Using Leica M lenses on the Panasonic G1 - some pictures

[sean_reid]

Advertisement (gone after registration)

I agree but would go further - it enhances the value of any M mount lens when you don't have to guess at its focus or rely on a rangefinder which may, or may not, be doing its job properly.

 

I'm afraid the G1 makes the Leica M focussing method seem well past its sell-by date.

 

Hi Mark,

 

Well, I would agree that live view can be useful for accurate focusing (WYSIWYG) but I personally *vastly* prefer the traditional rangefinder system for speed, the view through the finder, etc. There's no way an EVF could ever replace a rangefinder window for me. I realize other people, with other working styles, may actually prefer an EVF.

 

I've been using the R-D1 and M8 intensively, since 2004, for professional work and have very, very rarely had a mis-focused picture. And when that has happened, it has been my fault and not the camera's. I can't say the same for any auto-focus DSLR I've used.

 

I'm excited about the possibilities implied by the G1 and very interested in the camera but by no stretch of the imagination would I consider using it instead of the M8 for my own work.

 

Cheers,

 

Sean

[marknorton]

Yes, but you can imagine an optional mode where the horizontal mis-alignment of the rangefinder image is not set by the existing mechanical/optical method but instead by the camera's evaluation of actual focus at the sensor. Same viewfinder experience but when the images were coincident, you'd know you were getting the same result as conventional AF.

 

That said, I agree with you about DSLR AF; if you have 51 sensor elements, it will grab focus on pretty much anything it can find. Spot AF is best, IMHO.

[Advertisement]

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

[TimF]

f you have 51 sensor elements, it will grab focus on pretty much anything it can find. Spot AF is best, IMHO.

That's always been true of AF though hasn't it? As I recall the last generation of Nikon pro bodies didn't always have the actual AF sensors correctly aligned with the visible points on the focusing screen, which helps very little. During the brief time I had an F6 it didn't always hit the spot even in "spot AF".

 

As for the M8 focusing I agree with Sean. Maybe I've been helped by not having jumped in with the M8 when it first came out, but I've not had any focusing issues to date that weren't my fault. The sole exception to that perhaps was with my Canon 50mm f/1.2, but wide open the DoF on that lens is miniscule, and it's pretty soft at max aperture which doesn't help either! Doing a test with the camera on a tripod at about 5 ft distance proved there's no problem with it though.

[monza]

My eyes are good, but unfortunately not good enough to focus an M8 reliably with fast glass. An M3 gives me far more accuracy.

 

Since a large portion of my photos are shot at wide aperture, focus is critical. If I were a street shooter and used a wide angle at f/8 all the time, of course any RF would be fast and easy and the view thru a bright optical VF in such situations is unbeatable.

 

The ability to focus M lenses TTL is a huge advance, IMHO.

 

I agree with Mark: have a computer generated RF patch that is superimposed onto an optical viewfinder. The patch could even change intensity or color when the computer matched bits and focus was hit. Have an EVF and articulating LCD and every option is covered.

 

And of course focus shifts and back/front focus are then history.

[sean_reid]

Yes, but you can imagine an optional mode where the horizontal mis-alignment of the rangefinder image is not set by the existing mechanical/optical method but instead by the camera's evaluation of actual focus at the sensor. Same viewfinder experience but when the images were coincident, you'd know you were getting the same result as conventional AF.

 

That said, I agree with you about DSLR AF; if you have 51 sensor elements, it will grab focus on pretty much anything it can find. Spot AF is best, IMHO.

 

Hi Mark,

 

I use only the center focus point with all AF cameras (8 years now) and they still miss once in awhile.

 

As for your other idea, I was just thinking about that today. I was wondering if an AF type sensor system could be used to align one small rectangle (lit by LEDs or?) with another (in the middle of a window finder with frame lines). So the focus would still happen by superimposing the two rectangles but the method driving that would be different.

 

I'm also curious about LED (or ?) projected frame lines. If you have any ideas about both of these I'd love to hear them. Also, do you have any ballpark sense of how expensive it might be for a (lets say large Japanese) company to develop and manufacture these two systems? Expensive....not so much? Technology is your thing professionally, yes?

