5D Mk III vs D800/E, is the 5D3 better at anything?

[Northstar]

It was probably mentioned earlier but...in addition to the extra 2 fps, the 5d3 has many more cross type AF points and a much better LCD...and a great silent shutter which I'm finding quite handy.

In March I switched from Nikon to the Canon 5d3 because I had been wanting to move up to FF. I'm happy with the 5d3, it's a great all around camera but I wanted to stick with Nikon. I was disappointed when they announced the d800 specs. Like many other dark siders I was hoping for a follow up to the d700, with:
1. a bump in mp to 18-22
2. the Exmor sensor and more DR
3, the same AF
4. The same 8 fps
5. The same or better low light capability

Nikon would have killed canon with a camera like that...killed the 5d3 anyway, but I digress. So I switched to the 5d3 for its great all around capability.

With that said, I completely understand why some canonites are disappointed with this camera.....3-4 years since the 5d2 and canon didn't improve the IQ or meaningfully bump the MP's...and the result is only a 1/2 stop improvment in high ISO NR in raw. The 5d2 should have already had a great AF, so I discount that improvement somewhat.

So again, I get why some might complain about the 5d3, but for me, the 5d3 was pretty close to what I was looking for from Nikon, but they blew it IMO and now I'm a canonite.

 

[paul13walnut5]

What is a usable small aperture at 36MP pixel ptich? Do canon users with the 'mere' 22MP make more of hyperfocal focusing?

 

[Marsu42]

By definition you can't really have a contrasty image and high dynamic range. You get clipping but that is okay. The images have punch. HDR is either fake looking or flat for the most part because there is not enough contrast.

Well, there's hdr (like the horrible, outdated amateurish hdr look) and hdr (prevent clipping in natural looking images). And you are free to edit your shots locally to give more contrast to one part while more dr prevents white clipping or noisy shadows in another part. If it's about "punch" alone, a mobile phone camera would be a good choice

 

[markd61]

Go ahead and print a 24" x 36" poster print of a 5D3 image and a D800 image and see if you could blindly tell the difference. I bet you couldn't. How many of you actually EVER print bigger than that?

How many even print that big? I think I got a canvas done at 30x40" (entirely different resolution ball game) but even 15x10" is big for me these days.

I regularly print in A2 and have some commercial images on A0 - off 21mp

As always the subject matter is important. When I bought my 10D I promptly shot a tight headshot and printed it 40x60 ( I owned a photo lab at the time) and my clients were convinced it was medium format. Landscape scenes were another matter.
While I have to agree that with total attention to good technique the D800 will eke out a bit more res, I doubt that it would be easy to see even at 40x60 inches and certainly not at 16x20 and smaller.

 

[birdman]

My other point...people are all up in arms about Dynamic range...excuse me but aren't contrasty images much more pleasing to the eye in general? By definition you can't really have a contrasty image and high dynamic range. You get clipping but that is okay. The images have punch. HDR is either fake looking or flat for the most part because there is not enough contrast. Just my 2 cents...worth about 1 cent probably...lol.

HDR vs LDR is like 36 MP vs 22 MP. If you capture your image with HDR (or 36MP) you can always clip (downscale) to the lower quality. But you can never go the other way.

So if you prefer lower DR, you can shoot with a Sony/Nikon sensor, and clip the whites and/or add noise until the image looks like a Canon sensor (might have to experiment a little until you get the banding artifacts right).


Oh, I am a Canon shooter, probably will be until my 7D is old or broken, and I reserve my rights to critizise "my" brand whenever they are lagging behind. Right now, they are in terms of image quality. That might not matter to you or most buyers, but the fact is interesting nonetheless.

-h

Amen, brother. The only "qualm" I have with your analysis is that you may not know exactly how each camera differs (5d3 vs. d800) until you truly have each camera and shoot the same scene side by side. I (was) a 5d2 user until a few days ago. Sold it, and maybe I made a poor decision--because its just a hobby for me, anyway.

AF speed or AF points not critical for my style. But good, wide glass is...and this is where Nikon may have more of what I need. The 5d3 can be bought for $3150-3200 on Ebay all day long, I see. You may even find it for $3,000. It seems the D800 certainly spells out what I want in a newer DSLR. It's got less to do with 36Mp and more to do with higher IQ at lower ISOs. At least, that's what people say....

And finally, I do think Canon will be able to replicate the DR capabilities of the Sony sensor. Once they commit to solving this, what's left? Better wide glass? It came down to me not needing telephotos lenses as much as wide and medium FL. Canon has the better telephoto choices, and more of them. But every rose has it's thorn.

