rscheffler wrote:
About the 28 Cron on M9 and M240... It's a good question about how much of the vignetting character is a factor of the sensor and how much is profile related. I don't think I can state a definitive answer, but I can share my observations.
I shot a more controlled white surface test today, though IMO, still not ideal. Referring primarily to f/2 shooting, my feeling is:
M9's vignetting is slightly stronger and more 'top heavy' in that a horizontal composition will have stronger vignetting at the top left and top right corners than the bottom corners. This vignetting 'position' is consistent whether coded or uncoded. The M240's vignette in comparison, feels like it has been shifted up and is more centered, though the bottom 1/4 of the frame (all the way across) seems to have more mid tone darkness than the top 1/4 of the frame. The M240's vignetting position is also consistent whether coded or uncoded.
Of the two cameras, there seems to be a bigger difference in the coded corrections applied by the M240 than by the M9. Both pull back highlight values, compared to uncoded, but the M9 does so only slightly. The M240's correction is much more dramatic compared to shooting the lens uncoded. Uncoded, it's closer to that of the M9's vignetting character and also similar in terms of exposure/brightness range based on histogram values. Coded, and the image is more even with lighter vignetting, but as noted earlier, exposure seems to have been pulled back noticeably. I see this very obviously in the histogram and would guess it results in around a 1/2 stop highlight reduction. In fact, the histogram looks compressed at both ends, but more so on the highlight end.
Color shift on the M9 is a bigger problem, both uncoded and coded. Coded, the under correction results in some magenta on the left side and cyan/blue on the right side at f/2. Stopping down this color shift decreases but the bottom corners become increasingly more green, noticeable from f/4 on. Uncoded and the left side shifts magenta and right side cyan/blue, but past f/4 it's not terribly noticeable outside of scenes that require very neutral tones.
M240 color shift is much less. Uncoded it's really only noticeable wide open, with the right side being a more noticeable magenta shift compared to the left side's blue with some slight magenta. Beyond f/2.8 it's unlikely to be very noticeable. The M240's profile seems to be very effective at neutralizing color shift across the aperture range, with perhaps a hint of green/cyan along the very left edge that probably will rarely be noticeable in normal use. Shooting the lens coded or uncoded on the M240 results in no normally discernible color differences, such as WB or tint shift.
While I'm sure Leica will never do it, it would be cool to have an option for the M240 to shoot the lens coded and retain EXIF, but to disable vignetting correction. It would probably return some of the Cron's M9 magic to the M240 and avoid the terrible reduction of highlight values at wider apertures seen in the current profile....Show more →
very interesting and thanks for doing this Ron.
i'm quite surprised leica is choosing to pull back the highlights as part of lens corrections. that would definitely explain why many people think the lens isn't as magical on the m240. i think i would probably choose to shoot uncoded and use the lightroom plugin to correct color but not vignetting if i had this combo.
i would think that it would be feasible to make a firmware change allowing the user to specify whether they wanted full correction or just color correction.
sebboh wrote:
i would think that it would be feasible to make a firmware change allowing the user to specify whether they wanted full correction or just color correction.
I really wish Leica would do that and I have even requested on the L user forum that correction should include only color cast and not vignetting, or let the user choose as with the A7, but I understood from people close to Leica that it's not going to happen.
burningheart wrote:
zhangyue - Lovely Yosemite shots. Were you anywhere near the fires I heard about in the news last night?
Yeh, I was there, actually, it saved me for getting some interesting light during day time in 2nd day. The first day I was in glacier point around sunset after long drive from Bay area (my previous post), there is no cloud, nothing, and I see lots of haze in the air make me think I am back in China. I was little disappointed with sky and air at the time, but once sun goes down, it created interesting intensive red hue and atmosphere in air, the results turn out pretty good.
I don't know the fire has been started for several days which is the reason of haze
I will show some images later with fire going once I go through them wish it is under control soon
Nice images again, Like last one. it is already snowing?
burningheart - terrifying!! I don't miss snow at all.
Guys great images!
Ron - I didn't notice that much of a difference between my coded and uncoded 28cron shots. A bit of vignetting and slight blue/green colour shift in the bottom sides/corners (I have a filter on the lens, no hood). Not much else.. I'll have to shoot a few more shots in various conditions. Since your findings differ, I might just turn off corrections on all my lenses and see where it gets me in various lighting conditions vs corrections on.
