21mm Summilux on M9

[adan]

Advertisement (gone after registration)

Googaliser (and Thrice) - My mistake.

 

That is classic lateral CA you're showing, and it seems high in your lens compared to the one sample I've tried (as previously posted) on an M9. I'll let Sean Reid speak for himself if he weighs in on this thread - but I don't think he'd classify that as "mild" if that's what he'd seen in his sample.

 

As Wilson says, Leica should see your examples. Personally, I'd suspect the floating element is not adjusted correctly. I've seen samples from 75 'crons with the FE out of place, and CA and smeared corners are one of the obvious symptoms.

[Googaliser]

Googaliser (and Thrice) - My mistake.

 

You're trying to help and didn't have all the facts - so no apology needed (although its appreciated - forums can be pretty cut-throat - I always hesitate before posting...)

 

That is classic lateral CA you're showing, and it seems high in your lens compared to the one sample I've tried (as previously posted) on an M9. I'll let Sean Reid speak for himself if he weighs in on this thread - but I don't think he'd classify that as "mild" if that's what he'd seen in his sample.

 

I've sent Sean a PM

 

As Wilson says, Leica should see your examples. Personally, I'd suspect the floating element is not adjusted correctly. I've seen samples from 75 'crons with the FE out of place, and CA and smeared corners are one of the obvious symptoms.

 

I returned the lens today. They have another couple of copies in stock - so I have agreed to go back this week with my M9 (which they don't have in stock...) to do some more tests. Ideally I will try to shoot the same scene next Saturday morning. Normally I would be reluctant, but given the interest in this post I feel a responsibility now to get to the bottom of whether this is an inherent issue with the 21 'Lux/M9/LR system, or just a duff copy. If I were a betting man - I would put my money firmly on the latter.

I also left the files so they could also be sent to Leica for comment.

 

Thanks for all your input and software suggetions. I will keep you posted.

 

Rgds

[georg]

I second that, with such expensive lenses there should be no visible sample variation. it's within (very tight) tolerances or it has to be replaced/repaired, period! Maybe some element became missaligned due to transport!?

[Googaliser]

Maybe some element became missaligned due to transport!?

 

I doubt it - I have seen Russian dolls packed worse than the 21mm 'lux ! It is in a padded presentation box within a large foam box - all very impressive (although I would have prefered a reduction in price than a fancy box which I will never use)

[sclamb]

Here is a thread that I started on the issue with my 24/1.4 and a picture taken at f/8 where I was seeing colouration that I didn't expect to see at f/8:

 

http://www.l-camera-forum.com/leica-forum/leica-m9-forum/102127-purple-fringing-24-1-4-a.html

 

I have since run the same picture through C1 and applied the CA tool (the C1 crop shown in the thread did not have this applied but it still showed reduced colouration) and the purple goes away completely. There is also a Purple Fringing checkbox that does a good job too. Very nice tool and C1 definitely seems superior to me for converting M9 DNGs.

 

I sent the picture and crops to Nobby at Leica UK and he sent them to the head of lens design in Solms and the response was:

 

"In reply to the question of the customer you should know that the colour fringe the customer complains is a system (lens) inherent effect and there is no need to adjust the lens. To compensate colour fringe you can use a special tool in Adobe Lightroom 2.5 (resp. Camera Raw 5.5) DNG converting named defringing. By using this tool the effect of colour fringing is noticeable reduced."

 

I must say that what I see is significantly less than you are seeing in the 21/1.4 pictures. I suggest you send the picture and crops to Nobby at Leica UK and ask him to send them to Solms for evaluation.

 

Simon

[markowich]

Here is a thread that I started on the issue with my 24/1.4 and a picture taken at f/8 where I was seeing colouration that I didn't expect to see at f/8:

 

http://www.l-camera-forum.com/leica-forum/leica-m9-forum/102127-purple-fringing-24-1-4-a.html

 

I have since run the same picture through C1 and applied the CA tool (the C1 crop shown in the thread did not have this applied but it still showed reduced colouration) and the purple goes away completely. There is also a Purple Fringing checkbox that does a good job too. Very nice tool and C1 definitely seems superior to me for converting M9 DNGs.

 

I sent the picture and crops to Nobby at Leica UK and he sent them to the head of lens design in Solms and the response was:

 

"In reply to the question of the customer you should know that the colour fringe the customer complains is a system (lens) inherent effect and there is no need to adjust the lens. To compensate colour fringe you can use a special tool in Adobe Lightroom 2.5 (resp. Camera Raw 5.5) DNG converting named defringing. By using this tool the effect of colour fringing is noticeable reduced."

 

I must say that what I see is significantly less than you are seeing in the 21/1.4 pictures. I suggest you send the picture and crops to Nobby at Leica UK and ask him to send them to Solms for evaluation.

