Fun with the Leica SL (digital)

[ashwinrao1]

A birthday party to follow, with the 35 FLE

 

[uhoh7]

Those are really nice, Ashwin

Over at FM i begged for a summation of what's known about M lens performance on the SL and got this reply from Steve Spencer:

"Here is what we know from various tests. The 28 elmarit ASPH shows worse corner performance on the SL than the M240. We know this from 3 sources. Jono Slack has posted images showing this on the Leica Forum. Another poster in the link that Hiep originally posted and I reposted found the same thing, and those who have seen it say Sean Reid has shown the same thing. We also know that the 35 cron ASPH has worse corner performance on the SL than the M240. We know this from the link I reposted earlier and from Sean Reid. We also know than the 28 cron doesn't perform quite as well on the SL as the M from the link I reposted earlier. We also know that the 28 Elmarit ASPH was designed in 1993, the 35 cron ASPH in 1997, and the 28 cron in 2000. All designed more than five years before Leica came out with the M8, so it seems reasonable to assume those lenses were designed with the plan they would be used on a film camera not a digital camera.

So do we know that the SL has thicker cover glass? There is no source for it yet, but the impaired corner performance on these three lenses compared to the M240 is certainly evidence consistent with the idea that the SL has thicker cover glass. No we don't know this yet, that is why I said it appears to me, and it does appear to me that the SL has thicker cover glass based on its performance with those three lenses.

In addition to the above we also have evidence from Sean Reid that the new 28 lux ASPH shows notably better performance on the SL than the M240. Several people have commented on the test from Sean Reid and all report seeing fairly clear evidence that this lens performs better on the SL. Given that the 28 Lux ASPH came out this year and performs better on the SL that also came out this year, I don't know what you want to call it but I don't think it is unreasonable to say that it is likely that the 28 Lux ASPH was tested on the SL and that Leica knew it performed better on the SL than the M240. Now maybe you wouldn't call that being designed for the SL, but I would. It certainly appears to me that the 28 Lux ASPH was designed to work with the SL.

Putting these two lines of thought together it does certainly appear to me that the SL has a bit thicker cover glass (at least optically) and that the 28 Lux ASPH was designed to work with the SL (and its cover glass). That does make me wonder whether as Leica renews the 28 cron ASPH and the 35 cron ASPH (and maybe a slower 28mm) if these lenses won't work much better with the SL. I am optimistic they will, and further I expect the new M to have cover glass much like the SL."

Leica SL images and specs. Looks somehow familiar... - FM Forums

 

[fotografz]

Not sure what everyone is talking about regarding the skin tones, profiles, and software. I understand their value, and the SL will benefit from tweaks over time ... but ... just an observation from my own experiences:

All I see in the little boy/accordion shots is a slightly off WB compounded by a slight underexposure. A minor shift the Photoshop mid-tone color balance slider to the cyan side and tweak of mid-tone exposure fixes it in 3 seconds. No flushed skin tone or red skin shadows, and no crushed reds, better separation of the shirt and accordion with more tonal detail.

There is no exif data accompanying those shots, so I can't see what ISO or WB setting was used ... but with many of today's energy saving incandescent lights the color temp often falls outside/below the auto WB range of many cameras. I took note of this over time while shooting weddings as venues switched to the new bulb types. In many cases of low light, higher ISO images auto WB wan't enough to counter-act the strong red/yellow cast. Oddly however, manual WB test shots would get it closer ... often much closer.

BTW, the initial S2 images had a bias to the magenta which produced a pinkish skin tone which was improved in later updates, but the M240 was more complex, and I never could get natural skin-tones (at the time) regardless of PP work ... this SL doesn't seem to be suffering from that sort of interdependent color balance issue as far as I can tell.

This leads to a key SL question ... how easy/fast is it to set manual WB with a test setting shot for available light work?

- Marc

 

[scott kirkpatrick]

I don't go to FM because of a suspicion that the accuracy level is quite varied there. The response that uhoh7 quotes has some questionable statements. It may be conflating two 28 Elmarits, since when I purchased an M8 in 2006, the Elmarit-asph 28 had just been introduced to the market, and was advertised as the latest design from a renewed team. Its properties are such that it is extremely sharp out to about the radius of the M8's Kodak chip (1.3X). I'd like to know why it was said to be designed in 1993. The SX 28 reached the market this year as a mainline product, but was in reviewer's hands two years previous and sold as part of a collector's kit with a fancy body and three Summiluxes in 2013, so its design began well before the three years ago when "discussions that lead to the SL" are said to have begun.

A bigger deal is that the differences between the M240, SL, and Sony industry standard chips are not just cover glass thickness (Roger Cicala should get credit for sorting out that factor), but also microlens shape and placement, and well design. Possibly even pixel cell layout, as asymmetric cell layout played a role in the Italian flag syndrome that the M9 encountered.