Joakim - love the lines! Berlin Hbf is a photographer's delight, well, at least it would be for me, with how the lines and light interact. It's fascinating both above ground and below ground.
sebboh wrote:
very interesting and thanks for doing this Ron.
i'm quite surprised leica is choosing to pull back the highlights as part of lens corrections. that would definitely explain why many people think the lens isn't as magical on the m240. i think i would probably choose to shoot uncoded and use the lightroom plugin to correct color but not vignetting if i had this combo.
i would think that it would be feasible to make a firmware change allowing the user to specify whether they wanted full correction or just color correction.
Yes, it would seem uncoded would be the 'purer' route to the M9 & 28 Cron experience. The question for me, however, is whether I'm willing to give up the convenience of EXIF data. As well, I don't want to start fiddling with turning coding off and on because I'll forget to turn it back on for other lenses (been there, done that, too often). An alternative would be to alter or mask the lens's 6-bit coding to force a detection error and therefore no application of lens correction.
Referring specifically to the M240:
Since the camera doesn't know the physical aperture setting, I'm not sure Leica can make the profile reliably correct for specific aperture values. Maybe it does using the ambient light sensor and guessing? Otherwise, the profile correction would have to be a compromise so that it's effective with stopped down settings, like f/8 or 11 that require less correction, but still be sufficient for wide open. IIRC, lens code strength varied on the M9 based on ISO due to significantly increased noise at higher ISOs. On all apertures tested, the coded setting resulted in somewhat lower highlight values compared to uncoded. The difference was less at higher aperture values, but still there.
There is probably also a general difference in color response/bias between the M9 and M240 that affects images in general, and may be more noticeable when combined with the 28 Cron's particular characteristics. For landscape images, I'm particularly struck by how much deeper the blue sky is with the M9 vs. the M240. Greens are also richer in the M9. The lens's vignetting characteristics combined with the M9's profiling results in a more dramatic rendering. Of course tweaks in post can minimize the difference, but one area that is difficult is the difference in tonality along the lower half of the image, seen as a deepening/flattening of tones in M240 images. You can see it in the samples below - how the foreground grass appears more muted compared to the M9 versions.
Here are a couple examples (details in watermark), chosen in the hopes of bringing out the E6 'chrome' qualities of the M9 with the 28 Cron:
Here is the M240 file tweaked to approximate the stock M9 file above:
And another scene:
Other than trying to get WB in the ballpark, these DNGs were converted with default LR settings and therefore no tweaks of contrast, saturation, etc.
Oh, also discount the extreme edge vignetting as this was done with a non-Leica hood that after testing today, I can confirm does increase vignetting at wider apertures. But since it was on the lens for both cameras, that aspect is at least consistent...
Thanks for this Ron. This is a difficult comparison to make given the differences in the way the bodies render colours, and you've done an admirable job. To me the difference in the way the lens works on both bodies comes down to the vignetting. Even in your tweaked M240 shot, I don't feel that the falloff extends as far into the frame as the M9 shot. I think this is the largest part of it's character.
As far as the colours are concerned I think this is largely the difference that is true of the M240 vs M9 in general.
Gary, great images of Budapest! If you are still there, make sure you take in the hot spring baths. Its a nice relaxing way to end a day.
Ron, thank you for taking the time to do this test. It is always great when someone helps understand our glass and its characteristics better. I tend to shoot the cron with a polarizer a lot, so I may not have seen the difference in the sky that you have shown. Cheers.
Thanks Cal and Andrew. Good point about using a polarizer. Definitely more feasible now with live view and EVF options and would certainly address the matter of sky brightness. Have you found that it works OK for 28mm with a fairly even effect? The wider the lens, the more of a problem with the polarizing effect being somewhat patchy, if I recall... Might be worth getting a polarizer to try... Any particular brand preference?
I should add it's easier to see the differences by viewing the above images in an image browser and flipping between them...
Andrew Gough wrote:
Gary, great images of Budapest! If you are still there, make sure you take in the hot spring baths. Its a nice relaxing way to end a day.