 

Cheers,

 

Sean

[sean_reid]

My eyes are good, but unfortunately not good enough to focus an M8 reliably with fast glass. An M3 gives me far more accuracy.

 

Since a large portion of my photos are shot at wide aperture, focus is critical. If I were a street shooter and used a wide angle at f/8 all the time, of course any RF would be fast and easy and the view thru a bright optical VF in such situations is unbeatable.

 

The ability to focus M lenses TTL is a huge advance, IMHO.

 

I agree with Mark: have a computer generated RF patch that is superimposed onto an optical viewfinder. The patch could even change intensity or color when the computer matched bits and focus was hit. Have an EVF and articulating LCD and every option is covered.

 

And of course focus shifts and back/front focus are then history.

 

Focus shift could still be a problem but it might eliminate some other focus problems.

 

Cheers,

 

Sean

[howard_cummer]

Advertisement (gone after registration)

Hi Mark,

I so much agree with your further comment! Imagine having exact framing for all your Leica lenses and additionally accurate focus confirmation! Despite adjustment at the local Leica agent's repair shop my 90 Cron APO still back focuses on my M8 (which has also had upgraded framelines and the view finder adjusted). But on the G1 with the magnified live view the 90's focus can be confirmed beautifully. Also, I use my 135 Elmar on the M8 guessing that the field of view is slightly bigger than the rangefinder patch. There is real pleasure in putting the Elmar on the G1 and seeing exactly what the framing is. Also, since the viewfinder brightens as you stop down you can effectively focus at the taking aperture without difficulty up to f8 - f11 in good light. These are several pluses for using the G1 over the M8 for lenses outside the 24mm - 90mm frameline range provided by the camera. I have been using the G1 for several days now and am warming to it very much. Haven't used my M8 once since the G1 came to hand.

Cheers

Howard

[marknorton]

If you look through an M8 and block off the rangefinder and illuminator windows, all you see is a clear view ahead. What the M8 then does is to inject the framelines, rangefinder patch and viewfinder information into the viewfinder using light through the illuminator and rangefinder windows.

 

What I'd be interested in seeing is a G1-type EVF doing that job - the screen would be mostly black with white "openings" for the frame lines and a small patch replicated from a live view sensor. Because you now have a tiny video screen, the options for what you display and where you display it are simply down to the software you write - the framelines could be sized and positioned based on the actual lens mounted and the focus distance, still communicated back to the camera through the existing focussing cam. What's displayed in the viewfinder could now change according to the user's preference and operating mode, including, for minimalists, nothing. Finally, the central portion of the sensor output could be positioned in the viewfinder to provide the equivalent of the rangefinder patch moving as you focussed and coincident when correct.

 

That's the theory, lots of potential issues!

 

First, the Leica rangefinder optics provide a rangefinder patch image which is always in focus itself, short focal length, small aperture, big depth of field. Not so close up, try holding a rule in front of the rangefinder window. With this alternative approach, the "rangefinder patch" would be just as out of focus as the main image, so as you focussed, you'd see the rangefinder patch not only moving side to side/changing colour (or whatever choice could be made) but also going in an out of focus itself which you don't see now. That might provide additional rangefinder "snap"...

 

The sensor is viewing the world through a stopped down lens and while it can turn up the gain, the rangefinder patch would be noisy in stopped down or low light conditions. The rangefinder patch (which is in reality, tiny) might not be very high resolution.

 

Since the camera would be replicating a mechanical shift, there would be alignment issues - getting the rangefinder patch to be injected into the correct place when in focus but this would be possible for the user to do, similar to setting eye dioptre correction.

 

There would be packaging issues and increased power consumption, though the G1 method of sensing when the camera is put to the eye is a masterstroke. Add the wide-range built-in diopter correction in the G1 and you do think EVF has come a long way since the Digilux 2.

 

Then, there's the thorny issue of who is evaluating focus. On the G1 EVF + MF, it's me, with this method, it would effectively be the camera AF which as Sean points out might get it wrong. Would we simply be swapping one error prone system (misaligned rangefinder) for another (imperfect AF)?

 

Finally, would the complete thing convey the sense of qulity required or would it simply look and feel very contrived?

 

So, lots to think about to maintain the traditional rangefinder approach but surely worth it if it eliminates front and back focussing. As Sean points out, there would still be focus shift but since you would be focussing at working aperture, this would be covered as well.

