High ISO comparison 1D Mk II & Rebel XT

[Tom W]

As it's a rainy day, I decided to do some noise comparisons between the 1D Mk II and the new Rebel XT. I did my best to equalize the situation, using the same lens, tripod, same scene, and same framing (I had to zoom the XT a bit wider due to the smaller sensor). I shot raw with both cameras, and used DPP to process, taking care to set all parameters to the center position (i.e., saturation & contrast). I shot manual to allow the exact same shutter speed and aperture on both cameras at equal ISO settings.

I don't have the crops ready to post yet, but here's a description of my results:

At ISO 1600, the XT and 1D Mk II have very similar noise. Almost exactly equal in both amount and color/graininess. This bodes well for the XT, with its smaller sensor. There was a bit less detail on the XT, but not much. Probably a bit of on-camera noise reduction in the XT.

EDIT: taking a closer look, especially in the less dark areas on the image, the 1D2 seems to have a little edge at 1600 as well. Can't say exactly how much, but it's apparent when you view them side-by-side.

At ISO 800, the 1D Mk II enjoys about 1/3 stops worth of improvement over the XT. My ISO 1000 shots with the 1D2 were about equal in noise levels with that of the XT at ISO 800. Detail on the 1D2 was again slightly better than that of the XT.

At ISO 400, the 1D Mk II has about 1/2 stops worth of improvement over the XT in terms of noise. I compared the 1D2's ISO 500 and ISO-640 results with the XT at 400 and found that while the ISO 500 was a tinch cleaner, the 640 was a bit grainier than the XT. At ISO 640, the 1D2 still showed a tiny bit more detail as compared to the XT at 400.

I did not test below ISO 400 since both cameras are great performers at low ISO ranges. My belief is that the XT uses some on-camera noise reduction, but it's negative effect on detail is not really an issue except for the most demanding work. Very good performance IMHO.

[Tom W]

Here's a link to the test images:

http://www.pbase.com/photosbytom/1d2_xt_noise_test

[griff2]

Looking at the ISO 400 images, the 1DMkII images are IMO clearly superior. The XT's image appears slightly soft, the 1DMkII is sharper - you pay's your money..

[johneric8]

Looking at the ISO 400 images, the 1DMkII images are IMO clearly superior. The XT's image appears slightly soft, the 1DMkII is sharper - you pay's your money..

I agree that the MKII is a hair better but, not worth the money paid that is for darn sure. Look, I do weddings, calenders, portrait work and used both the 20D and the MKII. My conclusion is - If you have the money to blow go ahead and get it! But, as far as image quality, I cant see it as being worth the money to pay more for a MKII based on image quality. I'm sure some may disagree. I shot an event the other day with my 20D while my partner shot with him mkII. We compared images, and I came out on top by far that day. I wouldnt recommend paying extra for an mkII to anyone unless they just had a need for the latest greatest toy or where Proffesionals in the field and needed the extra pixels for very large prints. Everyone seems to forget that it wasnt long ago when pro's would shoot with just okay 35mm film cameras shooting national geographic with outstanding results. A digital rebel in the right hands can run circles around most cats with MKII's on this forum including myself.

[CyberDyneSystems]

mmmmmmmmmmm...... Laphroag ....... http://members.bellatlantic.net/~vze2hdzu/pics/beer.gif

[Tom W]

Noise is but one aspect of a camera. I was pleasantly surprised that my XT performed so well in that department, but it still takes a second seat in other aspects of operation. That's not to say it's bad - it's a very good performer, especially when it's price is considered. It outperforms my old 10D in many respects, and that was also a great camera (and I'd love to have been able to include that in the test as well).

But there are features and performance characteristics on the 1D2 that aren't matched by the non-1 series Canons including the incredible low-light & high-speed focusing capabilities, spot metering, big viewfinder, frame rate, and such. If you're shooting sports or quickly-moving objects, the 1-series is the better choice. The 1-series cameras do expand the range of situations where you can shoot with great confidence. If you don't shoot in those situations, then you might not need it.

