Dear Brogaard and Szabo,

I think what you have experienced is something related to the lens and not the imaging element itself. The lenses that you used do not exist (50mm f/1.2 and 80mm f/1.2) but in case you used the 50mm f/1.0 and the 85mm f/1.2, yes they suffer from a small amount of distortion.
The distortion has been seen by others too, as can be assesed by checking http://www.photozone.de/bindex2.html .
One of the best lenses to get straight lines would probably be the 90mm L TS-E, or one of the 600mm lenses.
Regards,
Andre  :p

Fascinating to see how the 1Ds captures good old fashioned barrel distortion so exquisitely.

The issue you have is totally with the lens and not the camera. The fact that the distortion is visible in the viewfinder proves this. The 1Ds image chip will reveal the flaws in your lens more clearly than a lower resolution digital or film camera. The camera cannot record a better image than the lens can create.

The EOS 1Ds i a superb camera, no doubt. Except when it comes to keeping lines straight. Please have a look at the two jpg's posted at this url:

Oval EOS 1Ds

As you scroll down the images of an ordinary paper you'll notice the lines move in an oval way just like a bellydancer. You will also notice the oval'isme is all over the photo. We haven't been testing any wideangle linse but suppose the problem is even worse here. You can actually see the line moving if you just look trough the cameras eyepice and keep a straight line in either side. We have never seen this failure on any film-based Canon camera including 1V or any earlier digital Canon SLR (and we have had them all). The problem has been noticed in all 3 EOS 1Ds we have been testing. We guess the failure arrives due to some different distance btw lense-back and CMOS. We have been trying to get an answer from Danish Canon rep. for over week - without any answers or returned calls at all. Our claim is that this failure is very seriuos when it comes to architecture and similar - you just can't use this camera at these situations. But also a close portrait will be too oval to be true.
You can very easy perform a similar test yourself - please do and let us and Canon know the answer.

Brogaard and Szabo

Oval EOS 1Ds

Hi,
As Andre already said this is a question of the lenses and in this case it is called BARREL DISTORTION and second I measured one of your photos 80mm? f1.2?and we are speaking of a distortion of 0,44%.
18pixels on the full distance on the upper border.
and the 50mm f1.2 has a deviation of 30 pixels at full length which are about 0.74% distortion.
Take a look at the page Photodo and there you can see this deviations.
I think you overreacted on this issue or I do not know what else you had in mind.  

Another clue was the statement that you could see it through the viewfinder as well.  Whats the common component in those two tests?  (Its not the imaging chip.)