Andrew
Thanks Andrew! This was a July vacation for my wife and myself. And yes, we certainly did enjoy the thermal baths (especially Szechenyi) to sooth our tired feet.
edit - iPhone pic of Szechenyi below.... I figured it would be a great opportunity to try out a Speedo but I just couldn't bring myself to buy one....
rscheffler wrote:
Joakim - love the lines! Berlin Hbf is a photographer's delight, well, at least it would be for me, with how the lines and light interact. It's fascinating both above ground and below ground.
sebboh wrotevery interesting and thanks for doing this Ron.
i'm quite surprised leica is choosing to pull back the highlights as part of lens corrections. that would definitely explain why many people think the lens isn't as magical on the m240. i think i would probably choose to shoot uncoded and use the lightroom plugin to correct color but not vignetting if i had this combo.
i would think that it would be feasible to make a firmware change allowing the user to specify whether they wanted full correction or just color correction.Yes, it would seem uncoded would be the 'purer' route to the M9 & 28 Cron experience. The question for me, however, is whether I'm willing to give up the convenience of EXIF data. As well, I don't want to start fiddling with turning coding off and on because I'll forget to turn it back on for other lenses (been there, done that, too often). An alternative would be to alter or mask the lens's 6-bit coding to force a detection error and therefore no application of lens correction.
Referring specifically to the M240:
Since the camera doesn't know the physical aperture setting, I'm not sure Leica can make the profile reliably correct for specific aperture values. Maybe it does using the ambient light sensor and guessing? Otherwise, the profile correction would have to be a compromise so that it's effective with stopped down settings, like f/8 or 11 that require less correction, but still be sufficient for wide open. IIRC, lens code strength varied on the M9 based on ISO due to significantly increased noise at higher ISOs. On all apertures tested, the coded setting resulted in somewhat lower highlight values compared to uncoded. The difference was less at higher aperture values, but still there.
There is probably also a general difference in color response/bias between the M9 and M240 that affects images in general, and may be more noticeable when combined with the 28 Cron's particular characteristics. For landscape images, I'm particularly struck by how much deeper the blue sky is with the M9 vs. the M240. Greens are also richer in the M9. The lens's vignetting characteristics combined with the M9's profiling results in a more dramatic rendering. Of course tweaks in post can minimize the difference, but one area that is difficult is the difference in tonality along the lower half of the image, seen as a deepening/flattening of tones in M240 images. You can see it in the samples below - how the foreground grass appears more muted compared to the M9 versions.
Here are a couple examples (details in watermark), chosen in the hopes of bringing out the E6 'chrome' qualities of the M9 with the 28 Cron:
Here is the M240 file tweaked to approximate the stock M9 file above:
And another scene:
Other than trying to get WB in the ballpark, these DNGs were converted with default LR settings and therefore no tweaks of contrast, saturation, etc.
Oh, also discount the extreme edge vignetting as this was done with a non-Leica hood that after testing today, I can confirm does increase vignetting at wider apertures. But since it was on the lens for both cameras, that aspect is at least consistent......Show more →
Ron very much appreciate your comparison efforts with 28cron, which are very interesting. I think as far as coding, vignetting, and color shift are concerned, your work has nailed down some issues--esp the lowered highlights on the M240.
What I really can't tell is what differences there are, if any, in handling fine details at the center, edges and corners. Those aspects interest me both in classic f/8 landscape mode with far details across the frame and in the lower corners, and also mid and close range at faster apertures. I suspect there are bokeh differences too--- but I know you have a real life, so thanks so much for what you've done so far.
Charles also reports the Cron is pretty decent on A7s afterall----I found it OK on A7.
took a 2 hr mtn bike ride today with my "ultralight" setup: M9 + 40 cron (m-rokkor version 1) L1020129-4 by unoh7, 40 cron f/6.7
My loop curls back under those high cliffs
L1020145 by unoh7, 40cron f/6.7
I like the little 40/2 but on the M9 edges and centers show some big differences--which are easy to spot at infinity with trees and leaves L1020156-3 by unoh7 40cron f/16
@Gary love the shots from Hungary!
@Joakim 24 lux station shots awesome!!
@Edward always very much enjoy your shots on the street. Cop in front of blonde billboard is wonderful!
Charlie - love the rendering of the 40, particularly nearer distance objects. But overall it has a bit of a gentler feel than what you've been posting with other lenses.
