Living with an M8 and M9 - the filter question

[jonoslack]

Advertisement (gone after registration)

Since the cyan drift (red vignetting) depends on the angle at which light passes the filter, it is difficult to imagine what difference the lens construction could possibly make. On the other hand, two nominally 35 mm lenses might have a slightly different actual focal length which might account for some slight difference in the amount of cyan drift observed.

 

Hi There

I would have thought that it was the vignetting that varied between the different lens construction - not the angle of light, and that it's the vignetting which controls the amount of cyan drift.

[mjh]

I would have thought that it was the vignetting that varied between the different lens construction - not the angle of light, and that it's the vignetting which controls the amount of cyan drift.

This isn’t your usual kind of vignetting that is caused by the lens – that kind of vignetting applies to all colors uniformly. Red vignetting is caused by the dichroic filter.

[jonoslack]

This isn’t your usual kind of vignetting that is caused by the lens – that kind of vignetting applies to all colors uniformly. Red vignetting is caused by the dichroic filter.

 

Of course, I understand that - but the amount of red vignetting will be related to the vignetting as a whole - and, of course, any correction which may be made in the firmware.

[stunsworth]

Since the cyan drift (red vignetting) depends on the angle at which light passes the filter, it is difficult to imagine what difference the lens construction could possibly make. On the other hand, two nominally 35 mm lenses might have a slightly different actual focal length which might account for some slight difference in the amount of cyan drift observed.

 

Michael, have a look at this shot of mine taken with an uncoded 35mm pre-ASPH Summilux with filter, and compare it to the uncoded aspherical Summilux a couple of posts below...

 

http://www.l-camera-forum.com/leica-forum/leica-m8-forum/95528-m9-some-more-reliable-news-2.html#post1002768

 

These are nominally at least the same focal length, and yet my lens doesn't exhibit the problems that the other one does.

[jaapv]

Since the cyan drift (red vignetting) depends on the angle at which light passes the filter, it is difficult to imagine what difference the lens construction could possibly make. On the other hand, two nominally 35 mm lenses might have a slightly different actual focal length which might account for some slight difference in the amount of cyan drift observed.

I agree, Michael, as I wrote, contrary to theory. It is just a phenomen I observed. Maybe the optical vignetting on the CV lens masked the cyan drift?

[mjh]

Michael, have a look at this shot of mine taken with an uncoded 35mm pre-ASPH Summilux with filter, and compare it to the uncoded aspherical Summilux a couple of posts below...

 

http://www.l-camera-forum.com/leica-forum/leica-m8-forum/95528-m9-some-more-reliable-news-2.html#post1002768

 

These are nominally at least the same focal length, and yet my lens doesn't exhibit the problems that the other one does.

Comparing these two images reminds me of a point I forgot to mention: Apart from slight differences in the actual focal length, focus plays a role as well. When the corners of the image are in focus, then the incident angles of light rays eventually hitting some point within these corners will show very little variation. If, on the other hand, the corners are out of focus, these parts of the image will receive light that hits the filter at more varied angles.

[jaapv]

Advertisement (gone after registration)

Aha!

[adli]

Comparing these two images reminds me of a point I forgot to mention: Apart from slight differences in the actual focal length, focus plays a role as well. When the corners of the image are in focus, then the incident angles of light rays eventually hitting some point within these corners will show very little variation. If, on the other hand, the corners are out of focus, these parts of the image will receive light that hits the filter at more varied angles.

 

I am not following you. How can focus affect the angle at which the light hits the filter when the filter is in front of the lens?

[earleygallery]

I've just had a eureka moment! Rest assured if Leica don't have the answer I do. Patent pending.....

[mjh]

I am not following you. How can focus affect the angle at which the light hits the filter when the filter is in front of the lens?

Focus doesn’t change the incident angle of light rays hitting the filter, but it does determine which of these rays will eventually hit some sensor pixel.

 

By definition, if an image is in focus, then all the light rays hitting some point in the image plane will originate from the same point in the scene. Conversely, if an image is out of focus, a point in the image plane will get hit by light rays originating from different parts of the scene and thus hitting the filter at different angles.

[adli]

Focus doesn’t change the incident angle of light rays hitting the filter, but it does determine which of these rays will eventually hit some sensor pixel.

