A Small Sigma 180 f/2.8 OS Macro Issue

[Ellen Schmidtee]

If Canon makes a good lens, it will sell. If the competitor is better, Canon needs to "protect their market".

That would stifle competition, which is bad for consumers. Canon should compete by making a better lens.

That's beside the fact that the above statement simplifies things, e.g. there's a competition on price/performance.

 

[Lawliet]

Canon also has to take care not to create precedence - include a fix for one lens and you'll be asked why all the other stuff doesn't get the same treatment. Suddenly the blame for flaky behavior gets shifted from the manufacturer to a 3rd party who doesn't even want to be involved or has the means to solve them as they can
t fix problems on the far side of the mount.
I wonder how long it will take until some lens declares itself capable of working with all 61Points without restrictions when in reality less optimism would be prudent. Same basic problem, but much subtler yet more troublesome effect...

 

[Random Orbits]

Whilst the issue of making a lens mount open or closed is an interesting argument, the fact of the matter is that you know full well that the EF mount is currently a closed system. On that basis, it is not fair to bang one's fists on the table and demand a solution to a problem caused by a competitor's unauthorised reverse engineering of Canon's product.

Reverse engineering when done properly is 100% legal and does not need to be "authorised".

And we as consumers are definitely within our rights to demand that Canon open up their lens communication protocol to make it easier for 3rd party integration.

It's your right to give your opinion. It is Canon's right to protect their IP, and it is their right to change their protocols without letting unlicensed users know. Ultimately, you have the choice of where you spend your money whether it be on Sony, Canon, Nikon, Sigma, Tamron, etc. Unless you have some legal ground to force Canon to make their IP open, then that's the way it is and you can only "vote" with where you spend your money. If enough people agree, then Canon might be persuaded if it starts affecting sales/profits, but I don't think that will happen any time soon.

 

[Ellen Schmidtee]

Whilst the issue of making a lens mount open or closed is an interesting argument, the fact of the matter is that you know full well that the EF mount is currently a closed system. On that basis, it is not fair to bang one's fists on the table and demand a solution to a problem caused by a competitor's unauthorised reverse engineering of Canon's product.

Reverse engineering when done properly is 100% legal and does not need to be "authorised".

I'm not a lawyer or U.S. citizen, so I might very well be missing something, but doesn't the DMCA put some limits on reverse engineering?


As Sigma completes Canon's line of lenses, IMHO, Canon would benefit from having their cameras avoid such SNAFUs. Examples would be fisheyes - until the 8-15mm, Canon made only a diagonal fisheye for FF, and Sigma made circular fisheyes for APS-C & FF + diagonal for APS-C.

 

[heptagon]

So can we conclude that Canon lenses should be protected like printer ink (more expensive than blood). This can be done with a little security chip built into the lens which identifies it as a genuine Canon. From a certain date on e.g. 1/1/2014 all other lenses will be disabled or set to full manual (no aperture control). It's about time for Canon to step up to these product pirates!

 

[Ellen Schmidtee]

There was a similar problem with Tamron lenses over a year ago - Canon apparently limited the functionality of focus points for some of it's lenses, and Tamron lenses which used the same IDs were hurt as well.

The more lens-specific-processing Canon adds to it's cameras, the more this kind of problems will occur, giving Canon bad rep as the brand that does not play well with others.

IMHO, the best solution for Canon would be to set aside a block of IDs for lenses by companies that reversed engineer the mount protocols, and not doing any lens specific processing for lenses that use an ID in that block.

 

[heptagon]

IMHO, the best solution for Canon would be to set aside a block of IDs for lenses by companies that reversed engineer the mount protocols, and not doing any lens specific processing for lenses that use an ID in that block.

This would be a very good idea for a start.

 

[Lawliet]

This would be a very good idea for a start.

They actually might have done that - but for an outside party the only way to differentiate between reserved for 3rd party and reserved for future lenses would be to ask.

But a basic problem remains: Canon doesn't do lens specific processing for fun, but because advanced techniques work only with certain lenses or at need specific data for others. That was a problem in the past when Sigma lenses only reported AF parameters for consumer bodies but lacked the second set for area-type sensors. Most of the time only the pro bodies were affected, and the 7D is using the center of most lenses, but if the 5d3 is an indicator for upcoming high end prosumer bodies...

 

[heptagon]

As far as i know, some (all) lenses aren't identified in the EXIF-Data of the image by lens-ID but by parameters like aperture and focal length ranges. Programs have to sort out which lens it actually is from that data. This might be one of the causes for the problem and i don't know how it is done for new lenses.

 

[Ellen Schmidtee]

This would be a very good idea for a start.

They actually might have done that - but for an outside party the only way to differentiate between reserved for 3rd party and reserved for future lenses would be to ask.

Or Canon could publish a list of lens IDs on it's web site that goes "ID A lens a, ID B lens b, ...., L through P reserved for reverse engineered lenses, ..."