 

[RLPhoto]

Correct me if I'm wrong here. Look at DPreviews stop wedge here.

http://www.dpreview.com/reviews/nikon-d800-d800e/19

Without any additional ADL or any extra processing in camera and nothing on. The d800 gets the same DR as the 5D3? Its showing this, so does that mean all the nikon is doing is pulling more from the blacks and pulling more recovery by a software curve at the cost of noise? Its seems like fake DR to me.

If yes, This means that the nikon sensor is not doing anything better than the canon. Just adding a curve which I could do to the canon in post and get the same result right? This also confirms my suspicion that Bayer sensor have hit there peak already and new sensor design like the Fuji S5pro are needed to gain True DR.

If not, Is this test valid?

Just seems like a false 14-stop DR from nikon by this test.

 

[Ivar]

Correct me if I'm wrong here. Look at DPreviews stop wedge here.

http://www.dpreview.com/reviews/nikon-d800-d800e/19

Without any additional ADL or any extra processing in camera and nothing on. The d800 gets the same DR as the 5D3? Its showing this, so does that mean all the nikon is doing is pulling more from the blacks and pulling more recovery by a software curve at the cost of noise? Its seems like fake DR to me.

If yes, This means that the nikon sensor is not doing anything better than the canon. Just adding a curve which I could do to the canon in post and get the same result right? This also confirms my suspicion that Bayer sensor have hit there peak already and new sensor design like the Fuji S5pro are needed to gain True DR.

If not, Is this test valid?

Just seems like a false 14-stop DR from nikon by this test.

This test is JPG.

 

[RLPhoto]

Correct me if I'm wrong here. Look at DPreviews stop wedge here.

http://www.dpreview.com/reviews/nikon-d800-d800e/19

Without any additional ADL or any extra processing in camera and nothing on. The d800 gets the same DR as the 5D3? Its showing this, so does that mean all the nikon is doing is pulling more from the blacks and pulling more recovery by a software curve at the cost of noise? Its seems like fake DR to me.

If yes, This means that the nikon sensor is not doing anything better than the canon. Just adding a curve which I could do to the canon in post and get the same result right? This also confirms my suspicion that Bayer sensor have hit there peak already and new sensor design like the Fuji S5pro are needed to gain True DR.

If not, Is this test valid?

Just seems like a false 14-stop DR from nikon by this test.

This test is JPG.

But it mentioned that it could be applied in post. Does that mean the RAW without any extra processing is the same DR, but must be processed to achieve the 14 Stop DR?

 

[TTMartin]

Correct me if I'm wrong here. Look at DPreviews stop wedge here.

http://www.dpreview.com/reviews/nikon-d800-d800e/19

Without any additional ADL or any extra processing in camera and nothing on. The d800 gets the same DR as the 5D3? Its showing this, so does that mean all the nikon is doing is pulling more from the blacks and pulling more recovery by a software curve at the cost of noise? Its seems like fake DR to me.

If yes, This means that the nikon sensor is not doing anything better than the canon. Just adding a curve which I could do to the canon in post and get the same result right? This also confirms my suspicion that Bayer sensor have hit there peak already and new sensor design like the Fuji S5pro are needed to gain True DR.

If not, Is this test valid?

Just seems like a false 14-stop DR from nikon by this test.

This test is JPG.

Then either Nikon has a really crappy in camera JPG engine to lose a 2.7 stop advantage, or more likely DxOMark scores are not accurate for Canon cameras.

DxOMark uses their own RAW decoder. It is my belief that DxOMark's RAW decoder cannot fully decode Canon CR2 files, so DxOMark scores are not accurate for Canon cameras.

The other possibility is that DxOMark is just marketing shill for Nikon.

 

[Northstar]

Correct me if I'm wrong here. Look at DPreviews stop wedge here.

http://www.dpreview.com/reviews/nikon-d800-d800e/19

Without any additional ADL or any extra processing in camera and nothing on. The d800 gets the same DR as the 5D3? Its showing this, so does that mean all the nikon is doing is pulling more from the blacks and pulling more recovery by a software curve at the cost of noise? Its seems like fake DR to me.

If yes, This means that the nikon sensor is not doing anything better than the canon. Just adding a curve which I could do to the canon in post and get the same result right? This also confirms my suspicion that Bayer sensor have hit there peak already and new sensor design like the Fuji S5pro are needed to gain True DR.

If not, Is this test valid?