So, grain of salt, folks.

scott

 

[ptomsu]

I don't go to FM because of a suspicion that the accuracy level is quite varied there. The response that uhoh7 quotes has some questionable statements. It may be conflating two 28 Elmarits, since when I purchased an M8 in 2006, the Elmarit-asph 28 had just been introduced to the market, and was advertised as the latest design from a renewed team. Its properties are such that it is extremely sharp out to about the radius of the M8's Kodak chip (1.3X). I'd like to know why it was said to be designed in 1993. The SX 28 reached the market this year as a mainline product, but was in reviewer's hands two years previous and sold as part of a collector's kit with a fancy body and three Summiluxes in 2013, so its design began well before the three years ago when "discussions that lead to the SL" are said to have begun.

A bigger deal is that the differences between the M240, SL, and Sony industry standard chips are not just cover glass thickness (Roger Cicala should get credit for sorting out that factor), but also microlens shape and placement, and well design. Possibly even pixel cell layout, as asymmetric cell layout played a role in the Italian flag syndrome that the M9 encountered.

So, grain of salt, folks.

scott

I also have some issues with the referred post at FM. I bought my 2/28 ASPH in 2007, it was one of the latest M lens designs by then and supposed to be more than excellent.

Which BTW it is on a analog, but more importantly also on a digital M, especially the M240 where I tried several times and was always pleasantly surprised. I would be surprised to see it replaced after such short time, especially when one thinks about common Leica M lens lifecycles. IMHO too many people are writing too much B.... about too many products in these days.

The 1.4/28 may be better, but it for sure is also much larger and heavier, letting the higher price aside. While this is the case, I am sure it was designed mainly with the M in mind, as I am sure we will see in the future a native 28 SL lens providing all benefits of the SL system, especially AF.

 

[ddanois]

One with the Zeiss Otus 55 @ 1.4

 

[uhoh7]

A bigger deal is that the differences between the M240, SL, and Sony industry standard chips are not just cover glass thickness (Roger Cicala should get credit for sorting out that factor), but also microlens shape and placement, and well design. Possibly even pixel cell layout, as asymmetric cell layout played a role in the Italian flag syndrome that the M9 encountered.

So, grain of salt, folks.

scott

That's why I make the quote, Scott, folks like you can weigh in on these ideas.

Do you have sources about the differences in microlens shape and placement, and/or well design in the SL sensor vs 240?

I agree with you guys on the 28 cron which I use all the time on my M9 and is fantastic on that camera. Apparently it is not quite as good on the SL?

This may be coming from Sean Reid.

I find it pretty hard to believe a new M would perform less well than the current M240 with older lenses. Apparently there is evidence the 28 elmarit ASPH and 35/2 ASPH are not quite as good on the SL.

 

[Godfrey]

...
Onto some additional photos, with the R-80-200 f/4 Vario Elmar. A great lens to add some length. Most of these were shot wide open...it's a sharp lens

- - - Updated - - -

Do get yourself up to LR 6.3, Ashwin. The profile Adobe has supplied is really very good.

I like these two photos. The 'sun at horizon' pair didn't seem up to your usual standard.

G

 

[glenerrolrd]

Having spoken with lens designers at both Leica and Zeiss ..its not the age of the lens thats the issue . Many of the Leica S lenses were based on existing designs well over 10 even 20 years old .

At Zeiss the newer lenses for the Sony A7 series have been formulated using the sensor glass as the final element in the design . The computer driven designs then “tweak” the original design to optimize for a given camera . If you want the best from an M lens ...use it on an M ....the sensor/glass was designed to use the M lenses.

The other factor is the ongoing improvement in lens coatings ..this is how the new ZF/ZE lenses have been improved over the prior versions .

Of course there are new lens designs that have improved performance overall . The OTUS for example . But don t immediately assume that a newer design is “Better” . Check out the tests at LensRental for examples .

One very nice aspect of the new Leica M Summiluxes is that they all have a similar aesthetic/color signature . I have the 21/24/28 and 35 wide angles for example ...processing files from any of these is near identical . These are my lenses for M street shooting ....for landscape you would want the 18/21/24 /3.4-3.8 wide angles because of the improved flatness of field and edge sharpness.

The 28/2 summicron has been a favorite since I used it on the M8 . It has a beautiful color signature and a very smooth bokeh . Terrific lens in high contrast lighting where deep color saturation (without punishing contrast) is desired . Pair it with a pre asph 50/1.4 summilux and you have classic “Leica” look images . I use Manos American Color as my reference if that helps clarify my intent .

The lenses render differently and it depends on the subject/lighting and desired aesthetic ...which is “best”.

My guess is that the rumored 28/2 is an upgrade in the lens hood as the original is awful and routinely replaced by either he old 35/1.4 asph hood or a cheap round screw in from eBay.