[sean_reid]

Thanks for the ideas Mark. You wrote:

"If you look through an M8 and block off the rangefinder and illuminator windows, all you see is a clear view ahead. What the M8 then does is to inject the framelines, rangefinder patch and viewfinder information into the viewfinder using light through the illuminator and rangefinder windows."

 

Yes, of course. I wonder about the electronic options specifically because I'm interested in the idea of an MFT camera that would:

 

- present the subject via an true optical window with frame lines (perhaps electronic)

 

- offer an electronic rangefinding system that was visually analogous to the current superimposed RF patches

 

Such a camera could potentially offer AF with MFT lenses and MF with both MFT lenses and traditional RF lenses.

 

I'm also wondering if, once the R & D was paid for, such a camera might be less expensive to produce than one with a traditional mechanical rangefinder. I very much like the mechanical RF system in the M7, M8 etc. but am interested in seeing a more affordable DRF (or similar) come to the market.

 

Regarding focus shift, I wonder how accurate an AF system could be if it was evaluating focus based on, say, an image from a CV 15 at F/5.6. I'd need to learn more about AF - how would the increased DOF affect RF accuracy I wonder.

 

Cheers,

 

Sean

[carstenw]

Rather than an EVF, I would *much* rather that the three rangefinder adjustments became invokable by a menu item, which used contrast measurements on the sensor, on a per-lens basis (coding necessary, probably), to find the best settings for each lens, so that the optical rangefinder would always be perfectly adjusted...

[TEBnewyork]

Sean,

I know you want a truly optical view but wouldn't an easy compromise be an electronic view with a sensor that is a little bit larger than what is recorded so that you can see outside the frame of your "live view".

 

While I also like the live view, and the "freeze" in the G1 viewfinder when you take the shot is a bit startling after using the M8, what I don't like about rangefinders is the ratio of what is in vs. out of the frame on longer lenses. That is partially why I sold my 90mm lens and stop at 75mm.

[Hacker]

I'm afraid the G1 makes the Leica M focussing method seem well past its sell-by date.

 

Yeah, who needs framelines that have focusing either optimized for near of far distances.

[sean_reid]

Sean,

I know you want a truly optical view but wouldn't an easy compromise be an electronic view with a sensor that is a little bit larger than what is recorded so that you can see outside the frame of your "live view".

 

While I also like the live view, and the "freeze" in the G1 viewfinder when you take the shot is a bit startling after using the M8, what I don't like about rangefinders is the ratio of what is in vs. out of the frame on longer lenses. That is partially why I sold my 90mm lens and stop at 75mm.

 

Hi Terry,

 

You should try the new frame lines - they're much better in that respect. Unfortunately, that EVF arrangement wouldn't even come close to what we have in a traditional rangefinder camera. I think there will be a place for an affordable DRF (perhaps in a slightly new form) that uses a true optical window finder and, in fact, I've been in discussions with a manufacturer who is open to some of these ideas.

 

A good optical window finder:

 

1. Is like looking through a window, not at a tv screen (as I first wrote about when I reviewed the Digilux 2).

 

2. Shows all distances in focus - near or far.

 

3. Remains bright even in low light

 

4. Does not show electronic artifacts

 

5. Shows the view both and outside the frame lines

 

An EVF could give us number 5 and may get better with respect to 3 and 4. But it will never give us number 1 or 2 and those have been important to certain photographers for a very long time. They're integral to certain ways of seeing the subject, composing, etc. This is one reason I use SLRs only when I have to.

 

Cheers,

 

Sean

[peterb]

In the M6 the finder projected a pair of red LED's you 'balanced' to get the proper metering.

 

I wondered whether an EVF could not project 'frame lines' in a similar manner onto a view and and offer either the matched diode metering as the M6 had or a projected patch for manual focusing (or quite possibly an electronic version of the rangefinder's focusing aid) along with a focus assurance light (as Nikon and Pentax do with manually focused Zeiss lenses made for their marques) when focus has been achieved. It might be a bit cluttered but it may achieve what you've been asking for.

 

Peter

[marknorton]

Oh dear, it's clear I'm not making myself clear...

