Proof of concept regarding self coding of lenses

[marknorton]

Originally Posted by rvaubel

Mike

Love that bike picture! And on my monitor I don't see any discernible cyan drift or vignetting. My RD1 would have a boatload of both.

Did you use a filter?

Rex

I don't think so, look at the colour of the saddle...

[sean_reid]

Originally Posted by marknorton

Sean

1) That said, I would be surprised if CV and Zeiss coded their lenses because they do not control - or can influence - what the camera does with the coding.

2) As for Leica opening up the interface, you'll have a sense from their discussions whether this is a possibility. You've commented though on the perfectionist streak in their work and I expect making the correction generic will an anathema to them.

3) The next step for Mike's coding work could be to look at the precision of the template - so that the black marks can be placed with 100% accuracy and the use of an epoxy ink which will be much more durable, for example:

1) I wouldn't be surprised if either or both offered it as an option (if they are allowed).

2) "generic correction"? What I'm asking for is just menu access to the existing data for Leica lenses (see the discussion in the 28s article and M8 part 4). There's no discussion of generic correction at all, nor should there be.

3) The current template needs a little tweaking. Epoxy ink, eh? Hmmm...

Cheers,

Sean

[rvaubel]

Originally Posted by robsteve

Some of the correction for cyan drift may just be a correction for the IR filter on the sensor itself.

Well this is a lively, realtime thread. A question for you lucky guys with M8s . Does the M8mkI softtware/firmware already have the cyan drift correction available to coded lenses? If so, does it appear to be working?

Rex
feeling deprived in Bezerkeley

[sean_reid]

Originally Posted by marknorton

I don't think so, look at the colour of the saddle...

Right...no filter, no cyan drift but that brings us back to purple. Again, I think the CV 15 will shape up to be a good Jamie profile lens.

Cheers,

S

[sean_reid]

Originally Posted by rvaubel

Well this is a lively, realtime thread. A question for you lucky guys with M8s . Does the M8mkI softtware/firmware already have the cyan drift correction available to coded lenses? If so, does it appear to be working?

Rex
feeling deprived in Bezerkeley

Not yet, version 1.10 early December and will be downloadable. Current version has some it seems but it's not the full Monty.

S

[devils-advocate]

Short answer: they don't know if it's legal, won't risk it, but may end up being able to license the system for their own lenses.


The long answer is based on the fact that we really don't, and can't, know whether Leica has applied for a patent and what it might cover. The patent process takes a really long time. The application isn't public during most of that process, so there is no sure way to know if you infringing a potential patent unless the inventor/applicant shows you their application and explains the patent to you. Even then, you don't know if it's a valid patent until it gets granted, and even then there are always a myriad avenues of challenging the validity of the patent. There's a reason a lot of patent lawyers own Leicas To add another twist, patents are national. An EU patent is of no value in the US or Canada until it is registered here. It's pretty counter-intuitive that you can be infringing patents you don't/can't know about, but that's how it works.

In other words, Cosina or Zeiss would be taking a real risk coding their lenses in this way. They should assume Leica has some patents pending, and would have a sufficient commercial interest to enforce those patents if/when they are granted. The only safe course would be to license the process from Leica, which Leica might not be interested in doing in order to protect the competitive advantage of their lenses.

That's where something like this thread comes in the equation....if the DIY solution is effective, end-users will take the 'coding advantage' out of thier purchasing equation. Leica may then decide that making fifty-bucks per lens off Cosina's sales sounds a lot better than $0, and opt to cut a licensing deal. This is much the same mechanism by which Napster and the other file-sharing sites forced the big music companies to put their content up for sale through legal downloads on the 'net, except that the amounts at play, and the players themselves, are likely too small for anyone to get into serious multi-jurisdictional IP-litigation.

Does that make any sense?

- N.

[mike prevette]

nope no filter. I see some vignetting, but it's possibly just shell shock after the few years shooting film with this lens. I just expect it to be there.

_mike

[rvaubel]

Originally Posted by sean_reid

Right...no filter, no cyan drift but that brings us back to purple. Again, I think the CV 15 will shape up to be a good Jamie profile lens.