 

[glenerrolrd]

One with the Zeiss Otus 55 @ 1.4
View attachment 114501

Couple of questions for you.

1. Were you able to get the Novaflex Nikon to T adapter ?

2. Have you encountered any difficulty with the EVF brightness when shooting in the Florida sun ? I shot Kite Boarders at sundown with my 90 M lens on the SL and I could see well enough to focus .

 

[Godfrey]

...
This leads to a key SL question ... how easy/fast is it to set manual WB with a test setting shot for available light work?

The FN button is assigned to custom WB by default, takes just a second to make a setting. They provide two tools as well, whole frame and selected area. Easy as can be.

G

 

[scott kirkpatrick]

...

Do you have sources about the differences in microlens shape and placement, and/or well design in the SL sensor vs 240?

I agree with you guys on the 28 cron which I use all the time on my M9 and is fantastic on that camera. Apparently it is not quite as good on the SL?

This may be coming from Sean Reid.

This stuff is quite scattered about. I pay closest attention to Sean Reid, Jono Slack (who is observant, but doesn't wish not to be thought a techie), and Sandy McGuffog. I've seen comments that Leica has worked on all three elements. Roger Cicala says that the cover glass thickness is 0.5 mm in the M8, 0.8 mm in the M9 and M[240], and he doesn't have ten SLs in house yet, so I don't know what he expects for it. The Sonys and Olympus/Panasonic M4/3 sensors have a 2 mm thick cover glass. He has taken a Sony apart and measured that. Sean has reviewed the 35/2.0 Summicron-asph and the CV35/2.5 on both M[240] and SL, as well as the 28 Elmarit-asph (the tiny one), the 28/1.4 Summilux-asph, and the CV 35/3.5 (which is rather compact). You'll have to read his report but I concluded that when he sees an improvement of the SL over the M[240], it is a small improvement that would be hard to see surviving in my handheld normal exposures. Sean didn't have a 28 Summicron handy -- the SX28 is fast and the EL28 is small, and that pretty much covers his requirements. I think Jono did provide some careful comparisons of corners with the SC 28 on both cameras, but unless Google is indexing him, you will have to crawl through an awful lot of LUF posts to find it now. Maybe the thread of M&R on SL has it.

scott

 

[Vivek]

Scott, I have taken the sensors apart in Sony cams to actually measure (unlike Roger, AFAIK) and I can assure you that the total glass thickness in the A7 cams is 2.5mm. Zeiss when they issued the Loxia also said that these lenses were corrected for 2.5mm.

 

[Godfrey]

Probably my last lens testing stint ...

I was interested in the color shifting behavior*with various R and M lenses after discussions here, and the light in the living room today provided a nice, evenly lit, white wall to work with. So I pulled out four R wide to normal lenses (Elmarit 19, Elmarit 24, Summicron 35, Summilux 50) and four M wide to normal lenses (Elmar 24, Color Skopar 28, Color Skopar 35, Nokton 50). The three Voigtländer lenses I exposed without lens code and then with lens code.

SL was set to manual exposure, fixed WB, ISO 200. I made three exposures per lens: one at f/4, one at wide open (whatever it might be), and one at fully stopped down (f/16 or f/22) compensating with exposure time to maintain as best possible an even exposure across the series. I picked f/4 as my reference "normal" aperture because, as it turns out, I tend to be shooting at around f/4 a great deal of the time. I could have just as easily picked f/5.6 or f/8, but I don't think it really matters much given what I see in the results.

Once all the exposures were made, I rolled them into LR and leveled the exposure at the center of*all the exposures to bring them all in line (within .3 stop) with the histogram peak at about 3/4 towards the white side. I also made a cursory WB exposure adjust at the same point for each exposure—in most cases, the changes were trivial to small, but there was some drift as the light varied a bit. This normalizes everything and makes evident the differences between the lenses.

To me, bad color shifting is when the color shifts polychromatically; in other words, if the whole frame shifts to a little blue or a little magenta, that's easy to manage. It's when the color shift goes blue in the middle and magenta on the edges, or vice versa, or one side green and the other side red, etc, that it's a problem.

A lot of lens characteristics were borne out by this test, but to cut to the color shifting results ...

• The only one that I would simply find unusable for color work is the Color Skopar 28mm f/3.5. It color shifts blue on center and magenta red on the edges, both with no profile and with the profile that previous experience has shown works best with it.
• All of the M-mount lenses shows more and deeper monochromatic color shift than any of the R lenses. Most also show more edge/corner darkening throughout the range, with only the Nokton 50 getting close to the R lenses in terms of even illumination at f/4.
• The only one of the R lenses that showed much corner/edge darkening is the Elmarit-R 19, and even there it is an acceptable amount given the extreme wide angle nature of the lens.
• All the Rs and the Elmar-M 24 show only a small amount of monochromatic color shift and both f/4 and fully stopped down are both very even illumination.

