Would you buy a new full frame camera for $1999 with the 5D Mark III sensor?

[dawgfanjeff]

It's certainly one I'd consider, and I'm glad I've held off on a 5D3, so I can compare them.   I wonder if this will cause some buyer's remorse from 5D3 buyers (which might still include me).
I went to a 7D from Rebel world for it's AF and FPS.  They myriad other features are great, and I do use most of them, but I can't honestly say that for portraits or landscapes, with same lenses, the pics are any better than I got from my Rebel XT (350D), or XTi (400D)-and isn't that ultimately what what we're looking for?.  In fact, factoring in my expectations and disappointment, they are *worse*.

Weather sealing would be great, the convenience of the controls from the 7D is a must, but yeah.  I'd be interested.  I could put the extra $1500 to terrific use:) 

[Sycotek]

God no - had 2 5D3's go back - no way in hell

[vuilang]

it really depend on what other features it'll have. but if it same as rumored and it's $2000.. I'd rather buy another 5d3 for (~$3200)

[bdunbar79]

After I receieve the 1D X, I won't buy another camera from Canon until the actual, true replacement to the 1Ds Mark III arrives.  IMO, the 5D Mark III isn't it (obviously) nor is the 1D X.  However, I agree most photographers sadly didn't want a 1Ds Mark III back when it was retailing, as the domination of the sales of the 5D Mark II over that camera proved.  However, maybe it was just a cost thing (ie the 1Ds 3 was so far more expensive than the 5D2 and the 5D2 was very good?).  Either way, my selfish thoughts are for a DIGIC V processor 1Ds Mark III.  What could we call it? 

[te4o]

Furthermore many of the people saying that the 5D3 isn't that much better in ISO are not comparing the cameras correctly in likely two ways. Camera manufactuers generally make up their cameras ISO ratings out of thin air. There are a few acceptable ways of rating ISO but most manufacturers chose the least regulated ones which so the manufacturer can essentially make up their ISO numbers as they see fit and when the 5D2 was released the ISO ratings were very optimistic to say the least with ISO 12800 being actually around what is commonly accepted as ISO 7000. With the 5D3 Canon has been less optimistic so ISO 12800 is actually a "true" ISO of 10000. The point is that people are comparing cameras using their RATED ISO which are on different scales not ISO measured on the same scale. It's like comparing a car going 0-60 kph vs a car that's going 0-60 mph, which doesn't make sense. The other issue is that after speaking to several Canon Techs and Reps, it seems that Canon specially designed the 5D3 to produce very low noise JPG's for photographers that have to shoot in JPG. Part of the innovation that they introduced was software but they also engineered the hardware to help in the process to make the noise coming from the sensor have far less speckle noise and have a very gausian distribution. This makes computer programs able to distinguish from the noise easier so when you apply noise reduction there is a slight advantage to the 5D3 of around a quarter of a stop more than the RAW data would suggest.

The 5D3 sensor is actually very advanced and has 0.55 stop advantage over the D800 in noise at higher ISO. Due to the fundamental technology that Canon uses in it's sensors, they cannot improve the low ISO performance very much.

This is exactly my experience ! Thank you Radiating for explaining this because I couldn't. But I was getting a lot better denoise-ification from 5D3 files at all ISOs, even the low ones than from the 5D2. The 5D3 RAWs are more pushable, richer and as easy to sharpen as the 5D2.
I won't buy a cheaper FF because I have the 5D3 already but would have considered if that one came out first. 

[noisejammer]

At low speed and reasonable ISO, the 5D3 has essentially the same image performance as the 5D2... so why not continue offering a 5D2 at $2k... perhaps migrate to a 60D style body and cut the production cost.... add some arbitrary firmware, a tilt screen and built in flash to make it more attractive.

Maybe in a year you can offer a plastic model using the 5D3's sensor (but that strikes me as improbable.)

The thing is to sell lenses not bodies. It's lenses that keep your customers loyal.

