M9 and Lens Coding

[gwelland]

Advertisement (gone after registration)

Since there seem to be so few M9 threads .... Here's a starter for ten ...

 

If we assume that the IR issue has been addressed with the anticipated camera, I'm inclined to believe that the need for coded lenses is likely to be even more important with a FF M9.

 

If we assume that vignetting is going to be fact of life with lens/sensor coverage, given that it happens today on film anyway with certain lenses, then I expect that it will be even more critical for the camera to know the characteristics of the lens fitted to render clean raw files.

 

Whilst today it doesn't much matter if you code lenses beyond 35mm, I suspect that all lenses will need coding for corrected output files.

 

My $0.02.

[Advertisement]

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

[lars_bergquist]

The importance of vignetting (as distinct from cyan corners) does decrease with increasing focal length. I do doubt that vignetting will be much of an issue with lenses longer than 50mm, because, like the M8, the M9 will apply some basic vignetting correction to all lenses, coded or not, and more correction only to lenses that need it. But I may be proven wrong in a week

 

The old man from the Age of the M8

[marknorton]

The old man from the Age of the M8

 

Excellent!

[adan]

Regarding vignetting (NOT cyan), I would expect wideangles to need correction more than they did on the M8, even though I'm sure Leica has addressed that in part through sensor design (with Kodak).

 

It may be necessary to have coding for the longest M lenses now, as well, since at some point the microlens offset that helps wideangles may start to have a shading effect (overcorrection) for 90's and 135s. Maybe only just the 135s. Maybe none.

 

50-75s are probably safe from vignetting, unless one wants even better results than on film (I don't care, myself - I usually end up ADDING some corner darkening to most shots anyway, just for that 'filmic' look).

 

In the end, we'll probably have to just wait and see what recommendations Leica makes (yeah, I know - "get everything coded"), and do some experimenting ourselves.

 

I'll be interested to see some results from, say, a 21 Super-Angulon, or a Visoflex + 400 - or the 15 c/v or WATE.

[pascal_meheut]

We'll see when the camera is here. The DMR was doing in-camera cyan vignetting correction with ROM lenses because the impact of light angle on the IR filter applies after the lens too.

 

But as the M9 sensor is more recent and as it seems that the IR filter is after the microlenses, maybe no cyan drift will occur at all.

 

Of course, vignetting is another story. But if coding is not really necessary, Zeiss and Voigtlander lenses will be more attractive than ever.

[gwelland]

As mentioned, I was referring to normal vignetting due to the light drop off, not cyan corners. I'm assuming that Leica have solved the IR problem and that UV/IR filters on the lens will no longer be necessary. That's a BIG plus if it happens.

[sandymc]

Advertisement (gone after registration)

I think its probably important to note that the cyan corners problem, and need for an external filter are separate issues. You can and will get cyan corners regardless of where the IR filter is. It's easy to put an IR filter on the sensor. Its difficult to have that and microlenses (ones powerful enough to avoid regular vignetting anyway) thin enough to prevent spherical distortion with wide angle lenses. To do that any not have cyan corners - well, lets just say my prediction is that assuming the Japanese site leak is correct, and the M9 has an internal filter, the M9 will have substantial vignetting and cyan corners without coded lenses. In fact, it would not surprise me at all to find it operates in "cropped sensor mode" only with some wides.

 

Sandy

[lct]

R wides vignette a lot at full aperture on FF DSLRs but i've never seen cyan corners with the latters so far. Are M lenses different from this standpoint?

[rebelfocus]

Note that for Auto ISO to work in the mode where the focal length is used to decide on the slowest shutter speed before bumping the ISO up then your lens would need to be coded to make this feature work.

 

I would also expect knowledge of the focal length to help in the vignetting removal process (it could probably be done anyway but would take more processing).

[jaapv]

R wides vignette a lot at full aperture on FF DSLRs but i've never seen cyan corners with the latters so far. Are M lenses different from this standpoint?

But most don't use DSLR lenses with UV/IR filters on them.

[jaapv]

R wides vignette a lot at full aperture on FF DSLRs but i've never seen cyan corners with the latters so far. Are M lenses different from this standpoint?

As for sensor filtering, the (in)famous incidence angle of M lenses is the explanation.

[lars_bergquist]

There will probably be a fly somewhere in the ointment.

 

The old man from the Age of Flypaper

[john_newell]

Wholly apart from whether the M9 needs coding for any purpose, the physical coding upgrade is simple and relatively inexpensive and, for me at least, having the camera recording the lens info in the exif data is a huge improvement over trying (and usually failing or goofing up) taking notes about which lens/aperture/etc. That obviously doesn't make for better pictures but it does make for better organization and I need all the help I can get.

[jaapv]

There will probably be a fly somewhere in the ointment.

 

The old man from the Age of Flypaper

I advise Eclipse 2 and Sensor Swipes #1

[lars_bergquist]

I am already keeping these items ready.

 

The old fly on the flypaper

[rfleica]

i don't plan to sell my m8 and would hate to keep changing filters. i hope the new firmware allows an "IR filter on" option to avoid colour drift in the corners on the m9.

[gwelland]

I'm inclined to agree with Sandy's assertion that Leica will have had to jump through some pretty significant loops to get FF working properly along with IR filtering. It will be interesting to see how that image processing works out.

[Bo_Lorentzen]

Coding makes sense, seems like one should just continue conding all lenses, as it ads datapoints to lightroom.

 

and I too hope M9 have a filter on mode, so I can leave the UV-IR on the lenses and swap to the M8 seamlessly.

 

.

[Advertisement]

Hallo,

Look here: Leica M

[carlmuck]

I think its probably important to note that the cyan corners problem, and need for an external filter are separate issues. You can and will get cyan corners regardless of where the IR filter is. It's easy to put an IR filter on the sensor. Its difficult to have that and microlenses (ones powerful enough to avoid regular vignetting anyway) thin enough to prevent spherical distortion with wide angle lenses. To do that any not have cyan corners - well, lets just say my prediction is that assuming the Japanese site leak is correct, and the M9 has an internal filter, the M9 will have substantial vignetting and cyan corners without coded lenses. In fact, it would not surprise me at all to find it operates in "cropped sensor mode" only with some wides.

 

Sandy

 

Now I've got to think about science, if the IR filter is an interference type, then cyan "shift" (actually loss of red, correct?) occurs due to the light's angle of incidence being high enough that the longer wavelengths (red's) can't make it through the notch. Is that correct? But the filtering at the sensor SHOULD exhibit less shift, since the light is, presumably, focused at that point?

 

So a highly corrected lens (say a 21mm summilux) that can arrange to get the light coming in "more perpendicular" at the film/sensor plane will need less filtering than an uncorrected one (e.g., CV 15).

 

carl.