go up, this becomes less of an issue, and I think the benefits of no AA filter outweigh the negatives, in terms of these mirrorless cameras.
Douglas: fixed the images, the two crop images are now clickable for full sized 100% crops.
wfrank: They do seem awfully soft on the NEX don't they? The examples are straight out of Lightroom, no sharpening other than LR defaults and I've applied no profile, auto or otherwise. The centre crop I made no WB adjustment but iirc I think I remembered to even up the temperature for the corner crop.
Thanks, Michael. The GXR certainly looks better. The tough thing about comparing cameras with and without AA filters is that it's an unfair playing ground unless you optimally sharpen each first. So, it makes it tough to know whether the differences in detail between the two cameras is a usual matter of aa vs. no-aa, or whether it's because the aa filter of the 5N is creating astigmatism with that particular lens. Either way, the GXR certainly looks better in this test, and I'm not under the impression that infinity wasn't reached on the 5N, but who knows?
michaelwatkins wrote:
This is definitely not a real world test, or an other world test, or even a test. Just an observation. Click images for full size 100% crops:
Call it what you want but that's really impressive detail on the GXR.
I'm very curious how the Zeiss ZM 21's, the 2.8 and the slower 4.5, perform on the GXR.
Definitely sharpening can improve the overall softness of the NEX image. I had the sharpening slider ~ 3/4s of the way over in LR and made some other adjustments, working quickly.
Pulling up the sharpness slightly on the GXR image addresses some mild softness in close where grass and tree branches become better defined (not shown in these crops), but it doesn't take much and if noise was a concern one could just leave well enough alone in that area at least, so maybe GXR has an advantage there beyond just its ability to resolve detail.
As I said earlier, I'm merely collecting my own observations here. It would be hard for me to recommend one over the other to anyone else, I'm just glad the one I kept fits me well. They are both impressive cameras for their size and the NEX all the more so for its cost.
Yeah, the GXR still looks much better. One thing to try is increase your detail slider all the way to 100 in LR3, which equals deconvolution sharpening and is good for cameras with an AA filter. Of course, you'll probably have to decrease the overall amount if you do this.
douglasf13 wrote:
Thanks, Michael. Do you think you could post those ZM 35/2 crops a little larger? I'm having a tough time seeing the differences. I would expect the ZM 35/2 to be a little better on the GXR, for sure. Granted, I believe that it has the longest exit pupil of all of the wide ZM lenses, so it should be the best of the ZM bunch on NEX, too.
Hi, I am curious.
About the exit pupil, is there a way to guess the magnitude (or the number) for the exit pupil distance from the lens technical specification for ZM lenses(lens diagram and etc)? I want to know if there is a way to figure this out.
Or, did you guess it from the ray angle column from here? http://diglloyd.com/articles/LeicaM/images-ZM/ZeissZM-ray_angle.gif
Do you know if the exit pupil distance is influenced by how far the rear element is from the sensor? I could not figure this out from my quick search on the internet.
I recently acquired summarit-m 35mm because I jumped to conclusion that the rear element of the summarit being so recessed towards the front of the lens is a good thing. (I also got a good deal on this.) This was why I chose the summarit over the zm 35s.
The exit pupil is the diameter of the virtual aperture when looking at it though the rear of the lens, and it is where the image is projected from. Here is a graph drawn up by Helen B. on the photo forum:
I see, so it's not exactly where the aperture blades are located, but how far it looks looking through the rear element... That would be pretty tough to guess from the specs.
Also, I noticed for the ZM 35 f2 and the ZM21 f2.8, there is one possible position to put aperture blades. But for the CV 15 and 21, there are 2 places for aperture blades may be located. http://www.cosina.co.jp/seihin/voigt/english/s-wide-e.html
I guess you kinda have to buy the lens and test it out.
Tariq: I did try out the ZM21/2.8 briefly on the GXR when I got the ZM18- it seems to work great.
I haven't had the opportunity to try the 21/4.5, I think it might be stretching it a bit. It seems a hard lens for digital.
Michaels findings echo mine, I wasn't particularly meticulous about it so I did not want to post any unsupported findings, it was a hunch and the NEX5N samples left me thinking I would have to reshoot to see what the "fault" was. But, the NEX5N works fine, the GXR does a little better in pulling out detail. I think Douglas nailed it in his comments. Both cameras are great little packages.
Thanks for those shots Michael - i'm playing with my 5N and a rented GXR M at the moment.
I think i'm seeing similar things with the GXR giving better results on the image periphery / tricky situations. I think my Nokton is showing less field curvature too, but need to test that properly as a very small focus change can skew that on the 5N.
I've been happy with my CV 15mm on the 5N, but it seems to work better on the GXR - less purple fringing and my copy has a little asymmetry which appears to be less noticeable on the GXR. Who knows?
I'll start a thread in a few days with my findings, but i'm thinking I may well be able to deal with the (perfectly adequate) GXR EVF if this A16 module materializes...
hoping for decent weather tomorrow for better testers as the light was changing from sun to rain every 5 minutes here today...
Sosua wrote:
hoping for decent weather tomorrow for better testers as the light was changing from sun to rain every 5 minutes here today...
What? Are you in Stockholm? Nah, just kidding. In Sweden we don't get the sun between the showers at the moment. NZ must be getting warm by now, at least the upper half of the north island.