[bdunbar79]

At low speed and reasonable ISO, the 5D3 has essentially the same image performance as the 5D2... so why not continue offering a 5D2 at $2k... perhaps migrate to a 60D style body and cut the production cost.... add some arbitrary firmware, a tilt screen and built in flash to make it more attractive.

Maybe in a year you can offer a plastic model using the 5D3's sensor (but that strikes me as improbable.)

The thing is to sell lenses not bodies. It's lenses that keep your customers loyal.

It'll probably end up being more of a 5D Mark II with a DIGIC V.

[LetTheRightLensIn]

Looks like Canon think that recycling the 5D Mark II's sensor technology continues to make sense.

Hogwash.

The technology in that sensor is approaching 4 years old now. Isn't it past time for something better?

Something with less read noise?

Something that is free of banding?

Yes and no. It sure was for the 5D3. But now just three months later the low-end brother of the 5D3 gets the exmor-killer? Why wouldn't they have just held the 5D3 back six months then. I think 5D3 owners would flip and all demand a new 5D3n soon and for free.

But yeah they sure as heck do need new sensor tech. They HAD been miles ahead of Nikon years ago, but they started talking arrogantly about sitting around on their throne having to do nothing and well I guess it has finally bitten them in the  , big time.

For low DR stuff it's not so bad at all, but with the modern sensors man you have a lot more freedom.

[LetTheRightLensIn]

If you don't think there is a noticable difference between the 5DII and 5DIII sensor then maybe you could tell me if you notice a difference between these photos:


That's a back to back comparison with the 5D3 resized to 5D2 size. The 5D3 sensor has a significantly better AA filter and so produces sharper images.

dangerous comparison, the lighting and focus has shifted just for starters

Furthermore many of the people saying that the 5D3 isn't that much better in ISO are not comparing the cameras correctly in likely two ways. Camera manufactuers generally make up their cameras ISO ratings out of thin air. There are a few acceptable ways of rating ISO but most manufacturers chose the least regulated ones which so the manufacturer can essentially make up their ISO numbers as they see fit and when the 5D2 was released the ISO ratings were very optimistic to say the least with ISO 12800 being actually around what is commonly accepted as ISO 7000. With the 5D3 Canon has been less optimistic so ISO 12800 is actually a "true" ISO of 10000. The point is that people are comparing cameras using their RATED ISO which are on different scales not ISO measured on the same scale.

This is true. If you compare them ISO for ISO the 5D3 gets an unfair disadvantage of nearly 1/3 of a stop.
So if it looks only 1/3 stop better SNR then it actually has more like 2/3rd stops better in reality (assuming DxO did things properly).

In any case here's a comparison between the 5D3 and the 5D2 rendered at the same resolution with the same ACR settings applied at a true ISO of 10084 for both cameras, this is a combination of a series of exposures using a method developed in consultation with the Cambridge Signal Processing Lab. I do consulting for a few camera review websites to develop testing methods and put this together for a project, with permission from all collaborators.

In any case feel free to spot the difference between the 5D2 and 5D3.

Feel even freer to spot the changes in focus and lighting and setup of the scene. 

The 5D3 sensor is actually very advanced and has 0.55 stop advantage over the D800 in noise at higher ISO. Due to the fundamental technology that Canon uses in it's sensors, they cannot improve the low ISO performance very much.

It's pretty feeble and outdated for late stage read noise so DR rather suffers at low ISO compared to Nikon and not even just the Nikon's that have exmor sensors, the D4 doesn't use any magic patents and it easily beats the pants off of the 5D3 for low ISO read noise. The low iso read noise on the 5D3 is actually among their worst efforts. Per photosite I think it may be worse than most to all they have done post 30D. It's even a tiny bit worse than on the older 5D2 model it replaced.

5D3 solved low iso horizontal banding but the vertical is as bad as with the 5D2 so it doesn't really make much difference in the end.

5D3 does have nicer noise at high iso than the 5D2 though since it is a lot more free of banding up there, etc. In certain scenes the usable difference is more than a simple SNR measurement implies.