[Tom W]

mmmmmmmmmmm...... Laphroag ....... http://members.bellatlantic.net/~vze2hdzu/pics/beer.gif

That one's empty. Not so for the Chivas, though. :)

[Tom W]

I've done a little work in the overexposure world with these two cameras. So far, at +2 stops overexposed, I'm able to pull them back down in RAW to normal. There's a very small detail advantage with the 1-series, but both deal with that degree of overexposure about the same. Basically, some highlight detail is lost in both images, but there isn't a great deal of difference. But, this test setup was the same as the original test, and the base exposure was a bit towards the center (about 2/3 stops below maximum on the histogram). I don't think the highlights were overexposed enough to really destroy detail. I'm going to test under more strenuous conditions to see what happens before I post any results.

[Pat H]

Especially considering the money I paid for my XT and the much higher price of the Mark II, the images are really not dramatically different. If you scrutinize them, sure, but really not noticeable at a casual glance. Definitely not worth 7x the price image quality wise. Seems that most of the $8,000 goes toward less restrictions, some more MP's, and the name on the camera. But then again, I'm sure those who actually own the camera would disagree.

[Cadwell]

That one's empty. Not so for the Chivas, though. :)

Chivas is an abomination. :evil: Some of the finest highland malts are blended (wasted) to produce it! Do you know, they even put 12 year old Strathisla in there? One of the best single malts you can (could) buy :evil:

The Laphroaig is quite nice if you fancy a drop of the peaty stuff though ;)

[Tom W]

Chivas is an abomination. :evil: Some of the finest highland malts are blended (wasted) to produce it! Do you know, they even put 12 year old Strathisla in there? One of the best single malts you can (could) buy :evil:

The Laphroaig is quite nice if you fancy a drop of the peaty stuff though ;)

The Chivas is a very mild blend - smooth, but water is smooth as well. Not my favorite, and I hate to see them waste the good stuff making it. Although I prefer the single-malts in general, among the blended scotches, I favor Dewar's the best.

I do enjoy the peaty flavor of Laphroaig. Which explains why there's none left.

[Tom W]

Especially considering the money I paid for my XT and the much higher price of the Mark II, the images are really not dramatically different. If you scrutinize them, sure, but really not noticeable at a casual glance. Definitely not worth 7x the price image quality wise. Seems that most of the $8,000 goes toward less restrictions, some more MP's, and the name on the camera. But then again, I'm sure those who actually own the camera would disagree.

Just to note, the camera tested was the 1D Mk II, not the 1Ds Mk II. Mpix were roughly the same, and the price differential is about $3100. I'd gladly perform a long-term test on the 1Ds Mk II if someone were to "loan" me one. :)

To me, it's amazing what 1 year of technology can do, bringing the smaller sensor so close to the noise levels of the 1.3X sensor (even at the expense of a tiny bit of detail). The Canon family is looking good across the DSLR range.

Digital is maturing, and is approaching film in the respect that the cheapest Rebel uses the same film as the 1V and is capable of producing identical images in most situations. It's in those less-simple situations that the 1V shines above the film Rebel. Likewise, the XT is capable of producing images very close to the 1D Mk II in most situations.

I'm still hoping to do some extensive highlight-recovery testing on the two cameras, as well as shadow recovery to see how much information is locked in there. But that will be time-consuming and I've got to take care of a few other things around here (like working, eating, and actually using the equipment).

[Pat H]

Just to note, the camera tested was the 1D Mk II, not the 1Ds Mk II.

My mistake :o

[Tom W]

I did some more overexposure shots today to see how the two cameras could pull data out of the presumably blown highlights. Of course, if the highlights are truly blown, there is no recoverable data. But, I shot RAW images with both cameras at +2 stops and +3 stops over what the camera considered normal exposure. The result was that both cameras were able to pull down the +2 stop overexposure with a moderate loss of image quality. At +3, the results were poor from both, but that's to be expected. Once a pixel reaches full white, it's not going to get any whiter or darker.