Cheers,

S

Your right, both the 15mm and 12mm VC superwides don't really need filters as they are both gee-wiz lenses that your not going to use alot were wedding clothing and real black sensitive subjects are the main focus. Although when you are taking a closeup of a gaggle of penquins or nuns, Jamie's profiles would come in mighty handy
Which is darn lucky because their are no filter threads anyway!

Rex

[rvaubel]

Originally Posted by sean_reid

Hi Nick,

I have no concerns about individuals using Sharpies, stickers, etc. I'm wondering if CV and Zeiss are legally allowed to code their lenses and/or adapters.

Cheers,

Sean

For the CV screw mount lenses they need to come out with an adapter that doesnt have a cut-away where you need to apply the sharpie coding. And, if they machine a recess into the face, the coding won't wear off. Ironically the screw mount lenses will be easier to DIY encoding than the bayonet models.

Rex

[sean_reid]

Originally Posted by devils-advocate

Short answer: they don't know if it's legal, won't risk it, but may end up being able to license the system for their own lenses.


The long answer is based on the fact that we really don't, and can't, know whether Leica has applied for a patent and what it might cover. The patent process takes a really long time. The application isn't public during most of that process, so there is no sure way to know if you infringing a potential patent unless the inventor/applicant shows you their application and explains the patent to you. Even then, you don't know if it's a valid patent until it gets granted, and even then there are always a myriad avenues of challenging the validity of the patent. There's a reason a lot of patent lawyers own Leicas To add another twist, patents are national. An EU patent is of no value in the US or Canada until it is registered here. It's pretty counter-intuitive that you can be infringing patents you don't/can't know about, but that's how it works.

In other words, Cosina or Zeiss would be taking a real risk coding their lenses in this way. They should assume Leica has some patents pending, and would have a sufficient commercial interest to enforce those patents if/when they are granted. The only safe course would be to license the process from Leica, which Leica might not be interested in doing in order to protect the competitive advantage of their lenses.

That's where something like this thread comes in the equation....if the DIY solution is effective, end-users will take the 'coding advantage' out of thier purchasing equation. Leica may then decide that making fifty-bucks per lens off Cosina's sales sounds a lot better than $0, and opt to cut a licensing deal. This is much the same mechanism by which Napster and the other file-sharing sites forced the big music companies to put their content up for sale through legal downloads on the 'net, except that the amounts at play, and the players themselves, are likely too small for anyone to get into serious multi-jurisdictional IP-litigation.

Does that make any sense?

- N.

Hi Nick,

It certainly does and it is quite interesting. Thanks.

Cheers,

Sean

[sean_reid]

Originally Posted by rvaubel

For the CV screw mount lenses they need to come out with an adapter that doesnt have a cut-away where you need to apply the sharpie coding. And, if they machine a recess into the face, the coding won't wear off. Ironically the screw mount lenses will be easier to DIY encoding than the bayonet models.

Rex

You missed my post on this. Already at work on getting some "no cutaway" adapters to test. May be soon.

Cheers,

S

[KJB]

I'm no IP expert but I believe it takes 18 months from first filing for a patent to publish. I did a quick patent search and was not able to find a patent by Leica Camera relevant to the coding of their lenses. I'd imagine the latest that we should see something publish would be 18 months from the first public disclosure by Leica on the 6 bit coding of lenses (whenever that was) as I can't imagine they would have filed after the announcement.

[rvaubel]

Originally Posted by sean_reid

You missed my post on this. Already at work on getting some "no cutaway" adapters to test. May be soon.

Cheers,

S

I missed one of your posts? The shame!

Anyway, if you mill some circlular depressions where the coding goes, the user can color them in with the appropriate lens coding. One size fits all!

Rex

[Riley]

Originally Posted by devils-advocate

The problem with patenting is that the exact nature of what the patent purports to protect depends entirely on the wording of the claims in the patent application. Unless and until we see Leica's application relating to this system/process, we have no idea what they might have tried to protect (the detection system, the code itself, the process of coding the lens - and how that is described, etc., etc). Fwiw I couldn't find anything that looked like an applicable patent on either the US of EU patent sites -- it takes many months after an application is received for these even to be posted.

