Just after getting my new Canon 30D I feel the need to see whether it has noticeably lower noise at all ISO levels than the 300D I've already got. So I set up a basic scene in my bedroom and did some tests. For convenience, and to make it realistic, I did not use a flash which means that some of the exposures are somewhat long. I used mirror lock up and the self-timer to make things easier. All the cropping and so on was done in TIF.
I've got a separate page of the same shots done for every ISO on the 30D and if someone in Sydney will bring a 20D around for half an hour I'll happily add results from that.
| Lens | Canon 50mm f1.4 | ||
|---|---|---|---|
| Aperture | f/5.6 | ||
| Exposure | ISO | 30D | 300D |
| 100 | 1.6s | 1.3s | |
| 200 | 0.8s | 0.6s | |
| 400 | 0.3s | 1/4s | |
| 800 | 1/6s | 1/8s | |
| 1600 | 1/13s | 1/15s | |
| 3200 | 1/25s | 1/30s | |
Here's a summary of what it looks like. Apologies for the difference in white balance, but I'm running the old hacked firmware on the 300D so can't use custom white balance. Images were all shot in large/fine JPG with least possible sharpening and any other parameters set to zero. The 30D used a custom WB to try to get closer to what the 300D does, but without a huge amount of joy. Note that 3200ISO on the 300D is 1600ISO underexposed one stop then adjusted on the PC, so it's not exactly right. But it is very noisy.
A 500kB summary image at 200% showing more detail than the thumbnails.
A 3MB, 1200x2400 composite of 100 pixel high slices of each image, at 200%
Material on this site is
copyright © Moz, 2003-2006 (mozphoto@moz.net.nz).
To enquire about using images please email me. Generally I'll allow free use for non-profit users.