Just seems like a false 14-stop DR from nikon by this test.

This test is JPG.

Then either Nikon has a really crappy in camera JPG engine to lose a 2.7 stop advantage, or more likely DxOMark scores are not accurate for Canon cameras.

DxOMark uses their own RAW decoder. It is my belief that DxOMark's RAW decoder cannot fully decode Canon CR2 files, so DxOMark scores are not accurate for Canon cameras.

The other possibility is that DxOMark is just marketing shill for Nikon.

Speculation...I've speculated for some time now that dxo could (as a very small company) be influenced (if you know what i mean) to purposely set up their testing procedure to favor one company over another in exchange for.....

Both nikon and canon are large global corporations with significant resources...this "influencing" could easily be done...imo. Again, I'm speculating here, but we've observed time and time again in the global corporate world and in politics all over the world - money influences decisions. Example, in the news today in the US...mitt romney raises $100 million dollars for his presidential campaign...um, that was just for the month of july AND why do citizens and corps give him that kind of money, we know why right?? Also, a couple months ago walmart was accused of a large scale bribary campaign in Mexico....humm.

Just some stuff to think about....and again, im just speculating and providing raw opinion. Thanks.

 

[markd61]

It recently occured to me after Nikon announced that with a grip the D800 shoots 6 FPS that I can't find a single advantage to the 5D3 over the D800.
Thanks.

Tell us how you like your D-800 or are you just going to tell us specs and DXO scores?

 

[Ivar]

This test is JPG.

Then either Nikon has a really crappy in camera JPG engine to lose a 2.7 stop advantage, or more likely DxOMark scores are not accurate for Canon cameras.

DxOMark uses their own RAW decoder. It is my belief that DxOMark's RAW decoder cannot fully decode Canon CR2 files, so DxOMark scores are not accurate for Canon cameras.

The other possibility is that DxOMark is just marketing shill for Nikon.

Speculation...I've speculated for some time now that dxo could (as a very small company) be influenced (if you know what i mean) to purposely set up their testing procedure to favor one company over another in exchange for.....

Both nikon and canon are large global corporations with significant resources...this "influencing" could easily be done...imo. Again, I'm speculating here, but we've observed time and time again in the global corporate world and in politics all over the world - money influences decisions. Example, in the news today in the US...mitt romney raises $100 million dollars for his presidential campaign...um, that was just for the month of july AND why do citizens and corps give him that kind of money, we know why right?? Also, a couple months ago walmart was accused of a large scale bribary campaign in Mexico....humm.

Just some stuff to think about....and again, im just speculating and providing raw opinion. Thanks.

I haven't seen A SINGLE evidence Canon doing better or even the same in DR department no matter the RAW converter.

DxO has clear methodics, while maybe not covering all the aspects, still being a scientific approach and very usable for one knowing what to do with that info.
Opposite to the forum - one "thinks" and "believes", with absolutely no evidence.

P.S. more than sure, the next gen Canon will have a lot of DR, definitely more than Nikon today.

P.P.S Now this is what I believe - Canon just wanted to use the same old tech for economics reasons. This is why there is no high MP camera currently from Canon - stuffing more pixels in and being less attractive in IQ and asking a premium do not exactly fit together. They compete with versatility, which is quite an argument too. Absolutely fine with that, becuse for most applications the IQ is already so good that it does very well even not being the top notch.

 

[RLPhoto]

This test is JPG.

Then either Nikon has a really crappy in camera JPG engine to lose a 2.7 stop advantage, or more likely DxOMark scores are not accurate for Canon cameras.

DxOMark uses their own RAW decoder. It is my belief that DxOMark's RAW decoder cannot fully decode Canon CR2 files, so DxOMark scores are not accurate for Canon cameras.

The other possibility is that DxOMark is just marketing shill for Nikon.

Speculation...I've speculated for some time now that dxo could (as a very small company) be influenced (if you know what i mean) to purposely set up their testing procedure to favor one company over another in exchange for.....

Both nikon and canon are large global corporations with significant resources...this "influencing" could easily be done...imo. Again, I'm speculating here, but we've observed time and time again in the global corporate world and in politics all over the world - money influences decisions. Example, in the news today in the US...mitt romney raises $100 million dollars for his presidential campaign...um, that was just for the month of july AND why do citizens and corps give him that kind of money, we know why right?? Also, a couple months ago walmart was accused of a large scale bribary campaign in Mexico....humm.

Just some stuff to think about....and again, im just speculating and providing raw opinion. Thanks.