 

[Ellen Schmidtee]

But a basic problem remains: Canon doesn't do lens specific processing for fun, but because advanced techniques work only with certain lenses or at need specific data for others. That was a problem in the past when Sigma lenses only reported AF parameters for consumer bodies but lacked the second set for area-type sensors. Most of the time only the pro bodies were affected, and the 7D is using the center of most lenses, but if the 5d3 is an indicator for upcoming high end prosumer bodies...

I think it would be appropriate to distinguish between two cases:

1. Lens specific processing is not applied, because 3rd party manufacturer did not pay for the benefit.

As long as the 3rd party manufacturer does not falsely advertise that the lens would benefit from lens specific processing, e.g. automatic illumination correction or area focus, I see no problem here. The consumer would make a conscious between features (price, IQ, lens specific processing), and do what he sees fit.

2. Lens specific processing is misapplied, because 3rd party manufacturer had it's lens ID itself as an other lens.

IMHO, both the OEM and 3rd party manufacturer should cooperate to avoid this case. When a consumer buys a lens, he does not know which lens specific features would be introduced in future cameras, and how much effort it would take to circumvent it's misapplication.

In other words, I expect manufacturers to apply defensive design to avoid this cases.

 

[Mt Spokane Photography]

So Canon doesn't provide a proper way for other manufacturers to make EF lenses without paying horrendous licensing fees. Then it purpously breaks compatibility with existing 3rd party lenses with new bodies. That doesn't sound good to me. All Canon would have to provide would be a proper way to ID the lenses. A simple manufacturer ID + item ID would be sufficient.

Third party manufacturers can assign their own lens ID, they do not need to use a Canon ID. They are merely trying to be clever, and it backfired.

However, the only way for a camera to know the optical and capable characteristics of a lens, is for the lens to be in the internal table that resides in the camera. So, Canon would have to maintain data for all the competitors lenses and issue camera firmware updates for them, and get complaints when problems occurred.
Thats like saying car makers should design cars to work with all transmissions from other makes. Then, whenever they want to make a change or improvement, they must get everyone else to change.
Its not so simple as one might believe.
But, the lens correction feature is not mandantory, if a user wants to use a third party lens, he can turn it off. Canon has actually provided for third pasrty manufacturers, you merely cannot use the lens correction feature in camera.

 

[Lawliet]

e.g. automatic illumination correction or area focus, I see no problem here. The consumer would make a conscious between features (price, IQ, lens specific processing), and do what he sees fit.

2. Lens specific processing is misapplied, because 3rd party manufacturer had it's lens ID itself as an other lens.

That works only if specific processing is optional. But those cases are only the tip of the iceberg, the real trouble starts when the camera needs the processing to work properly. Sure, you could set the AF to "dumb as a rock" as long as there is no proper ID, which likely would result in accusions of deliberatly crippeling 3rd party lenses. Or you shift it down as needed - looks fine on paper, but some of the Sigma lenses either front- or backfocusing depending on the orientation of the test target is uncomfortably unpredictable. And thats with just the center point at constant distance, more fun if you introduce real life variables.
Some cosmetic changes that have no impact on RAW files wouldn't justify to much worrying, I'm more scared of hidden troubles creeping up. ???

 

[Ellen Schmidtee]

e.g. automatic illumination correction or area focus, I see no problem here. The consumer would make a conscious between features (price, IQ, lens specific processing), and do what he sees fit.

2. Lens specific processing is misapplied, because 3rd party manufacturer had it's lens ID itself as an other lens.

That works only if specific processing is optional.

IMHO, illumination correction is optional, e.g. the user can turn it off.

Furthermore, what would happen if a customer would mount a new lens without upgrading to a firmware version that knows how much illumination correction to apply? What if the firmware version that knows that happens to have a bug?

I think it's reasonable for the firmware to fallback to no illumination correction for lenses it doesn't recognize.

But those cases are only the tip of the iceberg, the real trouble starts when the camera needs the processing to work properly. Sure, you could set the AF to "dumb as a rock" as long as there is no proper ID, which likely would result in accusions of deliberatly crippeling 3rd party lenses.

If that happens, Canon should reply those are unrealistic expectations, same as expecting it to buy such 3rd party lenses, invest man hours to measure it's light falloff, & add the appropriate illumination correction to the cameras's firmwares.

 

[Ellen Schmidtee]

So Canon doesn't provide a proper way for other manufacturers to make EF lenses without paying horrendous licensing fees. Then it purpously breaks compatibility with existing 3rd party lenses with new bodies. That doesn't sound good to me. All Canon would have to provide would be a proper way to ID the lenses. A simple manufacturer ID + item ID would be sufficient.

Third party manufacturers can assign their own lens ID, they do not need to use a Canon ID. They are merely trying to be clever, and it backfired.

As long as Canon doesn't set aside those numbers (before or after the fact), it's possible Canon would happen to reuse that ID for one of it's own lenses later on, causing the problem to arise at that time.

However, the only way for a camera to know the optical and capable characteristics of a lens, is for the lens to be in the internal table that resides in the camera.

That's not what I've suggested. My suggestion is to set aside lens IDs, then not perform any illumination correction. That would avoid the problem with practically zero effort and expense for Canon.

 

[danski0224]

Does Canon offer licensing options to 3rd party lens manufacturers?