My take on all this: By and large, most M lenses work well enough to use for non-color critical work. But, given the option, I'll pick an R lens over nearly any M lens in the sub-50mm focal length range for the SL and not worry about the extra bulk. They work noticeably better overall. The Summicron-R 35 and Summilux-R 50 @ f/4 and f/16 produce nearly perfect illumination across the entire field with no noticeable color shift.

I'm done with that. My lenses work well enough, and I'll pick R over M lenses for this camera. From this point on, I'm going to make photographs and let everyone else debate the minutiae of specific lens performance, color shift, cover glass thickness, etc.

G

 

[ddanois]

Couple of questions for you.

1. Were you able to get the Novaflex Nikon to T adapter ?

2. Have you encountered any difficulty with the EVF brightness when shooting in the Florida sun ? I shot Kite Boarders at sundown with my 90 M lens on the SL and I could see well enough to focus .

Actually, I have the Nikon to T adapter on order at B&H but it still hasn't arrived. I used the Novoflex Nikon to M and then the Leica M to T adapter as a work around.

As for the EVF, I find the brightness of the EVF a bit low as I bring it up to my eye but once I'm looking through it, the brightness is more than sufficient. I was concerned about this as well since I didn't find the Q EVF very usable in the Florida sun but SL is really nice.

 

[k-hawinkler]

Can one adjust the brightness in the EVF?

 

[Godfrey]

Can one adjust the brightness in the EVF?

The SL EVF and LCD can be adjusted for brightness ... SETUP -> Display Settings -> LCD Brightness
I believe there are Automatic and six steps from dark to bright to choose from.

G

 

[johneaton]

BTW, for those (like me!) who prefer to use an L-plate to switch between landscape and portrait mode on a tripod, it appears that the RRS Multi-Camera L-Plate fits and works well .

 

[D&A]

Dave,

I'm seeing excellent skin tones, both direct out of camera and with the LR 6.3 camera profile. I'm not seeing any teething pain here, not like there was with the M typ 240.

How much better it has to be before you're satisfied, I don't know, but I'm quite satisfied as is. A child, in indoor light, wearing a red shirt and holding a red toy up near his face ... And you want perfect skin tones with no adjustments? I don't think there any camera that can do that. Ever.

G

Godfrey, you are putting words in my mouth. Where did I write or say I expect, desire or in your words did I say I want perfect skin tones with no adjustment. Please dont attribute statements to me that I never said nor ever implied. I bring this up because once before you did something similar to something I posted and upon defending what I wrote, you retracted your words.

I only shoot RAW and I do and expect to adjust every file and have done so since the very 1st digital cameras. Many times I have the task of adjusting more than a 1000-1500 files on a given shoot, in some of the most difficult mixed lighting senario's one can imagine and most have to be painstakenly adjusted for use on national and sometimes international applications.

My comments although centered around the young boys skin tone, focused more on the reds in general in that series of images and gave my opinion that the reds appeared to be crushed. Again the key word is "appeared". I can't be certain unless I had the actual files to work with or in some cases listen to the opinion of the original photographer's view of their own files (in this case Ashwin and he posted some words regarding his files a few posts above).

Personally in a general sense I often find images of red flowers with large closely spaced pedals as a decent indicator if a color (such as red in this case) is crushed.

Lastly permit me to make this perfectly clear. I am not casting aspersions on the SL...on the contrary. Likewise my preference for color output at base ISO of the M9 over the M240 sometimes seems not to sit well with a few. Yet there are some of us who are of this opinion and it sometimes elicits a strong response. Its simply a preference and is based on a number of factors and again its not saying anything negative about the M240. There are sitiations where no matter how much one adjusts a certain camera's files with regards to certain skin tones or certain colors, under some lighting senario's, that it cannot emulate or match those of amother camera....or if it can, requires so much work with the files, that simply its becomes too much of a issue in pratical terms, regardless of profiles or actions in post processing. Thats why there are preferences with regards to output of one high end digital camera vs another.

Lastly if I feel the SL has some teathing pains , again thats just my opinion. I know of no perfect digital camera, especially when first released as a new model and more times than not there are aspects that are addressed down the road via firmware updates. Whether that be operational, or output or some other factor. My definition of teathing pains may be different than how others characterize it and the word doesn't imply a defect or inadaquacy of the product, but simply that the product could possibly benifit from some tweaking or attention/addition to certain details.

Dave (D&A)

 

[GrahamWelland]

Seems to me that this is the ultimate R series camera that Leica promised back when I was shooting their Leica Digilux cameras. I sincerely hope that those who kept their R lenses vs converting them will take advantage of the platform.

say what you will about Leica, they get there eventually but is it too late?