Nikon vs Canon -The MegaPixel Issue
So, I was screaming mad when I learned that the D3's chip size was actually slightly smaller than the D2x (i.e., the D2x is 12.4MP, the D3 is 12.1; the D3 is, though, essentially full-frame). But, does it really matter? Well, that depends.
If you were someone who dropped $20-$30k on a medium format back, just to get a native resolution of 22MP, this camera's got to frustrate you, especially when you realize that the color depth of the Canon and Nikon is going to be so comparable to your MF back, that you'll feel comfortable shooting chocolate with these cameras.
So, what of the difference in chip size?
(Continued after the Jump)
The image specs and lens choices are the same as outlined in the noise section of this piece. In fact, since we're all worried about the size of the Nikon as smaller than the Canon, I'm putting up the Canon iso3200 here first to compare to the Nikon files I'll be showing.
| Here again, so it's on the same page, is the Canon, at iso3200. |  |
| Here's the same Nikon iso3200 file, but it has been rezzed up to be the same size as the Canon file. No sharpening has been applied. It appears there is less noise, and it's smoother overall. I know it's not as sharp, but when you're so deep into this photo, either the viewing distance of the viewer or the reproduction abilities of the output device will all but take care of this concern. |  |
| Here's the same Nikon iso3200 file but it has been uprezzed and sharpened to my tastes. It appears equal to, or better than, the Canon image. In other words, by up-rezzing the file to the size of the Canon file, I can actually achieve a more pleasing final result than the native file results from the Canon. |  |
| Just for comparison purposes, here's the same Nikon iso3200 file as above, not uprezed, not sharpened. (in other words, it's the same file, just uprezzed, and then secondarily sharpened. |  |
These results, for me, essentially kill the megapixel issue for me. Size, with the proper chip and internal camera processing software, is no longer a comparison point for me. I'm sure we'd have problems if there was noise at the ISO's I need, because enlargement would yield more noise, but since that's not a concern, it's amazing what a 12.1MP camera can look like when compared to a 22MP camera. My yardstick had always been based upon a conversation I once had with an editor at National Geographic. We were discussing, very early on when it was not digital cameras, but the scanners being used to scan film - that a file that was 60MB at 8-bit was scanning at pixel-to-grain 1:1 using Kodachrome 25 as the benchmark. So, anything over 60MB was just increasing the number of pixels that hold one grain of film, and thus, is overkill/redundant. I see that, properly done as Nikon has, it's not an issue at 12MP.
- Nikon vs Canon - Introduction
- Canon - A first look
- Nikon - A first look
- Nikon vs Canon -The Noise Issue
- Nikon vs Canon -The MegaPixel Issue
- Nikon vs Canon - Shooting Tethered
- Nikon vs Canon - The LCD Screen
- Nikon vs Canon - External Ports
- Nikon vs Canon - Buttons and Access
- Nikon vs Canon - Card slots
- Nikon vs Canon - The Future
- Nikon vs Canon - Conclusions
Please post your comments by clicking the link below. If you've got questions, please pose them in our Photo Business Forum Flickr Group Discussion Threads.
16 comments:
I'm posting here because it's not really relevant to your conclusion, but I want to mention this section.
I realize that what you decide is more "pleasing" is entirely up to your own aesthetics, but the smoothness that you see in the uprezzed Nikon file is actually a loss of color detail. (Ironically, giving the plastic doll a more plastic appearance.)
What happens to the Canon File when up-res'd to the same degree?
Really, man, you need to be more convenient. I can't read further after that section.
You are rezzing the nikon and it looks unreal or fakey. the canon looks really good and it is not rezzed at all. canon is still the one to beat and you are doing your best to beat it but the results are not showing what you want them too. the E3 is a hell of a camera but it is not a canon or medium format. jerry e
sadly this comparison is without meaning. 1. Compare resolutions at ISO100, 2. The details in the picture are not sufficient enought to show a difference. 3. If you look on other pictures compared 1DS III and D3, like imaging-resource.com has good pictures for a fair comparison from both cameras, then you see that the D3 has lower noise as the Canon, but as usual due to a higher Noise Reduction. So the D3 looses lot of Details.
4. The D3 is noise wise the best camera Nikon ever produced but they can't do magic...
Yep, I think the IR has done good work with their samples. Compare these two photos:
http://www.imaging-resource.com/PRODS/D3/FULLRES/D3hSLI00100.HTM
http://www.imaging-resource.com/PRODS/E1DSMK3/FULLRES/E1DSMK3hSLI0100.HTM
And look at the green cloth. You can clearly see the texture of the cloth from the 1Ds Mk3 picture, but with D3 the texture looks rather ugly. And I bet no sharpening can bring it out.
After looking at the samples at imaging-resource.com I'm quite surprised by the difference in Dynamic Range. The Canon seems to handle highlights much better than the Nikon.
The flaws in this test are many, but so is drawing conclusions from imaging resource photos. Straight out of the camera jpegs are apples and oranges. Nikon "normal" is less sharpened than Canon "normal" which has ugly edge effects. Both introduce tone curves, noise reduction, sharpening, etc. You simply can't draw any conclusions from them, except as to in-camera jpegs. To maximize quality, you must start from raw. Remember, this is "light hearted" and there is a boxing ring!
I have to agree with the others here. I don't think the Nikon image is "better" either. I personally prefer the Canon image, in comparison he Nikon's look washed out & the uprezzing has lost detail. But, it's hair-splitting at full size pixel peeping; and the images could be brought so close that when printed it makes little difference what sensor captured the image.
That said, Nikon has obviously taken huge strides forward & can now compete with Canon at these high ISO settings. That's good for us.
The Nikon and the Canon have different uses. I would use the Canon for large ad print work, but I don't want to shoot hundreds of wedding photos on the Canon. I don't need the large size print and I don't need to waste all that digital storage space. The right tool for the right job.
I love it. I can tell what kind of camera's you guys shoot just by your posts! If your pissed off that the once hailed "best in class image" is now second best to Nikon go buy a D3. I mean that is why you are reading this right, because you have already heard that the D3 is the best camera on the market now and you are doing your research to get one.
Although personal bias may play a role in the evaluation of the presented samples (and I will admit I am biased also), I think for the examples given, the Canon still comes out better. The noise in the Canon file is more fine grained than the noise in the uprezzed and sharpened Nikon file. I think its a great result from the Nikon, but at 21-22MP, I think the Canon is still better at ISO 3200. Based on the examples in the article.
Curious that the opposite comparison wasn't done, downsizing the Canon file to match the resolution of the Nikon file. Wouldn't this result in lower noise for the Canon?
it's funny, but I would have preferred you up-rezzing barbie's hair with the nikon. Now why haven't you shown us that I wonder?
Since you're doing post processing on the Nikon image, it would be interesting to also do something with the Canon image. Of course in that case instead of scaling it you'd want to run it through a noise reduction filter. That way you'd be looking at a more realistic comparison of the best you can do with an image coming out of a post processing workflow.
On the issue of file sizes too big for jobs that don't require it, there's always the option of shooting them at sRAW and Medium to Small sized jpegs.
And for all those people who say they love their 1.6x crop because it gives longer reach on the tele-end, I say, crop form the FF and you'll not only get better "reach", you'll get better bokeh as well.
compaq evo n110 battery
compaq evo n110s battery
compaq pp2101x battery
Post a Comment