The difference between the two cameras was small - the 1D2 had what appears to have a bit more range in pulling good data from the top end of the exposure range, but I think that this might be more dependent on the shape of the respective tone curves. My guess is that both cameras are limited at the high end by the 12-bit per channel RAW image, so that there can't be a huge difference. It was nice to know that with a RAW image, there is some latitude at the top end.

At the dark end, I underexposed both cameras at -2 stops again in RAW. The 1D2 was able to pull more detail out of the underexposed area, as expected due to its slightly lower noise floor. The difference again wasn't huge, but noticeable.

For fun (OK, not that fun), I also took some overexposures in JPG on the 1D2, at +2 and +3 stops over. Neither allowed any recovery close to what the RAW image could do. I simply could not get the detail to return, nor could I get the colors very close to the original. If you're exposing well, JPG is fine. If you're going to flirt with the extremes, you'll have less latitude to recover.

I must make one point on the overexposure - my subject was a particularly bland but detailed page from a DeLorme topographical atlas. When metered, it gave a pretty good lump in the middle of the range with very little information in the top or bottom 1/5 of the chart. In other words, correct exposure had it's highest values about 1 stop below full exposure. That means that when I overexposed this scenario by 2 stops, it was probably more equivalent to overexposing a typical scene by 1 stop. I don't think a 2-stop overexposure of an outdoor scene could be salvaged, at least not in the brighter parts of the image.

[Michaelmjc]

They are very comparable, I didn't think my little XT was capable of keeping up with the big guy.

[defordphoto]

They are very comparable, I didn't think my little XT was capable of keeping up with the big guy.

This is not surprising in the least. They are both using an extremely similar DIGIC II processor.

[Tom W]

They are very comparable, I didn't think my little XT was capable of keeping up with the big guy.

I didn't expect the XT to play in the same ballpark as the big boys either, but it is a surprisingly good performer (and I'm sure that the 20D performs similarly, but with more flexibility). It's missing the things that separate the 1-series from the rest, but it has the capability of providing images of the same caliber in most situations.

I keep mentioning "most situations", because I've been in enough situations where the speed, focusing, and handling of the 1D2 were very useful in obtaining the images I wanted. But that doesn't represent the majority of everyday shooting. For the most part, the XT and 20D really will fulfill one's needs.

[defordphoto]

Absolutely Tom. If I wasn't shooting in the capacity I am shooting, I'd probably just have the 10D and two 20Ds rather than the MKII. As my business expands, I may find the need for an ultra-high megapixel camera, but that will be on down the future sometime. I've printed some amazing enlargements with 8.5mp.

Canon has a perfect lineup of cameras right now. Something for everyone, and it's good to see the Rebel camera line stepped up in quality over the original.

[willg]

Absolutely Tom. If I wasn't shooting in the capacity I am shooting, I'd probably just have the 10D and two 20Ds rather than the MKII. As my business expands, I may find the need for an ultra-high megapixel camera, but that will be on down the future sometime. I've printed some amazing enlargements with 8.5mp.

Canon has a perfect lineup of cameras right now. Something for everyone, and it's good to see the Rebel camera line stepped up in quality over the original.

easy there on the quality of the original rebel...it can produce some amazing shots..hell the d30 and d60 can still produce amazing shots too

[EOSAddict]

The Chivas is a very mild blend - smooth, but water is smooth as well. Not my favorite, and I hate to see them waste the good stuff making it. Although I prefer the single-malts in general, among the blended scotches, I favor Dewar's the best.

I do enjoy the peaty flavor of Laphroaig. Which explains why there's none left.


Just picked up a bottle of quarter cask high strength Laphroaig - nice.

Danger of going off topic on this thread but I like the idea of a Canon AND whisky lovers forum!! :lol: :lol: Personal favourite Balvenie Single Barrel 15yo. And don't lower the tone with the word blend!

[rklepper]

I absolutely love my XT.

[ed2day]

Thanks for the comparison, Tom. Interesting. I would have expected to see more noise difference between the two--they look roughly the same to me. The big guy looks noticeably sharper--more precise focusing perhaps? A lot of what you are paying for in the 1-series is gobs of high speed buffer memory--that's the only way to get sustained high frame rates.