 

By definition, if an image is in focus, then all the light rays hitting some point in the image plane will originate from the same point in the scene. Conversely, if an image is out of focus, a point in the image plane will get hit by light rays originating from different parts of the scene and thus hitting the filter at different angles.

 

OK, then I get it. Out of focus parts might get light from a wider angle and thus are more prone to get cyan drift.

[BerndReini]

Michael, I believe it would be possible that the lens design influences cyan drift since the IR spectrum of light behaves differently. For example, when you shoot pure infrared images, you have to experiment where your plane of focus is instead of relying on the rangefinder. I would think that even the coding of the lens could influence it. In the other thread, someone suggested as well that the two aspheres in the first summilux could increase the cyan drift. The old lux renders colors somewhat differently from the new design.

[jaapv]

Bernd, that is right, and many lenses exhibit a hot spot at certain apertures. That would be lenses with less IR light towards the corners. And probably less cyan drift.

[chris_tribble]

Thanks everyone - I feel clearer though not necessarily happier. I think that a period of filter swapping might lie in wait for me... Heigh ho. Let's hope we don't face a nightmare scenario of having to use different strengths of IR filters with different bodies!!

 

The horrid prospect is that of running a cropped sensor M in parallel with full sensor - I did this for a bit with 1D mk 2 and the 5D - it wasn't comfortable. But two M9's in one year? Ouch... If any of the price guestimates are right (and it's inevitably going to cost more than the M8/2 did when it came out (that was €5000???) it looks like I may have to sell body parts...

[mjh]

Michael, I believe it would be possible that the lens design influences cyan drift since the IR spectrum of light behaves differently.

Sure, but behind the filter, there shouldn’t be any IR to speak of, so differences in the way the lenses handle IR shouldn’t make a difference. The reason for the cyan drift is that for some incident angles, the filter isn’t just blocking (reflecting) IR, but also some red light.

[lxlim]

Thanks, Chris for bring up this thread. Very useful.

 

Thanks Michael for sharing.

[Goldie]

I am going to give my opinion on the OP's question by coming in from a different angle.

Firstly we are assuming the M9 is full frame and it uses the Maestro chipset: I think the second assumption is a given, since Leica have publicly said Maestro will be their chipset of choice in all their German made cameras, the first assumption is based on our hopes and three years of innovation in the sensor industry.

Given the power of the Maestro chip I think that Leica will have tried to discard the "embarrassment" of the IR filter by three improvements; better micro-lense implementation, enhanced cover glass filtration and last but not least, in camera processing, made possible by the capacity of Maestro.

If this were the case then IR filters would become redundant with the M9.

However by giving users a choice as to the Maestro's filtration function, Leica could give us seamless lense changeover either from a film M or an M8.

[hoppyman]

Steve in my opinion that is not a conclusive comparison. You would need to shoot the same shot identically with each lens. The white wall in the second shot also serves to make the effect more obvious. Differences in white balance and exposure are also involved. It would be easy to construct a worst case comparison with a plain white surface. Of course as a practical matter, the only thing that is important is if you are satisfied for yourself in what your own lens does.

Michael, have a look at this shot of mine taken with an uncoded 35mm pre-ASPH Summilux with filter, and compare it to the uncoded aspherical Summilux a couple of posts below...

 

http://www.l-camera-forum.com/leica-forum/leica-m8-forum/95528-m9-some-more-reliable-news-2.html#post1002768

 

These are nominally at least the same focal length, and yet my lens doesn't exhibit the problems that the other one does.

[stunsworth]

Steve in my opinion that is not a conclusive comparison. You would need to shoot the same shot identically with each lens. The white wall in the second shot also serves to make the effect more obvious

 

Geoff, I agree there's no controlled experiment in the comparison. I'm about to go to bed, but I'll look for an example with a white background in the morning (morning in the UK that is). I've never seen cyan corners when using this lens with a filter whatever the background and aperture, I'll take a closer look in the morning.

[Advertisement]

Hallo,

Look here: Leica M

[doug_m]

I like having some sort of filter on all my lenses. I also have all the IR/UV filters I will need for the foreseeable future and do not shoot film. So why would I want the M9 to not need filters- especially if I will be keeping my M8 and M8.2?

 

 

PS I would trade one of them in but fear the trade in amount will be extremely low.