 

Simon

 

illuminating comments from leica about the IQ of their 5000.- euro lenses. i am not going to take this and will send mine back.

peter

[Googaliser]

Advertisement (gone after registration)

illuminating comments from leica about the IQ of their 5000.- euro lenses. i am not going to take this and will send mine back.

peter

 

I agree with you - I am shocked if this is a verbatim in context quote.

[georg]

Simons sample shows an amount of CA which is propably inevitable in this specific configuration (oblique light rays passing through a 0.8mm thick cover glass) and there's a reason why Leica (as well as Schneider/Rodenstock with their "digital" lenses) talked about taking the cover glass into account in the optical design of the S-lenses. Zeiss/Alpa had the very same problem with their Biogon 38mm, even on sensors sensitive to oblique light rays (like the 33MP Dalsa) it showed some aberrations that are unknown with film.

But "Googaliser" should send his sample to Solms and tell us what's their "excuse" - I think it's a serious optical defect due to some kind of missalignment.

[sclamb]

The CA I see is at the edges of the frame and where there is high contrast. I don't see any green just the purple colour, and the bulk of the frame is fine.

 

I agree it is odd that the head of lens design thinks this is normal, but then I don't understand the challenges and trade-offs of making a 24/1.4 or 21/1.4 lens, and how much CA one should accept. Who am I to argue a point I don't understand. Perhaps there is a reason that LR and C1 both have CA and purple fringing removal tools, maybe in that it is an inevitable result of this type of fast, wide prime lens design.

 

Simon

[markowich]

The CA I see is at the edges of the frame and where there is high contrast. I don't see any green just the purple colour, and the bulk of the frame is fine.

 

I agree it is odd that the head of lens design thinks this is normal, but then I don't understand the challenges and trade-offs of making a 24/1.4 or 21/1.4 lens, and how much CA one should accept. Who am I to argue a point I don't understand. Perhaps there is a reason that LR and C1 both have CA and purple fringing removal tools, maybe in that it is an inevitable result of this type of fast, wide prime lens design.

 

Simon

 

just to set the record straight,

i own 11 nikon lenses, 7 hasselblad (H) , 4 schneider (alpa mount), 2 rodenstock (alpa mount) and 5 leica M lenses. most of the problems i have had so far occured with leica. i am not going to buy any more from them before they shape up their QC.

the m9 is an interesting camera, but the output reminds me of DB one or two generations ago. i see noise at base iso, in dark-bright transition zones, similarly to Aptus 75, at 1 stop beyond base iso. i sincerley hope that M10 will feature a state of the art sensor, usable above 250 iso. at this point the only reason i keep bothering with leica M is the weight factor.

peter

[Mauribix]

i sincerley hope that M10 will feature a state of the art sensor, usable above 250 iso. at this point the only reason i keep bothering with leica M is the weight factor.

peter

Ok Peter, dunno what kind of tools (lenses apart) you're used to, but "USEABLE ABOVE 250 ISO" seems a bit exaggerate IMHO., do you really think that the M9 and M8's (as well) sensors output are unusable?

I might be missing something, but that's absolutely an unfair statement in my experience with these cameras.

[hoppyman]

Hi Googaliser. I don't believe that you are seeing chromatic aberration from your lens there. I think that this may be related to 'bloom' from the sensor when saturated (clipped or nearly so). I'm sure that is not unique to the Leica sensors of course. Certainly possible that the Raw correction happening in camera or the Raw converter may need correcting, but I'm not convinced these are what you are seeing. Wolfgang at the Akadamie loaned me one of these to shoot with at the last Forum meeting in Hessenpark. Here are a couple of 100% uncorrected crops from on my M8.

You can clearly see the odd bloom at some overexposed edges in the interior shot (@1.4) It is not on every edge and see especially the edges from the outside example(stopped down a little). Please keep us informed here as you investigate this further but hold onto that lens!

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[arthury]

Googaliser,

 

I have never seen CA surfacing in this manner. They almost look like someone took a paint brush and hand-colored those magenta edges. I am not saying you did that ... but, that's how they looked to me.

 

Most CA has a blur edge because of the nature of how CA is formed. These edges do not have the characteristic blurry edges of a typical CA.

[ohnri]

The WATE exhibits far less of this artifact than the 21 lux does at identical focal length, f/stop, shutter speed, same M9, same scene, etc.

 

Best,

 

Bill

[ho_co]

Googaliser--

First, you should send Leica your results and comments.

 

Second, I don't know Nobby; I don't know whether Peter Karbe indeed responded; and although I don't know the example Simon sent, he says it showed less of the defect you're showing. So there's no way to generalize from his lens to yours (or to peter's).

 

Third, as someone pointed out above, we are discovering that classical optics doesn't cover all the aberrations we are becoming aware of with digital.

 

My suggestion above that you try another RAW processor was intended to try to help overcome the problem, not to introduce a different workflow. I find Capture One a pain, but 1) if you have a specific problem and 2) Capture One takes care of that problem automatically, that may be a fact to work with.

