Official Leica Statements

davidada · November 22, 2006

Advertisement (gone after registration)

Sean, you are totally correct please Art is not about 100% color correct high fidelity images, it is about the artists vision. I just looked at a M-series pinhole adaptor on ebay and thought I would like to try that and see what image it draws.

November 22, 2006

Hi davidada,

Take a look here Official Leica Statements. I'm sure you'll find what you were looking for!

sean_reid · November 22, 2006

1) So, my 5D without weather sealing is apparently full of rainwater and dust, as I have used it both in the rain and in the desert?

2) There is so much more to it than just weather sealing or not. Leicas have never needed it before, and Leica owners have put these cameras through hell. I am certain that the M8 matches the M7 for sealing, yet we have never heard of an M7 dying of water or dust?

3) I think before requesting a possibly quite expensive feature, it would be wise to be sure that it is necessary.

1) It means your number hasn't come up yet or the camera has not been used in heavy rain. Unsealed Canon DSLRs *do* fail in heavy rain, as many pros know.

2) The M8 is a very different camera from the M7. I've already explained why.

3) When Leica asked me, two years ago, which features I thought would be essential for the digital M, weathersealing was near the top of my list. They considered it but didn't do it.

Again, you're welcome to disagree but I will continue to argue that this is an important feature for many kinds of professional work and the M8, unabashedly, is designed and sold as a professional camera.

Sean

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

ptomsu · November 22, 2006

2) The M8 is a very different camera from the M7. I've already explained why.

Sean

Yes Sean, the M8 is a digital camera, compared to the M7 using analog media we call film - just kidding :-))

I would agree that the M8 should be weather sealed, as normal professional cameras are.

Peter

scott kirkpatrick · November 22, 2006

You do have to wonder what the conversation would be about now if everyone had continued to press for a fixed sensor. Putting it on the table, I'm one who believes that the new orthodoxy of the last couple of weeks that there was never a choice is the spin/knowledge of the moment and is likely flat wrong if you brought a few good sensor enginers to the table for a couple of days and told them to figure it out -- remember the first sesor Kodak delivered had a fix for IR.

I agree with you that good engineers could invent in-camera solutions to IR filtration in a short back-focus environment for the M8. And I hope they will, but for the M9. To put this in would seem to require 6-12 months delay and some development cost. Neither seems like something Leica will (or should) consider at the moment.

scott

sean_reid · November 23, 2006

2) The M8 is a very different camera from the M7. I've already explained why.

Sean

Yes Sean, the M8 is a digital camera, compared to the M7 using analog media we call film - just kidding :-))

I would agree that the M8 should be weather sealed, as normal professional cameras are.

Peter

That's just it, though. A digital camera is much more likely to be damaged by moisture than a film camera.

Cheers,

Sean

sean_reid · November 23, 2006

One of the factors that seems to be missed in the ongoing discussion of lens coding, is that the Leica solution obviates the need for the lens to be manually chosen every time a lens is changed. Whilst I'm all for providing choice, and a manually selectable option would be useful in a pinch to use a non-coded lens, I'm not sure that I care too much for having to make a menu change everytime I switch lenses. Even with two bodies (which I don't yet have) I will want to switch lenses on occasions and fiddling with the menu will soon grow tiresome. Much better just to get the lenses coded and be done with it.

Hi Ian,

The convenience and speed would be one of the USPs (unique selling points) of the Leica coded lenses. With the firmware change I envision, there would be three options:

1. Lens detection disabled (this currently exists)

2. Lens detection enabled (this also currently exists)

3. Manual lens selection (followed by a menu of Leica lenses one can specify)

Photographers who own only coded lenses could ignore the third option and it would have no effect on their work with the camera. Most of them would likely leave the camera permanently set to option 2.

Photographers who own some uncoded Leica lenses could choose to use option 3 in order to get the benefits of the in-camera corrections.

I will verify this in one of the tests I'm doing for the 28 mm lenses article but I suspect that most 28 mm lenses using a 486 will show a similar degree of cyan drift in the outer zones. The cyan is coming from the filter, not from the lens. I suspect that this is also true for the 24/25 and 21 mm lenses. As those articles are written, I'll test the latter two focal length groups as well. My previous tests of these lenses suggest that neither the Zeiss nor the CV lenses vignette *less* than the Leica lenses (to any appreciable degree) at equivalent focal lengths.