Tariq: no worries. I was eyeballing the 21/2.8 for speed and had it as my no1 alternative if the crop had been 1.3X or less... On the GXR corners looked great, but I only snapped and chimped a few casual shots outside the store. I deliberately framed a parked car and there was no corner deterioration at wide open on APS-C, and no smearing on the GXR.
michaelwatkins wrote:
Definitely sharpening can improve the overall softness of the NEX image. I had the sharpening slider ~ 3/4s of the way over in LR and made some other adjustments, working quickly.
For better sharpening in LR, the first thing I'd try is to pull the detail slider to 100. This sharpens the finer structures better than just increasing the sharpening strength.
By the way, your first crops at the top of this page say ZM35/2.8 while you say it's the ZM25/2.8.
AhamB, I noticed the labeling error late last night myself. The centre crops should be ZM25/2.8, the edge were ZM35/2 but when I have a moment I'll double check and also put together for my own curiosity the opposite set of crops.
Following the advice you and Douglas offered, I did revisit the sharpening and noted a little improvement with detail maxed out although sharpening still needed to be pushed hard to get anywhere close to the GXR Mount A12 result.
All in all I'm very impressed with that the Mount A12 can do with these lenses and look forward to what the Mount A16 might show us when it shows up next year.
Speculation / wishful thinking mode on:
Can we engage in wild speculation and wonder aloud if the 12MP anti-alias filter-free module can keep pace, at low ISO sensitivity anyway, with the 16MP sensor in the NEX-5N, will a anti-alias free 16MP based sensor in the Mount A16 give the 24MP sensor in the NEX-7 a good challenge as well?
Probably the differences will be closer, but at the same time if the Mount A16 inherits the good behaviour towards RF lenses the A12 generally seems to show, it should be quite a winner for Ricoh and RF lens users. Now if only they'd update a few other parts of the system...
Hi, Michael. Seeing as how the nex-7 looks to be a step back from the 5N with rangefinder lenses, I wouldn't be surprised if the A12 keeps pace with the nex-7, let alone the A16.
BTW, I ordered another ZM 35/2 for my 5N, and I'll try to get around to posting some corner crops, just so we can make sure your lens reached infinity. I'm assuming it did, but it could be good to have a second sample.
kosmoskatten wrote:
Tariq: I did try out the ZM21/2.8 briefly on the GXR when I got the ZM18- it seems to work great.
I haven't had the opportunity to try the 21/4.5, I think it might be stretching it a bit. It seems a hard lens for digital.
Michaels findings echo mine, I wasn't particularly meticulous about it so I did not want to post any unsupported findings, it was a hunch and the NEX5N samples left me thinking I would have to reshoot to see what the "fault" was. But, the NEX5N works fine, the GXR does a little better in pulling out detail. I think Douglas nailed it in his comments. Both cameras are great little packages.
On pg. 2 of this thread here: https://www.fredmiranda.com/forum/topic/1043757/1#10065426 samples from "an antique" 21 f/4.5 Biogon were posted that look great... and I would assume that since that is may be the original version and would a much more symmetrical design than the current Zeiss version, it would be even more demanding on digital. Over on one of the NEX threads, I think the 'best wide angle for 5N' someone posted some ZM21 f/4.5 images that looked pretty good in terms of corner sharpness. Definitely much better than the ZM21 f/2.8. Edit: Henrik beat me to it with the links above.
michaelwatkins wrote:
Speculation / wishful thinking mode on:
Can we engage in wild speculation and wonder aloud if the 12MP anti-alias filter-free module can keep pace, at low ISO sensitivity anyway, with the 16MP sensor in the NEX-5N, will a anti-alias free 16MP based sensor in the Mount A16 give the 24MP sensor in the NEX-7 a good challenge as well?
Probably the differences will be closer, but at the same time if the Mount A16 inherits the good behaviour towards RF lenses the A12 generally seems to show, it should be quite a winner for Ricoh and RF lens users. Now if only they'd update a few other parts of the system...
Michael, I think the resolution difference between the rumored AA-less A16 vs. the NEX 7 will be pretty close. The biggest difference should be edge/corner resolution in favor of the A16. If it continues and/or improves on what we're seeing with the A12, I suspect a significant number of alt users considering the NEX-7, yet disappointed with the image quality vs. the 5N, will instead opt for the GXR. While the NEX is a slicker overall experience, I think ultimately image quality is the primary concern in these circles, or at least image quality that is as faithful to the capability of a given lens as possible. Therefore, despite some compromises with the GXR, such as the lower rez and external EVF, possibly slower operational speed that may be compounded by the larger files created by the 16MP sensor, IMO it will be the preferred choice.
The GXR is where I'm definitely leaning now thanks in part to your sample images posted above. While they were done at different times, I feel the edge crop is quite indicative of what I also saw with some of my strongly symmetrical lenses on the 5N (as can be seen on my website). It would have been interesting to see the ZM35 comparison crops also at wider apertures, such as f/4, because I feel it would show an even greater difference favoring the GXR. My experience was that by f/8 much of the edge/corner smearing seen in my ZMs on the 5N was minimized and therefore it's possible we're not seeing the full degree of difference between the 5N and GXR based solely on f/8 crops. I would be very interested in whether the A12 retains good edge/corner sharpness with the ZM35 f/2 at wider apertures.
NZ must be getting warm by now, at least the upper half of the north island. 