Archive for May, 2010

Canon 7D native ISO tests

Wednesday, May 26th, 2010

Recently I picked up a Canon 7D which I’m really enjoying so far. While researching the camera I found various references to both 100 and 160 being the native ISO and as I do quite a bit of still photography from a tripod was keen to determine the optimum ISO setting and also the best ISO multiples to use for available light shooting. Between various forum posts there seemed to be about a 50/50 mix of people suggesting one or the other was the native ISO so I decided to do my own test. I printed an ISO 12233 test chart replica from Cornell University on an A4 piece of paper and set it up inside a lightbox in a room with no direct sunlight. I shot using the 7D with a Canon 70-200mm f/2.8L IS USM II lens at f/2.8 at around 100mm focal length, tripod mounted with a remote cable release, manually focused once at the start of the sequence with the camera set to aperture priority mode. While I’m sure there are plenty of things that could be done better I thought this was typical of the kind of lighting situation where I’d be tempted to knock up the ISO speed. I imported the RAW files in Adobe Lightroom 3 beta 2 and didn’t perform any operations other than changing the colour balance and exposure equally on all shots to lighten them up for printing.

The results? ISO 100 and 160 were very close, but I felt 160 was just a little more defined so I might settle on that for tripod shooting. I don’t have an expert eye but I couldn’t spot any nasty steps in the ISO range that other’s have reported, in my mind it was a pretty even transition with the each ISO speed. ISO 400 remained excellent so I wouldn’t hesitate to go that high for quality shots. What I’d set out to find is if automatic ISO was a useful mode and I feel that it is. From what I gather Canon don’t release specific details on the native ISO of their cameras but the results made me wonder if on a modern DSLR it’s a myth, I’d assume the ISO setting equates to a sampling period for the sensor and can’t think of any reason it couldn’t be infinitely variable over a range. Of course at the lower end of the range there could be an optimum level such as 160 below which the sensor becomes saturated and needs to be attenuated which may explain why ISO 160 appeared marginally better than ISO 100.