So let's say that the M8 is programmed to apply "X" correction for vignetting and "Y" correction for cyan drift to a file made with a given lens such as the Elmarit 21/2.8. I suspect that if those same two corrections are applied to a Zeiss 21 or CV 21 file, the result will be less vignetting and less cyan drift. If it's correct that the cyan drift is correlated to angle of view than that cyan should follow about the same pattern with all lenses of a given focal length (allowing for the fact, of course, that not all lenses of the same nominal focal length actually show the exact same FOV). As I said, I'll be testing the set of 28s soon to see if this is the case. The Leica's correction for vignetting may not completely eliminate vignetting in the Zeiss or CV lenses (and it is slight in all three 28s for example) but it will very likely improve it.

If the above is correct, and I suspect that it is, the corrections that are designed to be applied to a file made with a given Leica lens of a certain focal length are also likely to benefit files made with other high quality lenses of the same focal length. I don't think anyone expects Leica to create custom corrections for other maker's lenses. *But* since Leica's design has put us in the position of *needing* to use IR cut filters for color work, it would be an excellent move on their part to open up the "cyan drift" correction to lenses other than their own coded ones, especially those that are 28 mm and wider. Again, I'm against the idea of a closed RF system, in principle and in practice.

While the corrections may be designed specifically for Leica coded lenses, lets let photographers empirically find out how well these corrections might work for other lenses as well. Leica can and should clearly state that these electronic corrections may give undesirable results when used with lenses other than those specified. The photographer individually can then take personal responsibility for whatever results may come from using "non-specified" lenses. My gut instinct is that the results will be quite good.

Some feel that this closed-system approach will benefit Leica by increasing the number of people who choose to buy Leica lenses instead of those from other makers. Hasselblad is no doubt thinking along the same lines with some of its new introductions. But trying to force people into a closed camera system can backfire for a number of reasons. Remember that Leica has promoted the coding as an *option* (with benefits) for M8 owners. With recent developments it's likely to be no longer an option but a *requirement* for wide angle color work. Here are examples of people who may object:

1) Photographers who *prefer* the drawing characteristics of certain non-Leica wide angle lenses, and who own various examples, may not wish to buy a camera that cannot use those lenses successfully for color work.

2) Photographers who are opposed to closed camera systems may reject the Leica just as they would reject the new Hasselblad system.

3) Photographers who are new to digital rangefinder photography but interested in it (and this is potentially a large and important group) may reject the M8 because of the cost of adopting the system as a whole. Let's look at concrete examples of the costs involved for photographers who would need to begin a wide-angle M system from scratch (as was the case recently for Guy Mancuso and Jono Slack, for example). These examples don't even include the cost for filters, etc. nor do they include the costs for longer lenses. The list below also only include lenses that will perform well on the M8 (except for the potential IR issues).

Scenario One: All Leica, slower 28

Leica M8: $4800

Leica 21/2.8: $3500

Leica 24/2.8: $2900

Leica 28/2.8: $1500

Total Cost: $12,700

Scenario Two: All Leica, faster 28

Leica M8: $4800

Leica 21/2.8: $3500

Leica 24/2.8: $2900

Leica 28/2.0: $3200

Total Cost: $14,400

Scenario Three: M8 with Zeiss, slower 28

Leica M8: $4800

Zeiss 21/2.8: $1300

Zeiss 25/2.8: $1150

Zeiss 28/2.8: $1000

Total Cost: $8250

Scenario Three: M8 with Zeiss and CV, faster 28

Leica M8: $4800

Zeiss 21/2.8: $1300

Zeiss 25/2.8: $1150

CV 28/1.9: $ 500 (including adapter)

Total Cost: $7750

Scenario Three: M8 with CV, slower 21, 25 and 28

Leica M8: $4800

CV 21/4.0: $ 385 (including adapter)

CV 25/4.0: $ 330 (including adapter)

CV 28/3.5: $ 340 (including adapter)

Total Cost: $5855

Scenario Three: M8 with CV, slower 21 and 25, faster 28

Leica M8: $4800

CV 21/4.0: $ 385 (including adapter)

CV 25/4.0: $ 330 (including adapter)

CV 28/1.9: $ 500 (including adapter)

Total Cost: $6015

Every Zeiss ZM lens I've tested has performed very well. The CV lenses also have surprised many people with how well they perform despite being inexpensive. Is it better for Leica that a person buy an M8 and some CV/Zeiss lenses or that he or she not buy an M8 at all? Many here already own a wide range of Leica lenses but there's a broader market out there who do not already have that lens collection in place. I'm interested in seeing digital rangefinder photography be as accessible to as wide a range of photographers as possible.

