Sigma digital lenses with Canon full sensor

[cristina_fumi]

Does anybody know if those Sigma lenses (15-30 and macro 105 f 2,8)

will work with a full frame sensor such as the one of the Canon 1ds

mark II or the new 5 D?

[ben_rubinstein___mancheste]

yes, film also.

[cristina_fumi]

Thanks!

[mark u]

Both lenses work fine on a 1Ds Mk II (at least so long as the 105 isn't a very old one - see below), and I'd guess they will work equally well on a 5D, although no-one has had the chance to test them to prove the point as yet. Sigma lenses to be avoided for full frame use are those with DC (note Cee not Gee) in the name, since they only have a reduced image circle (also a feature of Canon EF-S lenses), and lenses which are more than 5 years old. The newest DG lenses are ones that have additional coatings on the rear elements designed to reduce reflections from shiny sensor optics, and are less prone to flare when used on a DSLR.

 

So far as the older lenses are concerned, many of them were built before Sigma improved their electronic design for compatibility with all EOS bodies. Some of them may work with the more recent bodies, and some of them can still be re-chipped to do so, but you need to research this on a case by case basis. These older lenses will work with older bodies, and even with e.g. a 1D (original).

 

It's worth pointing out that Sigma's new electronic design appears to be based on firmware that can be easily updated rather than hard coded logic chips that were used in earlier lenses. Also worth bearing in mind is that Canon have to retain compatibility with all their EF lenses - some dating back to the late 1980s - when they release a new body. That limits the changes they can make without having to offer an undesirable expensive lens rechipping programme of their own.

 

You may have seen (e.g. in the white paper for the 5D) that the 5D incorporates some new focussing algorithms. However, a closer reading suggests that these combine different algorithms already used in different bodies - some were designed to prioritise focus speed, and others focus accuracy. The new algorithms combine to offer the best of both worlds.

 

Information provided by Birger Engineering, a firm that provides EF mount adaptors so that EF lenses can be used on e.g. TV cameras, gives a good insight into how the EF lens protocol works. In essence, there are commands for the lens to inform the body about its characteristics, and there are commands to make the lens focus, adjust aperture and switch on IS. The lens generates responses, confirming that the command has been received and acted on or reporting any failure. So far as focus and aperture adjustment are concerned, there are commands that make small step adjustments, and others that make bigger moves. It's probable that early EOS bodies didn't make full use of these different commands and responses. There is also an ability to adjust the rate at which data passes between the camera and the lens (the baud rate).

 

http://www.birger.com/Merchant2/merchant.mvc?Screen=ef232_softintf

[erik widding]

I want to preface this response with a disclaimer that I work for Birger Engineering, the manufacturer of the Canon EOS lens adapter that Mark mentioned in his post...

 

The suggestion Mark makes regarding the use of our protocol definition to better understand what information is available to the camera body from the lens, and what sort of control information the camera is capable of providing to the lens, is for the most part a good one. But it should be understood that the EF232 lens adapter is not a passive unit, and it does in fact provide for a layer of abstraction between the user and the lens. There is not a one to one corelation between the Birger lens controller protocol and the actual Canon EF protocol. The Birger protocol was in fact developed with the intention of supporting multiple lens types, such as Sigma SA and Nikon AF-S.

 

I would suggest that a much better resource for information regarding the capabilities of the Canon EF lens platform would be the "EF Lens Technology" chapter in the "EF Lens Work II" book published by Canon.

 

 

Regards,

Erik Widding.

[jbq]

"Also worth bearing in mind is that Canon have to retain compatibility with all their EF lenses - some dating back to the late 1980s - when they release a new body."

 

I don't think that's a problem. In my little mind bodies have a lot of relative flexibility, and it shouldn't be hard in theory (just annoying) to special-case the handling of an old buggy lens in the firmware of new bodies. On the other hand it seems to me that lenses have fewer adaptative capabilities, and that developing a new lens in a way that's still compatible with old bodies (and probably with their quirkiness) might actually be a harder task.

[mark u]

Erik - your comments are noted, and of interest. However, the issue here is whether Canon might be able to engineer incompatibility between a new body and current third party lenses without simultaneously causing problems for their own lenses. My reading is that the "hidden" features of the protocol that caused problems for older Sigma and Tokina lenses are now all in the open, and third party lens incompatibility is probably a feature of the past.

[erik widding]

Mark wrote:<BR>

<B><I> the issue here is whether Canon might be able to engineer incompatibility between a new body and current third party lenses without simultaneously causing problems for their own lenses </I></B>

<P>

From our experience having built what is effectively "just another camera body", I can say that it is possible that incompatibility was engineered into the body. But it is much more likely that this happened by accident, or by simply not testing with third party lenses.

<P>

For various reasons we do not test or support the use of our EF mount product with anything but Canon branded lenses. Six years ago we did the initial protocol work and came out with a product that was compatible with about 20 lenses. It came to our attention, as popularity for our product took off, this spring that we were incompatible with many of the newer lenses from Canon. We did make both electrical and software changes to remedy this issue. We are now compatible with 35 lenses - all that we have ever tested. We support software upgrade of our product in the field because we assume an incompatibility will eventually surface.

<P>

When we released the newest version of our product we were told by customers that some Tamron and Sigma lenses do work and some do not. We do not keep record of which. Some of the lenses that do not work were reported to have functioned with the older model, which had in retrospect obvious protocol "mistakes".

<P>

Even cameras and lenses all manufactured by Canon speak slightly different "dialects" of the protocol. We have spent a great deal of time and expense drawing this conclusion. This is something that has evolved over 25 years of development where many different people on many different teams have done work. Add to that a bunch of third parties (such as my company) that have learned to speak by immersion, without formal specification, lessons, dictionaries or the like, and confusion will ensue.

<P>

Our test set of 35 lenses started out with those we had easy access to and those we thought customers would want to use. Over the years it has been augmented with lenses that at one point or another were reported to us as incompatible. Canon has manuactured approximately 125 unique lenses (including all revisions of similar lenses). I would suspect that their test set includes every single one of these - except those they know speak exactly the same way as one already in the set.

<P>

Regards,<BR>

Erik.<BR>

<BR>

---<BR>

Erik Widding<BR>

President<BR>

Birger Engineering, Inc.<BR>

<BR>

(mail) 100 Boylston St #1070; Boston, MA 02116<BR>

(voice) 617.695.9233 x207<BR>

(fax) 617.695.9234<BR>

(web) http://www.birger.com<BR>

[cristina_fumi]

I really appreciated! Thanks to you all!

Cristina