Funny you should ask about those other performance parameters. Of course I took note of those aspects and can report that the 28 Cron delivers slightly better wide aperture fine detail sharpness in the mid zone of the frame on the M240. As you know, wide open it's not corner to corner sharp, but the M240 extends this slightly. At f/8 it's a wash. What is also noticeable is a slight increase in detail resolution with the M240, and if the files are sized to M9 pixel dimensions, fine detail is slightly better rendered, particularly diagonal lines are smoother with finer slight jaggies than the M9. I'll have to run a comparison some day with the 21SEM since it's a lens that already performs very well wide open and may reveal more in terms of sensor performance differences. According to theSuede the M240 is supposed to have better sensor stack glass quality, which may allow for better/truer detail sharpness into the edges...
rscheffler wrote:
Joakim - love the lines! Berlin Hbf is a photographer's delight, well, at least it would be for me, with how the lines and light interact. It's fascinating both above ground and below ground.
sebboh wrotevery interesting and thanks for doing this Ron.
i'm quite surprised leica is choosing to pull back the highlights as part of lens corrections. that would definitely explain why many people think the lens isn't as magical on the m240. i think i would probably choose to shoot uncoded and use the lightroom plugin to correct color but not vignetting if i had this combo.
i would think that it would be feasible to make a firmware change allowing the user to specify whether they wanted full correction or just color correction.Yes, it would seem uncoded would be the 'purer' route to the M9 & 28 Cron experience. The question for me, however, is whether I'm willing to give up the convenience of EXIF data. As well, I don't want to start fiddling with turning coding off and on because I'll forget to turn it back on for other lenses (been there, done that, too often). An alternative would be to alter or mask the lens's 6-bit coding to force a detection error and therefore no application of lens correction.
Referring specifically to the M240:
Since the camera doesn't know the physical aperture setting, I'm not sure Leica can make the profile reliably correct for specific aperture values. Maybe it does using the ambient light sensor and guessing? Otherwise, the profile correction would have to be a compromise so that it's effective with stopped down settings, like f/8 or 11 that require less correction, but still be sufficient for wide open. IIRC, lens code strength varied on the M9 based on ISO due to significantly increased noise at higher ISOs. On all apertures tested, the coded setting resulted in somewhat lower highlight values compared to uncoded. The difference was less at higher aperture values, but still there.
There is probably also a general difference in color response/bias between the M9 and M240 that affects images in general, and may be more noticeable when combined with the 28 Cron's particular characteristics. For landscape images, I'm particularly struck by how much deeper the blue sky is with the M9 vs. the M240. Greens are also richer in the M9. The lens's vignetting characteristics combined with the M9's profiling results in a more dramatic rendering. Of course tweaks in post can minimize the difference, but one area that is difficult is the difference in tonality along the lower half of the image, seen as a deepening/flattening of tones in M240 images. You can see it in the samples below - how the foreground grass appears more muted compared to the M9 versions.
Here are a couple examples (details in watermark), chosen in the hopes of bringing out the E6 'chrome' qualities of the M9 with the 28 Cron:
Here is the M240 file tweaked to approximate the stock M9 file above:
And another scene:
Other than trying to get WB in the ballpark, these DNGs were converted with default LR settings and therefore no tweaks of contrast, saturation, etc.
Oh, also discount the extreme edge vignetting as this was done with a non-Leica hood that after testing today, I can confirm does increase vignetting at wider apertures. But since it was on the lens for both cameras, that aspect is at least consistent......Show more →
Ron - great post.. Not that different overall, if you know how to correct the image to your liking.
One thing that troubles me though. Do you remember around a hundred pages back when we talked about sensor reflectivity, and flare etc etc.. This was in relation to flare and veiling flare and ghosting. For some reason, my feeling is still that in certain lighting conditions the M240's sensor reflectivity (sensor stack whatever you wanna blame for the problem) causes the 28cron to lose it's magic to a greater degree, whereas it's a really good match for the M9 in almost all lighting conditions.
I haven't played with the 28cron on the M240 enough to pin down what the difference is, and how to provoke it. But just from general shooting with the 28 on both cameras, I still think there's a bigger difference then contrast/colours to be found through various light conditions. Maybe it's more noticeable where I live, due to the quality of light here?
Maybe contrast/colours really are the key to making the 28cron look better on the M240?
Gary Clennan wrote: Andrew Gough wroteGary, great images of Budapest! If you are still there, make sure you take in the hot spring baths. Its a nice relaxing way to end a day.