(Yes, I'm drafting part of an article in this post. <G>)

Cheers,

Sean

sean_reid · November 23, 2006

Advertisement (gone after registration)

BTW, just for fun, here's a crop from the CV 28/3.5 on the M8. Very cloudy day so very soft light, ISO 640, hand-held, processed in C1 at default settings, no sharpening or other post work other than making this crop from the 100% file. This is a tiny lens that costs less than $300.

Cheers,

Sean

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!

sean_reid · November 23, 2006

Again, CV 28/3.5 on the M8. Very cloudy day so very soft light, ISO 640, hand-held, processed in C1 at default settings, no sharpening or other post work other than making this crop from the 100% file.

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!

Asher Kelman · November 23, 2006

Leica also tells you to be careful with astronomical radiation...
You simply can't seal rotating devices perfectly and permanently, not with simple rubber-sealing like in so-called weather-sealed "professional" cameras. Even a 1D or a D2 isn't REALLY sealed, like Leica M and R.!

Hi Georg,

No problem sealing things that rotate! No issue at all! Don't know where you have that idea. The issues I think come from modern corporations being governed by guys from that last great merger or great stock increae.

No issue discussed here is at all beyond the Leica engineers. It's just that the thought was, let's give people what they want, a Leica with a digital sensor. However, delivering anything into the 21st century has to meet the current expectations.

Asher

http://www.openphotographyforums.com

marknorton · November 23, 2006

With the firmware change I envision, there would be three options:

1. Lens detection disabled (this currently exists)

2. Lens detection enabled (this also currently exists)

3. Manual lens selection (followed by a menu of Leica lenses one can specify)

The menu selection will also require you to specify whether a filter is being used.

I have a different take on this. The imperative to do this is driven far more by the needs of JPEG users than DNG users.

Doing the correction in-camera goes against the grain of being able to work with raw data and fix things in C1. I am sure the DNG purists would rather have a tool in C1 to perform the cyan correction rather than try to fix the camera's best efforts at doing so when the filter characteristics are not as expected, the focal length is off or the user has incorrectly selected the lens type.

Given that, the DNG should contain data uncorrected for cyan and that opens the door for different plug-ins to handle different lenses and/or filters. If the meta-data were to contain the lens type (from the coding) plus a flag showing whether or not there's an IR filter in use, this could be automatic when the image is initially "developed" and then available for fine tuning.

In that case, there's no real reason to be able to select the precise lens type in use in the camera manually. It can be done in C1 instead with the coded lens giving the advantage of fast-tracking the process.

Which just leaves JPEG. Since the cyan correction only applies to 21, 24 and 28mm lenses, it would likely be sufficient to have a menu for "21, 24, 28, Other" for uncoded lenses.

scott kirkpatrick · November 23, 2006

The menu selection will also require you to specify whether a filter is being used.

I have a different take on this. The imperative to do this is driven far more by the needs of JPEG users than DNG users.

Doing the correction in-camera goes against the grain of being able to work with raw data and fix things in C1. I am sure the DNG purists would rather have a tool in C1 to perform the cyan correction rather than try to fix the camera's best efforts at doing so when the filter characteristics are not as expected, the focal length is off or the user has incorrectly selected the lens type.

Given that, the DNG should contain data uncorrected for cyan and that opens the door for different plug-ins to handle different lenses and/or filters. If the meta-data were to contain the lens type (from the coding) plus a flag showing whether or not there's an IR filter in use, this could be automatic when the image is initially "developed" and then available for fine tuning.

In that case, there's no real reason to be able to select the precise lens type in use in the camera manually. It can be done in C1 instead with the coded lens giving the advantage of fast-tracking the process.

Which just leaves JPEG. Since the cyan correction only applies to 21, 24 and 28mm lenses, it would likely be sufficient to have a menu for "21, 24, 28, Other" for uncoded lenses.

Because the DNG saves only 8 bit representatives from the full 12-14 bit original intensities, this increases the need to fix anything that can be fixed accurately before the reduction to 8 bit resolution (not "range," which remains 14 bit). So I think RAW and JPGers benefit equally.

I strongly agree that "filter on/off" belongs in the menu.

scott

scott kirkpatrick · November 23, 2006

BTW, just for fun, here's a crop from the CV 28/3.5 on the M8. Very cloudy day so very soft light, ISO 640, hand-held, processed in C1 at default settings, no sharpening or other post work other than making this crop from the 100% file. This is a tiny lens that costs less than $300.

Cheers,

Sean

Hi. it's nice to see such a nice picture from the least pretentious of the 28 mm lenses. In your review, will you compare them all in bright light as well as in soft light? Update your list of "sunny day" lenses?

scott

marknorton · November 23, 2006

Because the DNG saves only 8 bit representatives from the full 12-14 bit original intensities, this increases the need to fix anything that can be fixed accurately before the reduction to 8 bit resolution (not "range," which remains 14 bit).