 

 

So far in this thread, we have some good guesses, some solid experience and some pontification, but we haven't solved the problem or even described it in a way that we all accept.

 

I second the recommendation that the lens whose results you've posted be returned to Leica with your sample shots for evaluation. And I know that everyone who has posted here would be very interested in the upshot.

[markowich]

Simons sample shows an amount of CA which is propably inevitable in this specific configuration (oblique light rays passing through a 0.8mm thick cover glass) and there's a reason why Leica (as well as Schneider/Rodenstock with their "digital" lenses) talked about taking the cover glass into account in the optical design of the S-lenses. Zeiss/Alpa had the very same problem with their Biogon 38mm, even on sensors sensitive to oblique light rays (like the 33MP Dalsa) it showed some aberrations that are unknown with film.

But "Googaliser" should send his sample to Solms and tell us what's their "excuse" - I think it's a serious optical defect due to some kind of missalignment.

 

i do not recall having seen any CA on photos taken with my rodenstocks/schneiders with Aptus 75 and P65+. but of course, this is an unfair comparism since those are slow lenses.

peter

[georg]

That's exactly what I meant, the Digitar/HR-lenses take the cover glass into account (as another optical element) as well as the S-lenses or the fixed front-filters from some Leica-lenses (180mm Summicron, 70mm Summarit) - those are all nearly perfectly flat optical glasses, but it seems to make a difference.

But these lenses have to be used with a special optical element (to "simulate" the cover glass) when used with film.

[markowich]

Ok Peter, dunno what kind of tools (lenses apart) you're used to, but "USEABLE ABOVE 250 ISO" seems a bit exaggerate IMHO., do you really think that the M9 and M8's (as well) sensors output are unusable?

I might be missing something, but that's absolutely an unfair statement in my experience with these cameras.

 

maurizio,

well, 'unusable' is a tough word, also film at 3200 was 'usable'...but what i observe with the M9 is a huge reduction of DR when leaving base iso, combined with ugly color speckle noise (when pixel peeping of course). i claim that this is just not state of the art anymore sensor technology. i find it really unfortunate since as a concept the M9 should be a low-light camera. still the output at base iso is very good but i think twice when dialing in >iso 250.

peter

[hoppyman]

I'm following this with interest to learn from. After reading the posts from Thrice and quotes from his expert, I can see that I was wrong to assume that the issue came primarily from sensor 'bloom' in this instance. I reviewed the 40 frames that I have from trying this lens on my M8. Processing with default settings in ACR 5.4 at that time. I had shot a mix of contrasty daylight at medium apertures and a selection at f/1.4 inside usually deliberately trying to provoke any weaknesses by having a bright light source in frame and the edge of the frame where possible.

I looked at all 40 DNGs at 100% now and offer these results to try to contribute to the discussion.

 

The worst case occurs when I have out of focus high contrast edges, usually to clipped or nearly so, at the edge of the frame. I can see both purple and green clearly in many instances.

I found a couple of frames where the high contrast edge against a very bright area also fell in the sharper focus. I can just detect a thin fringe in these cases.

 

In most instances where I have well defined edges but no clipped or at least bright background I can detect almost none or no effect AT 100%

 

I think that the OP is right to consider the practical results from his specific system. I do not know if the specific lens varies from the norm for this design.

 

time.quote=hoppyman;1105520]Hi Googaliser. I don't believe that you are seeing chromatic aberration from your lens there. I think that this may be related to 'bloom' from the sensor when saturated (clipped or nearly so). I'm sure that is not unique to the Leica sensors of course. Certainly possible that the Raw correction happening in camera or the Raw converter may need correcting, but I'm not convinced these are what you are seeing. Wolfgang at the Akadamie loaned me one of these to shoot with at the last Forum meeting in Hessenpark. Here are a couple of 100% uncorrected crops from on my M8.

You can clearly see the odd bloom at some overexposed edges in the interior shot (@1.4) It is not on every edge and see especially the edges from the outside example(stopped down a little). Please keep us informed here as you investigate this further but hold onto that lens!

[ATTACH]171354[/ATTACH]

 

[ATTACH]171355[/ATTACH]

[Advertisement]

Hallo,

Look here: Leica M

[sclamb]

Whilst not related to the 21/1.4 issue described initially, here are some pictures from my 24/1.4 (pardon the messy clothes!). DNGs processed using C1 with no changes made, and no CA or Purple Fringing reduction functions used.

 

Full frame, ISO 160, 1/500s, f/1.4:

 

 

Crop showing extent of CA with dark area against overexposed background:

 

 

Full frame, ISO 160, 1/90s, f/4.0:

 

 

Crop showing extent of CA with dark area against overexposed background:

 

 

These don't look bad at all and nowhere near as bad as the OPs 21/1.4 pictures. If I switch on the Purple Fringing option in C1 then it all but disappears.

 

I am sure the OP can get a better performing 21/1.4. His pictures look like there is a serious problem other than pure CA.

 

Simon