 

I'm not suggesting for one minute replacing the M viewfinder with an EVF, simply using one to inject those elements into the view which are currently done using the frame lines, the red LED display and the rangefinder patch. The main viewfinder would be straight-through, as currently.

 

Once you have the ability for the user to see the main view with the information you choose from programming to put on the EVF superimposed, you can emulate the existing viewfinder but improve it by making the framelines much more accurate, provide additional information to the user and create a rangefinder patch sampled from the centre of the live view sensor and shift it laterally to indicate out of focus.

 

Based on Sean's 5 criteria, that would give #1 and #2. #3 could be handled by adjusting the EVF back-light brightness to track the brightness coming in through the viewfinder as it does today where the framelines, rangfinder patch and red LED are all adjusted for brightness based on light coming from the same general direction as the the viewfinder is looking. #4 is a currently a problem because the EVF is a strobed device and you see the red, green and blue lines and your persistence of vision keeps them "displayed". #5 is met because the view forwards is unchanged.

 

Specifically, you can see here the front of the viewfinder with the angled internal reflecting surface. The brass carrier holds the framelines and light coming in through the front of the camera is reflected off the curved mirror through the slits in the frameline masks into the prism where it is superimposed on the viewfinder image. The rangefinder patch and LED display are handled similarly.

 

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!

 

So why not replace the brass carrier with the EVF? Whatever light the EVF emits is handled just like the red light from the LEDs now and is superimposed on the view forwards.

 

If the rangefinder patch is now a sample from the real live view sensor, it will show actual focus, through the lens which is the overriding advantage of the G1 arrangement over the M rangefinder.

[farnz]

Mark,

 

Quite ingenious!

 

But wouldn't there be a difference in scale depending on the focal length of the lens on the camera? The image on the sensor will be naturally magnified by a longer lens but the optical element of the VF will not, so how would the two images correlate or line up?

 

Pete.

[marknorton]

Pete

 

Indeed, and the size (number of pixels x and y) of the sample taken from the live-view sensor would change according to the focal length of the lens, possibly shooting distance as well since the focal length changes with shooting distance. That sample would then be scaled to match the magnification of the optical viewfinder.

 

The camera would need in any case to sense the position of the rangefinder cam for parallax correction to a good level of accuracy - 3 mm movement and 8 bit accuracy at least, would be my guess.

 

Then you get into the issue of legacy lenses - a lot of work, mainly software but pointless if it involves a lot of complex adjustments, a case of swapping one can of worms for another.

 

Might be that mimicking the current rangefinder patch is too much work and/or the results don't justify it, so that a focus assist arrows and dot might be an alternative.

[jank]

Front or back focus certainly could be corrected making the rangefinder a near perfect focusing system..In electronic age it could be done as follows:

one has to know which lens is on. Since focusing error varies from lens to lens, the e.g. ser. number could identify the particular one.

The finder mechanics could be made fine adjustable by use of micromotor, perhaps by piezodevice.

Each lens , if desired could be individually calibrated , position of motor memorized and correction then applied.Should the aperture setting be known the focus shift could be corrected easily.

This approach might take care of finder error and focus shift.In addition the parallax could be also corrected usng additiona piezo device.Power consumption of such circuitry is certainly minimal, small microprocessors consume only few uA and piezodevices or micromotors would run only during focusing or aperture change.

I think, eventually all lenses will be individually electronically identified.

The older ones will have coarser corrections, the new ones would naturally be near prefect.

Jan

[carstenw]

Ah, so that's what you meant! I saw a bit surprised before you cleared it up. Yes, this sounds like exactly what we have discussed earlier, but with more details. Do you really think it is feasible without increasing the cost significantly, and not losing any accuracy? That would be great, especially if combined with some kind of per-lens tuning in firmware.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[marknorton]

Yes, we discussed using an LCD to create adjustable framelines before and I've banged on for a while about the camera needing to sense actual focus instead of the parallel universe of the rangefinder and focussing cam.

 

I still come back to the question of how well the camera can determine correct focus (and then make the two images coincident) rather than the user but I don't see an easy way of having the user evaluate actual focus within the viewing scheme we're keen to preserve. I can see real issues for example, with the camera trying to determine actual focus in marginal light and marginal subjects.