JohnL
Very confused
6 bits gives 2^6 = 64 possible different values, so what exactly is the information contained in this encoding?
I would assume focal length is one item. With the new tri-Elmar, I'd guess that the focal length options might be something like "less than 16mm", 16mm, 18mm, 21mm, 24mm (probably good for 25mm also), 28mm, 35mm (perhaps to include 40mm?), 50mm, 75mm, 90mm and maybe 135mm. This gives 11 different choices.
Max aperture might be another part of the encoded information, with options something like f/1.4 and faster, f/2, f/2.8, f/4 and slower. Lenses having a speed between the stops might be coded to the nearest full stop. This gives 4 different choices.
Combining these two items gives 11 x 4 = 44 different possibilities. This is more than can be encoded in 5 bits, and needs 6 bits, which suggests my thinking might be on the right track.
Presumably shorter and faster lenses are more prone to vignetting, and would likely need more correction in post-processing.
Does anyone *know* what the coding logic really is?
I would assume focal length is one item. With the new tri-Elmar, I'd guess that the focal length options might be something like "less than 16mm", 16mm, 18mm, 21mm, 24mm (probably good for 25mm also), 28mm, 35mm (perhaps to include 40mm?), 50mm, 75mm, 90mm and maybe 135mm. This gives 11 different choices.
Max aperture might be another part of the encoded information, with options something like f/1.4 and faster, f/2, f/2.8, f/4 and slower. Lenses having a speed between the stops might be coded to the nearest full stop. This gives 4 different choices.
Combining these two items gives 11 x 4 = 44 different possibilities. This is more than can be encoded in 5 bits, and needs 6 bits, which suggests my thinking might be on the right track.
Presumably shorter and faster lenses are more prone to vignetting, and would likely need more correction in post-processing.
Does anyone *know* what the coding logic really is?
Honus
carpe diem
John,
The 6-bit code simply identifies the lens. The firmware in the camera would then have the lens-specific data to make any needed image adjustments and generate the EXIF data. By my count, there are 45 existing and older lenses that quailfy for encoding, leaving 19 more lenses to be developed before they have to add an additional bit.
Robert
The 6-bit code simply identifies the lens. The firmware in the camera would then have the lens-specific data to make any needed image adjustments and generate the EXIF data. By my count, there are 45 existing and older lenses that quailfy for encoding, leaving 19 more lenses to be developed before they have to add an additional bit.
Robert
jaapv
RFF Sponsoring Member.
Honus said:
And that is assuming they don't use the position of the mechanical frameline-sensor as an additional 3-way "bit", making for 192 possibilities.
Gabriel M.A.
My Red Dot Glows For You
What Honus said.
JohnL
Very confused
If it identifies a specific lens, then it is a closed system, i.e. it makes no allowance for using non-Leica lenses. Ah, well ...
Maybe we can have other lens coded according to the nearest-spec Leica lens. At least we'd get the EXIF data. We can still fix any residual vignetting in post.
If they use the frameline sensor position, it will tell them "the lens in use may be 28mm or it may be 90mm, or perhaps the photog is using an external viewfinder, or even the full frame" ... doubt this is a useful option.
Maybe we can have other lens coded according to the nearest-spec Leica lens. At least we'd get the EXIF data. We can still fix any residual vignetting in post.
If they use the frameline sensor position, it will tell them "the lens in use may be 28mm or it may be 90mm, or perhaps the photog is using an external viewfinder, or even the full frame" ... doubt this is a useful option.
jaapv
RFF Sponsoring Member.
JohnL said:
Sorry, this one I don't understand. The lens will tell the camera: this is either a 35 or 90, another dot will tell exactly which. I don't think the dots or the frameline sensor convey the way the photog. is viewing the scene or if an external viewfinder is mounted. It just sets the lines to one of three positions. The dots specify the lens.This is the way the camera sees the positions of the long Tri-Elmar btw. What I mean is that the frameline sensor limits the possible focal lengths to two and then the lenscoding needs less to specify the lens type.
As to the corrections: it may be preferable to do the corrections in PS anyway, even with Leica lenses, it depends on skill I suppose. As for correcting non-Leica lenses to the nearest coded lens, it might work or might not. I think it is not just vignetting being corrected, but CA as well.
JohnL
Very confused
If the lens mounted is not a Leica M-mount lens, possibly an LTM lens with an adapter and not even an M-mount lens at all, and even more so if the photog is using an external finder, there's no assurance at all that the framelines bear any relationship to the lens on the camera. Additionally, some 40mm lenses (for example) bring up 50mm framelines, while some bring up 35mm framelines. I agree that if the frameline info could be relied on, then then it could reduce the number of bits required for additional info.jaapv said:
As an almost-100%-RAW-shooter I'd tend to agree that corrections are probably best made in post, but it would still be nice to have some lens info in the EXIF, even if its faked a bit.
sdai
Established
JohnL said:
I'm sure if you're willing to pay, someone will make a coded adapter for you.
jaapv
RFF Sponsoring Member.
I doubt that Leica will accomodate third-party lenses or other not supported lenses. The framelines are correct for all supported lenses. As any others will not be encoded there is no danger of upsetting the system.
Last edited:
JohnL
Very confused
I'm sure you're right in that Leica will not accept other brand lenses for coding, but it seems likely there will be a cottage industry for coding lenses, and this need not be limited to Leica products. In fact IIRC Jorge was planning to offer this service at one time ... not sure if that's still the case.
boojum
Mentor
Resurrecting a corpse, but here is my best guess as to what is happening: binary coding. The white dot is an off-switch, or a zero. The black is an on-switch, or a one. So, white-black-white-black-black-black is 010111. I doubt that Leica translates this to a higher number system as the firmware will read the six digits, look for a match in an internal table of these binary codes and at that entry will be all the data for the lens. Done deal. This may have been answered later. I am working my way back to front. The frameline could add two more levels to the table depending upon the frameline position. I. E., framline one for 64 entries, frameline two for 64 more entries and three for the final 64 entries.
If correcting for the nearest Leica lens works manually, and it seems to, coding should yield the same result.
I swore I would never, ever revert to computer speak.
Added later: I have coded a Jupiter 8 50mm f/2.0 as a Summicron 50mm f/2.0 ver III, Leica code 11817 (010111). The M240 I tested it on accepted it as automatic lens identification and correctly as a 50mm f/2.0 and is so noted in the EXFIF data. Winner, winner, chicken dinner. I think that answers the question. And this is not rocket surgery, if I can do it anyone on this forum should be able to do it.