However, from what I've read and saw in person, the extra special coatings on the Sigma, actually seem to make it a bit slower than 1.4. It almost seemed that it was about 3/4 of a full stop slower than the Canon.
I'm still wavering, but at this point, I'm leaning towards the 50L 1.2. I was able to shoot some MIGHTY dark bars with this and no additional lights with the 1.2 wide open.
So for when you need very low light...at this point, I'm leaning towards getting the canon 50L 1.2.
There are some topics relating experiences with F1.4 lens in Canon digital cameras, I did not find now, but the conclusion is surprising:
A member of CR did a test with an F1.4 lens (at F1.4) mounted on a Canon DSLR camera with all manual settings, and photographed a target with controlled lighting. Then he repeated the shot with everything the same, except that he put duct tape on the electrical contacts of the lens, so the camera could not identify which lens was being used.
He found that the picture looked darker when the camera could not identify which lens was being used. Neuro then gave a simple explanation and terrifying:
Digital cameras (unlike film) do not capture well the light that hits the sensor at a very tilted angle. How F1.4 lenses (and more luminous) many light rays arrive at quite tilted sensor, and do not penetrate well into the photodiodes. To circumvent this problem, Canon pushes the ISO (in secret) to simulate the use of light rays that exists in the film. Ie. When you select ISO100, the camera secretly push the ISO up to 153 (for example) and compensates for the lower utilization of light rays.
If you select ISO 1600, the camera pushes secretly up to ISO 2129 (for example), and compensates for the lower utilization of light rays. In this case, the noise will be larger than an ISO 1600 "true".
If the lens is not manufactured by Canon itself, ISO not be pushed secret, and the lens will appear darker than a Canon lens that had the ISO pushed.