« Random Excellence: Eddy Pula the Famous | Main | Memories of Dick Zakia »

Thursday, 19 April 2012


Feed You can follow this conversation by subscribing to the comment feed for this post.

Regardless of whether one gives any credence to the overall DxO score, their underlying testing methodologies do appear sound.

As a long time Canon user who owns a lot of L glass and flashes, I am seriously disappointed in what Canon has been able to come up with after nearly 3 years of "work" on 5DMkII. Either complacency got the better of Canon engineers or perhaps their Nikon counterparts are just more hungry to regain the top spot.

Whatever the case may be, I have decided against upgrading to 5DMkIII. And since I am no fan of Nikon color palette or ergonomics, I think I am eagerly awaiting any news from Sony on their version of full frame upgrade. I seriously hope they knock it out of the park this time. More power to the underdogs.

I tend to find myself fascinated by these numbers even though I think that for the most part these cameras are also so good that the differences are mostly irrelevant.

I promise not to pay attention to DxOMark numbers nor pixel peep. I promise ....

I feel so INFERIOR! How do I get off shooting with the 106th & 107th ranked cameras!

"How do I get off shooting with the 106th & 107th ranked cameras!"

Woo, fancy. Mine is only #126.


Yikes, looks like Canon has really lost their touch. This just isn't competitive.

How about them apples! 10/49!? Way to go Canon!

So when, I have to ask as an architectural shooter, are Canon going to get their act together with respect to DR and deliver a body that does justice to the 17 & 24mm TS-E lenses?

Not only that, but it is behind an APS-C camera nearly four times cheaper. Oh, the shame :-) ... also the difference between it and the arch rival D800 above is equal to the difference between it and lowly D3100 below. Whoever would by this piece of light leak???

or alternatively there is more to this sport than the sensor rating alone.

I don't pay much attention to DxO's ratings. I grew up on transparency film, any camera from the last several years has way more dynamic range. I'd like to hear from Ctein on DxO's methodology.

The 5DIII is the best camera I've owned. The improved AF system alone is worth the upgrade. I've already shot one event with it, so it's paying its way. Unfortunately, photos from most of the events I cover are not for public release. However, it does nicely with pretty pictures too: http://smu.gs/HWDYIR

As for the light leak, I'll worry about that when I take photos at high ISO with the lens cap and the top LCD illumination on.

An outstanding stills performance from a specialized video camera. This glass is half full -- not half empty! (:-)

Dear Mike,

Relative to some others, the exposure range isn't so great, but a smidge under 12 stops is still a very good exposure range.

It isn't always about grading on a curve.

pax / Ctein

The 5D Mark II had three serious flaws:

1) banding at low ISO
2) poor autofocus
3) soft video

From what I've gathered based on the reviews that are trickling out, #1 has not been addressed at all and #3 is still true, but for a different reason. Only the autofocus--which was terrible, but I could live with--has been fixed.

Don't get me wrong, the 5D2 was and is an incredible camera, but it's galling that image quality (both stills and video) in the 5D3 is still compromised, and at a higher price point than the 5D2.

It's very disappointing personally, since I'm planning to get my 5D2 converted for UV photography, and will need a replacement for "normal" work.

think of it this way. the 5dmkiii is a d700 with more megapixels, a 100% viewfinder, and better video. it's "what the d700 should have been!"

The reactions to these news are now at the level of say, the Porsche driver's pain when his car scores a pathetic 178 mph top speed vs. some Ferrari's measured 195. All in countries with speed limits in the 65-75 mph range.

Finally digital is approaching stability. There is no "need" to replace a camera every two years anymore. Good.

Whoah! I have enough trouble taking a decent photograph, never mind taking one that also makes the most of 14 stops of dynamic range

As someone who bought a camera because he moved over the tipping point based on a DxOMark, I think there's some serious validity behind it.

I've got a GF1 and a S95, both great cameras, both make great prints -- if you know how to use them. In situations where you need that extra DR or low light performance, you're either managing the situation (e.g. wait for light to change, or change the light yourself) or putting up with the result.

In the end, with my K-5 -- that extra DR and high ISO performance does make a difference. Probably mentally more than anything. Likewise when I used a D700, the feel of not having to worry about the camera's sensor performance to get the best result is ... liberating.

I'd compare it to say, taking a photo on film, which you know would be on limit (e.g. pushed, low light) and knowing you've got something, but something not perfect.


so canon released a new camera with a hole in the roof and a sensor, which equals the D3-sensor of 2007 (+ double resolution).
thats embarassing for canon, especially so since they usually make sure to have slightly "better" specs at least on paper...
watching the wildlife in the forums on these occasions is quite a ride!

Dang. I guess I'll have to send mine back. Never mind that it's the best camera I've ever owned.

Last time i checked amazon is that most camera on the top are Canon ... it is replacement business. Unlike Nikon, Canon has a lot of other fishes to fry. Some MBA thinking here may say "cash cow". Unlike Kodak, Canon really has other businesses that are also core. We knew that even using Sony A900 is good enough even now, innovation might be not that important. There is no Apple in this field. May be Canon is ok, even without using $ to do better R&D in this market and not even enter the EVIL segment.

I always think that Canon is a marketing company.

Dear TBanner,

I've been a steady defender of DxOMark's methodology and results. Their experimental procedures are good, their theoretical underpinnings are sound, and in the small handful of cases where I've been able to run comparable tests I get results close to theirs that also match well to my in-the-field photos.

In the very few cases where their results are substantially at odds with other test sites that I've been able to investigate, they come up right, and the other sites wrong.

Use the Google box on the side to search for "dxomark ctein" and you'll likely turn up many of those writings.

Please remember that DxOMark is doing the equivalent of "film testing." 'Cept today that means the sensor+camera system. They aren't telling you about anything else. It is not a total camera review, doesn't claim to be.

pax / Ctein

Banding. I don't see any banding with my 5DIII, and the banding issue with the Mark II was way overblown. I used the Mark II for almost three years without it ever being an issue for me or for clients.

What I find most disappointing is that the 5D2 lags the D800 by 2 f-stops of DR (that's a decent GND there) despite having pixels 1.5 times larger, which is a significant size.

They appear to have improved the body over the 5D2, but the sensor appears only marginally better. During that time Nikon, who already used the better sensors in each respective class with the D3S/X, have progressed much in the sensor department.

I use the 5D Mark II daily, and banding is _always_ an issue at ISO 100 in low light. Given that I shoot at ISO 100 in low light quite frequently, I see banding all the time.

Interestingly, the banding isn't fixed from frame to frame--if I average three or four shots, the banding disappears--and it's lost in random noise from ISO 400-3200 unless I do some serious pushing.

My camera's been in for service a number of times (I am _very_ rough on my gear), including a main board replacement and several calibrations specifically aimed at taming banding.

Banding is not unique to my camera, or one manufacturing run--it is a well-known flaw in the 5D Mark II and for people who run into it, it's a huge problem.

I'd come to the conclusion that Canon were more interested in developing the Video DSLR market than that old-fashioned single-frame photography rubbish that Nikon still seem to hooked up on...

Dear Ctein,


T Bannor (not a flag)

The comments to this entry are closed.



Blog powered by Typepad
Member since 06/2007