Yes, good point...

sean_reid · November 23, 2006

Hi. it's nice to see such a nice picture from the least pretentious of the 28 mm lenses. In your review, will you compare them all in bright light as well as in soft light? Update your list of "sunny day" lenses?

scott

Hi Scott,

Yes, indeed (if I can find contrasty light in late November in Vermont <G>).

Cheers,

Sean

sean_reid · November 23, 2006

1)The menu selection will also require you to specify whether a filter is being used.

2) I have a different take on this. The imperative to do this is driven far more by the needs of JPEG users than DNG users. Doing the correction in-camera goes against the grain of being able to work with raw data and fix things in C1. I am sure the DNG purists would rather have a tool in C1 to perform the cyan correction rather than try to fix the camera's best efforts at doing so when the filter characteristics are not as expected, the focal length is off or the user has incorrectly selected the lens type.

3) Given that, the DNG should contain data uncorrected for cyan and that opens the door for different plug-ins to handle different lenses and/or filters.

4) If the meta-data were to contain the lens type (from the coding) plus a flag showing whether or not there's an IR filter in use, this could be automatic when the image is initially "developed" and then available for fine tuning.

5) In that case, there's no real reason to be able to select the precise lens type in use in the camera manually. It can be done in C1 instead with the coded lens giving the advantage of fast-tracking the process.

6) Since the cyan correction only applies to 21, 24 and 28mm lenses, it would likely be sufficient to have a menu for "21, 24, 28, Other" for uncoded lenses.

Hi Mark,

1) That's true. In fact, even with the coded lenses I would think we're going to have to specify whether or not a filter is mounted so I would expect that a selection for this would be needed for coded or uncoded lenses.

2) It would be important for JPEG work as well but I think we both note that Scott's point about the RAW processing in-camera is relevant. That said, I agree with your idea about this correction being available in RAW conversion and have been thinking the same thing as alternate way of making more wide lenses usable with the M8.

3) That's worth looking at.

4) Ideally, the meta data would include the lens type with manual lens selection as well. That would speed working with a large number of files (as with a wedding, for example).

5) No reason to not to have that information fast-tracked (in C1) even for non-coded lenses. The coded lenses would still be faster in the field. Also, it may be important because one cannot always remember what lens was on the camera for a given picture.

6) I've also been thinking about that and it would seem that this is only needed for 28 and wider lenses. I want to be sure of that based on some testing however.

Thanks for the thoughts.

Cheers,

Sean

pvfb · December 31, 2006

Hi

This is my first post on the a Leica forum, so apologies if what I say has been said before, but then again if somebody at Leica reads this... and takes note...

I've used Leica's for nearly 50 years (I have a picture with my father's III? aged two to prove it!) and currently have two M6s and a selection of lenses, some of them duplicated; the 35s by design and the 50s because they came with the bodies.

I've ordered an M8 and pre-purchased a coded 28 (2.8) to go with it, what I don't really undertsand is why I am being "forced" to have my lenses coded, I quote from the downloaded manual "When using a lens without 6-bit coding, the recognition function of the camera must always be turned off, to prevent malfunctions." this is going to be a real pain if you have a mix of lenses. Particularly for me as the one coded lens I have as a wide is one the camera needs to know about. Do Leica really need the £75 a time that much?

Not withstanding the technical merits in pre/post processing or even whether the imaging chip is driven differently, just knowing which lens I was using is a feature I've found iuseful in the EXIF data on my DSLRs, and I don't see that this would be a particularly complex feature to add, it's certainly a customer friendly feature.

The coding boils down to a 6 bit number ie 0-63 and I can only assume a non-coded lens will return either a 0 (or 63) or an error to the camera, so it knows it has a non-coded lens and could optionally prompt for manual input before defaulting after a few seconds.

Providing an on-camera list of every lens to scroll through would be a "bit silly" but, why not allow the user to store a list of say 6 lenses (if the screen has enough lines (making it scroll at this stage may not be that easy) and the code value or at worst enter the code value. This would mean publishing a list of codes, but I doubt they will remain secret for long

This approach could be extended to coding filter presence, or overiding it if they were for example to allow you to set a default IR on/off value for coded lenses.

Even a simple button press to overide an erroneous auto detection would be an improvement.

... and for the record I run a company that among other things makes image processing systems, so I'm not going to take any "the software engineering is to complex" **** from a marketing *******, ...... just thought I'd get that in first in case the "there is no problem with IR" department decide to comment.

Peter

pvfb · December 31, 2006