I haven't seen A SINGLE evidence Canon doing better or even the same in DR department no matter the RAW converter.

DxO has clear methodics, while maybe not covering all the aspects, still being a scientific approach and very usable for one knowing what to do with that info.
Opposite to the forum - one "thinks" and "believes", with absolutely no evidence.

P.S. more than sure, the next gen Canon will have a lot of DR, definitely more than Nikon today.

P.P.S Now this is what I believe - Canon just wanted to use the same old tech for economics reasons. This is why there is no high MP camera currently from Canon - stuffing more pixels in and being less attractive in IQ and asking a premium do not exactly fit together. They compete with versatility, which is quite an argument too. Absolutely fine with that, becuse for most applications the IQ is already so good that it does very well even not being the top notch.

Did you even view the stop wedge? ???

 

[Ivar]

This test is JPG.

But it mentioned that it could be applied in post. Does that mean the RAW without any extra processing is the same DR, but must be processed to achieve the 14 Stop DR?

This is like high ISO - some prefer more (or less) cleaning as opposed to default settings. Moving a shadow slider is no more difficult task. The key is if the source file breaks down during that process or not.

So yes, it needs to be processed - but it MUST contain the info to be extracted in the source.

A photo camera processor is always weaker than a PC, thus allows the latter using more complex and quality algorythms - this is why for the ultimate quality the RAW is preferred over in camera JPG engine (which is still good enough these days for majority of tasks).

 

[rumorzmonger]

I've had my D800E since the beginning of July, and here are a few unscientific observations I've made...

Dynamic Range - the 5D Mark III has a bit more dynamic range straight out of the camera, but the D800E files give you a whole lot more dynamic range to work with if you want to spend the extra time in post-processing.

Image Sharpness/Acuity - it's close, but I would give this one to the 5D Mark III if it were not for the lower pixel count; the amount of fine detail sharpness and resolution is better in the Canon files, but the Nikon file has 28% more real estate which helps it when you make larger prints (24 x 36"). I really expected to see better sharpness with no AA filter in the D800E, and Canon seems to be getting better performance out of their lenses with the new lens correction profiles loaded into the 5D mark III, so this comparison might surprise a lot of people.

Focusing - the 5D mark III is vastly superior in AF speed and, more importantly, in accuracy. The D800E isn't bad - it's about the same as the D700, but the Canon is in a different league altogether.

Metering - I would call it a draw... both cameras are pretty good, but neither one is perfect.

High ISO Noise - I have no idea - I normally shoot at ISO 100.

 

[bdunbar79]

I was prohibited from using a flash last night and shot several frams at ISO 12,800 on my 5D Mark III and only applied 20% NR and the noise was gone. I can't complain.

 

[tron]

I was prohibited from using a flash last night and shot several frams at ISO 12,800 on my 5D Mark III and only applied 20% NR and the noise was gone. I can't complain.

This is really good! I would never try 12800 with my 5DMkII. The top ISO I use is 6400.
Did you use Lightroom or photoshop?

I ask because it has been said that only Canon's DPP latest versions exhibit the maximum sharpness... (reported on other threads on this site a few months ago)

 

[Razor2012]

I was prohibited from using a flash last night and shot several frams at ISO 12,800 on my 5D Mark III and only applied 20% NR and the noise was gone. I can't complain.

This is really good! I would never try 12800 with my 5DMkII. The top ISO I use is 6400.
Did you use Lightroom or photoshop?

I ask because it has been said that only Canon's DPP latest versions exhibit the maximum sharpness... (reported on other threads on this site a few months ago)

I took a ton of shots at my daughter's recital and alot were @12800 and they looked great. Never had to use flash once. I was using my 70-200 2.8II with Av and auto ISO.

 

[bdunbar79]

I was prohibited from using a flash last night and shot several frams at ISO 12,800 on my 5D Mark III and only applied 20% NR and the noise was gone. I can't complain.

This is really good! I would never try 12800 with my 5DMkII. The top ISO I use is 6400.
Did you use Lightroom or photoshop?

I ask because it has been said that only Canon's DPP latest versions exhibit the maximum sharpness... (reported on other threads on this site a few months ago)

Even better. All I literally did was import it to Adobe Camera RAW, apply NR, and saved it to maximum JPEG. I'll show you later when I get home.

 

[wickidwombat]

you should give topaz denoise a go
you can apply differenct degrees of noise reduction to shadow, highlight and colour channels as well as custom noise reduction to deal with pattern banding if you have to push 4 stops of shadow

they do a 30 day trial version, its worth playing with