AndrewThanks Andrew! This was a July vacation for my wife and myself. And yes, we certainly did enjoy the thermal baths (especially Szechenyi) to sooth our tired feet.
edit - iPhone pic of Szechenyi below.... I figured it would be a great opportunity to try out a Speedo but I just couldn't bring myself to buy one....
Hahaha.. That's the first thing I did when I went to Serbia after high school graduation. Bought the Speedo and headed to the beach.... And then refused to take off my pants when we got to the water. My legs from the knees up were so white, I didn't think about it until all my cousins were stripped down and bronzed from head to toe
AGES ago, thanks for the memories
Adam, what a visual...
Gary, simply STUNNING images. Love them. Your 21mm work impresses me most. You're so consistent with that lens...
adamdewilde wrote: Gary Clennan wroteAndrew Gough wroteGary, great images of Budapest! If you are still there, make sure you take in the hot spring baths. Its a nice relaxing way to end a day.
AndrewThanks Andrew! This was a July vacation for my wife and myself. And yes, we certainly did enjoy the thermal baths (especially Szechenyi) to sooth our tired feet.
edit - iPhone pic of Szechenyi below.... I figured it would be a great opportunity to try out a Speedo but I just couldn't bring myself to buy one.... Hahaha.. That's the first thing I did when I went to Serbia after high school graduation. Bought the Speedo and headed to the beach.... And then refused to take off my pants when we got to the water. My legs from the knees up were so white, I didn't think about it until all my cousins were stripped down and bronzed from head to toe
AGES ago, thanks for the memories
Hi Adam, I agree it's possible to tweak the M240 & 28 Cron but I think many prefer the starting point offered by the M9. Maybe sensor reflectivity is a factor in certain situations, but it would probably mean shooting the two cameras side by side for an extended period to determine this, which I'm not going to do. So far, my feeling is the primary factors for day to day differences between this lens on each camera boils down to:
- inherent vignetting character with each camera, whether coded or uncoded
- mid tone contrast
- mid tone color saturation
- color profile biases/weighting for certain colors
The last three points are probably more a software factor than physical camera/lens property.
Looking at a number of full rez files converted from DNG, I can see pluses and minuses for each camera. The M9 has that little bit more pop and saturation and the vignetting contributes a subjective feel. In some instances, there is also better shadow detail separation and definition where sometimes the M240's shadow values posterize unattractively (though can be minimized with the green shadow fix LR plugin). The M9 offers a simpler, more direct route to a slide film kind of look. The M240 offers a slight but discernible improvement in fine detail resolution and off-center lens performance. Its vignetting character is tamer and more symmetrical in the corners, which might appeal to those coming from DSLRs with a vignetting phobia. Then there are the host of M240 refinements over the M9 that perhaps don't directly influence image quality but subjectively affect the shooting experience.
IMO, the M9 has the better default wide aperture look with the 28 Cron but may not be obvious or noticeable to the untrained eye in situations that aren't side by side comparisons. Stopped down and the differences are more subtle, with the M9 retaining slightly more vignetting. There is of course also the color profile difference.
Is it enough for me to return to shooting the M9? No. If I really wanted the M9/28 Cron look, I could spend a few minutes in LR to create a preset to apply in one click to M240 & 28 Cron images. But it's good to know how it performs and reminds me there's not really much of an image quality difference between these two cameras in average use conditions (which also isn't why I got an M240).
As with every successive camera model, the refinements result in compromises that each of us have to weigh based on our type of work.
Ron - Fair enough. I'd actually give up shooting the 28cron before I gave up the M240 for the M9. I think the improvements of the body far outweigh the cons brought on by a few lenses (of course if I needed wides and my camera performed like an A7, that would be a different story). To be honest, I'm more neglecting the 28cron in favour of the 35FLE then the fact that it's not fantastic on the M240 (in some lighting conditions). Because at the end of the day, as you proved, LR goes a long way when you know what you're doing.
Also one more question. Why would you think the two points listed by you are software related rather then the relationship between the lens optics and sensor reflectivity?
- mid tone contrast
- mid tone colour saturation
Wouldn't one assume that contrast and colour saturation would be similar across the whole sensor, or that Leica would want the sensor to render the contrast and colour the same? If this was done to compensate for lens rays and a non-transparent sensor, then it would effect all lenses to some degree (non-transparent as in, not film), and if so, I haven't noticed this.