Wednesday, April 15, 2009

No Difference between Canon 5D and Canon 5D Mark II at 1600 & 3200 ISO

I just got a Canon 5D Mark II in my hands from an official Canon source that I shall not mention here. Here's the question I had: is there any difference between the Canon 5D and the new Canon 5D Mark II at 1600 ISO and 3200 ISO? The 5D Mark II can also shoot at 6400, 12800 and 25600 ISO. But I wanted to know if any improvement had been made at the two highest ISOs they share: 1600 ISO and 3200 ISO.

I shot photos of one of my kids' rain boots laying on the basement floor in very dim natural light. I used a Canon 24-70 f/2.8L lens on both the 5D and the 5D Mark II. I shot all photos at 51mm and f/2.8. All photos were shot in RAW format and then prepared for this report with Lightroom 2 and Photoshop CS2. Now, I know the 5D Mark II comes with some new fancy algorithms for reducing image noise. But if you shoot almost everything RAW, like I do, those don't really matter much.

The one thing I did that some people might object to was crop the same amount of the scene for both cameras, even though a true 100% magnification section would take in less of the scene with the 5D Mark II compared to the 5D because it has 65% more pixels per image. (At the end of this test I'll include one of those comparisons in case you're interested.) My reasoning was that if I were looking at a specific object in a 5x7 print from each camera (for example, the bride's face when she's with the groom on the dance floor at the reception), I would be interested in the amount of noise I saw in her face in the 5x7 image, which would be the same size for each camera (more pixels doesn't change the relative size of objects in the same 5x7 photo!).

Note: I didn't use a tripod for these tests; so you can't tell much about the resolution of details from these photos; but some camera movement will not affect the amount of digital noise in the image. I did no sharpening when preparing these images for this report.

At 1600 ISO

Canon 5D Mark II at 1600 ISO (a higher quality version of the above image)

The details are a bit clearer in the Canon 5D Mark II shot; but it's not clear whether that is due to the greater number of pixels in the 5D Mark II image, or whether there was more camera shake when I took the 5D shot. Regardless, the point of this test is to compare the amount of noise in the two images. Look at both the shadow and highlight areas. Do you see any significant difference? I don't.

At 3200 ISO

Canon 5D Mark II at 3200 ISO (a higher quality version of the above image)

Somewhat ironically, the image from the 5D has the better detail in this comparison, probably due to camera shake. But again, this would not affect image noise. So, if you look at the noise in the highlight and shadow areas, which camera is best? Is there really any difference?! I think it's truly a dead heat.

Now, the 5D doesn't offer the higher ISOs of the 5D Mark II....6400, 12800, and 25600 ISO. So, let's look at a similar image taken at 6400 ISO with the 5D MarkII:

Canon 5D Mark II at 6400 ISO (a higher quality version of the above image)

Pretty noisey. Makes me wonder if I underexposed a shot at 3200 ISO with a 5D by one stop if I'd do any worse once the exposure level was raised in Lightroom?

Anyway, let me show you the 100% magnification crops for the 5D and 5D Mark II at 3200 ISO so you can see if it makes a difference--i.e., if it makes the difference between the 5D and 5D Mark II any clearer:

Canon 5D Mark II at 3200 ISO 100% magnification

Canon 5D at 3200 ISO 100% magnification
(higher quality version of above image)

The 100% magnification view doesn't make much of a difference to my eye. Let me know if you see something I'm not.


I think there's the assumption out there in the Canon user community that the 5D Mark II probably has less image noise than the 5D at 1600 and 3200 ISO because it can take relatively "acceptable" photos at 6400 ISO, and can also take photos at 12800 and 25600 ISO. The results of this test have convinced me that this difference doesn't actually exist. And unless I want to take a bunch of photos at 6400-25600 ISO, have an extra 8 megapixels per image, and shoot HD video, then a lightly used 5D selling for $1300 versus the 5D Mark II selling price of $2700, might just make a lot of sense!

Addendum: I made the statement that the noise reduction in the 5D Mark II doesn't affect RAW images. While I believe this is true, I can't find mention of this in the 5D Mark II manual. So, in the interest of full disclosure, I will include the fact that the 5D Mark II's high ISO noise reduction setting was at "standard" for these tests, which is second only to the highest setting: "strong" noise reduction.

Note: you may also be interested in my follow-up blog post describing my experience using the 5D Mark II to photograph a wedding.

Note 2: I finally got a hold of the 5D Mark II again and did a follow-up comparison of images at 1600 and 3200 ISO with those from my 5D. My results were a bit different this time...

Michael Grace-Martin is a professional wedding, portrait, event, stock, and fine art photographer based in Upstate New York. He is also the author of this blog. All images and text are (c) Michael Grace-Martin Photography. His main website is:


  1. Thanks so much for your review. I have had a 5D since its introduction and would only consider an upgrade for better high ISO performance. Decision made!

  2. Your tests are inconclusive because your sample pictures are mostly shaky and blurry. Please redo the test using a tripod and a cable release. I'm really interested to see the results


  3. I think that the presence of camera shake actually gets the focus off of looking at how details are resolved and, instead, looking solely at the noise in the image. However, I've heard from a few of you now who *really* would like the test re-done without camera shake. I'll have to wait until I can get a hold of the Mark II again. But when I do, I'll re-do the test using a tripod to make y'all feel better!

  4. AMOUNT of noise and CLARITY of noise are two different things. Maybe you need to research a little and yeah redo the test. Your experiment has two many uncontrolled variables -- most notably camera shake, which make me wonder if you even used same shutter and f/stop -- and therefore inconclusive results.

  5. So I finally re-did the test. You can see the results at