If it calms a few minds, I'd assess the chance of Leica taking legal action against individual users (assuming they ever get a patent granted) at Z-E-R-O, for all the common-sense reasons that would likely come to every non-lawyer's mind. While a commercial infringement might be a different question,we can all relax and go back to our sharpies.

generally speaking, legal patent provisions allow people to use the quoted patent material for their own use. The legal innuendo begins when people attempt to make money from the patent material. In other words, there is no case for a legal engagement between Leica and Leica users provided that they do not attempt to profit from it.

Edit to Add: carry on

Riley

[monza]

I don't think there is any worry that a numbering system can be infringed upon, either via copyright or patent.

Such a numbering system for lenses isn't really patentable. It's nothing more than assigning a number to specific product, the same thing as UPC code is assigned to a specific product. Figuring out which products and which codes match, doesn't infringe anything. Neither does utilizing this information in one's products.

In a previous business of mine, we commonly reverse engineered binary codes that were much more 'hidden', so to speak, than this lens numbering system. Some of these binary codes were revealed by adjusting DIP switches, some were stored in EEPROMs, some in firmware. Many were difficult to deduce, and all required extensive testing to confirm.

We had to know most, if not all, of the combinations for our product to work in the host computer system...all the way from 0000 to 1111, for example.

If Zeiss or Cosina figured out Leica's coding system there is little Leica could do. There is no inherent way to protect a simple numbering system.

What is protectable (via copyright or perhaps patent) is Leica's software image correction algorithm. Offering up a simple 6 bit number that this software utilizes, doesn't mean someone is duplicating or shipping products with the same software.

Patents for electronics are now taking about 3 years in the US.

[Riley]

patents are about protecting the rights of owners of ideas
copyright is about protecting the rights of the owners of material
registered design is about protecting the owners for the use of a design
licenses are how organisations trade rights

you will find that olympus also have a coding system, and this somewhat precedes leicas

there isnt a discussion in this, provided no one tries to make money from it, people are free to use the system. any other way of working things would see research and development of almost everything shut down

Riley

[mr_big]

Here I am with a 75 cent Sharpie marking up over 8k worth of camera equipment to make it work... seems kinda strange if you ask me. Thanks to Mike and the others on this thread I have sucessfully coded my 35mm and 90mm lenses, but I still have a question: Has anyone done any testing to see if this makeshift coding actually works? What if the black marks only do part of the job, like bringing up the lens info, and the white marks do the actual vignette corrections? So far we've all only been dealing with the black and leaving the other spaces blank.

[chetccox]

There are only 2 possibilities for the lens sensors, on or off (1 or 0) If black triggers the sensor to one of these states then white cannot do so as well. So the answer is no, the white codes make no difference.

[mr_big]

Thanks Chet, that make sense. So it looks like you guys have saved me $250.00. Now I'd like to see what Sean has up his sleeves regarding the CV mount for my 28mm 1.9.

[scott kirkpatrick]

Originally Posted by rvaubel

Well this is a lively, realtime thread. A question for you lucky guys with M8s . Does the M8mkI softtware/firmware already have the cyan drift correction available to coded lenses? If so, does it appear to be working?

Rex
feeling deprived in Bezerkeley

Rex, I'm equally deprived in Jerusalem, but Sean's articles provide grist for a little quantitative analysis, which I plan to continue in a new thread once he is comfortable with the comparison of vignetting effects with and without the firmware enabled. Since he now has all or most of his lenses pretending to be appropriate Leicas, this will probably take some more shooting time.

There is no simple rule for predicting vignetting (in response to your earlier post #159) or its corrections. In the M8, the offset microlenses reduce overall vignetting dramatically, before any firmware tricks come into play. They may even increase brightness at the edges for focal lengths 50mm and longer above what the same lenses would see at the center of the frame. There are two places where a radial color shift occurs. The M8 cover glass is a photometric filter -- it reduces red to where a Si sensor will aproximate human sensititity and it reduces IR even more (but not to zero!). Towards the edges, light entering at an angle sees a thicker filter. Same color response, but more of it. That also contributes to a green tint at the borders. I would expect the existing M8 firmware to correct for both of these effects, but not for an IR-cut filter on the lens, which is a new thing. Looking at some of the test shots, the existing firmware seems to help in the mild cases, but the effect on color may be mostly just brightening the edges.

scott