Not so sure the 5D3 has 1/2 stop SNR over the D800 also don't forget the 5D3 has the most color-blind CFA ever used in a DSLR. It does hold together better from around ISO6400 and up though.

[LetTheRightLensIn]

That said I definitely do agree with radiating that the 5D3 sensor IS better than the 5D2 sensor in some ways.
Just did a quick test and it actually does seem to grab better detail when using same lens from same location to same target. I only compared the in cam jpgs so far, not the RAWs, but the 5D3 definitely showed more details on the target bill ISO100 test and the ISO6400 jpgs (NR off on both) definitely looked better from the 5D3, no large scale chroma subtle patterns looking ugly and same detail for less noise. I will compare RAWs next.

And DxO does give the 5D3 something like just about 2/3rd of a stop better SNR than the 5D2.

And 5D3 has less high ISO banding and less super large spread out many, many pixel sized chroma blotches than the 5D2 on darker areas of in cam jpgs (this probably holds for RAW too).

It does measure more color blind though for sure according to DxO that is worse.

It measures a touch worse low ISO read noise, that is worse, but not in any meaningful way.

5D3 has vastly less horizontal banding at low ISO but it has about the same vertical and so long as any banding is there it catches the eye so the practical difference in the end is not so much even though the 5D3 eliminates one type entirely.

Also looked at 7D and it sure as heck seems to have a reach advantage to me compared to the 5D2 despite the claims of a few that it's nearly a wash because the 5D2 pixels are better.

[takoman46]

where the option for "nope, I'm broke after buying a 5d3?!?"  anywho i'm in that camp, but I wouldn't be objective to picking up a camera like this to go as my second body if and when the time comes to retire my 7d for good.

"Like" 

I admit I'll be a little upset if the 5D Mark III sensor shows up in a $1999 body in the same model year. But then I'm sure that even if it is the same sensor, Canon would probably spend a few extra dollars to re-formulate the algorithms for sensor efficiency and/or performance to make sure it doesn't come close to the noise handling of the 5D Mark III. Realistically, I'd expect 5D Mark II sensor characteristics or maybe even lower at 7D sensor characteristics.

[LetTheRightLensIn]

Comparing ACR Process 2012 for 5D2 and 5D3 though I'm not seeing the detail advantage for the 5D3. With Neutral +4 sharpended, 0 NR in cam jpgs the 5D3 ones at ISO100 showed definitely more and crisper grabbed detail actually. Using ACR with all settings to 0, I can't see much, maybe 5D2 is a hair crisper per pixel but 5D3 has a hair more pixels, seems pretty similar detail and all between them.

Yeah at ISO100 honestly can't see much between the 5D2 and 5D3 for captured detail using ACR. So I don't see anything to match those saying it pulls in noticeably more detail than the 5D2. OTOH I definitely don't see anything to match those saying the 5D3 is soft and no good either. Honestly they seem pretty much the same using ACR, in that regard. Using in cam jpg neutral profil at +4 sharpening and 0 NR the 5D3 jpgs have less moire/artifacting and pull more detail than 5D2, at least for areas of very strong contrast (it might be the 5D3 fuzzes over low contrast detail more with in cam jpgs, at least with certain profile, didn't check that here though).

When reach limited 5D2/5D3 seem awfully similar but the 7D very clearly brings in much more detail than either.

And yes the 7D holds a clear advantage over the 5D2 when reach limited under any circumstance. Even at ISO6400 it pulls in noticeably more detail and if you normalize, even just using the simplest method, size of captured object between them the 7D delivers it with LESS noise and much less artifacting and more detail. Hands down 7D beats 5D2 when you are reach limited whether you shoot ISO100 or ISO6400. Compared to the 5D3 it also hands down pulls in more detail at ISO6400, however, in this case, if you compare them at same target size scale it doesn't have a noise advtange over the 5D3, it's a bit worse, although the detail is still a touch better and with less artifacting from debayer for sure. So I'd still say the 7D wins over the 5D3 too when you are reach limited be it ISO100 or ISO6400, although by a lesser amount at ISO6400.

All compared at 8MP, a non-reach limited comparison, the 5D3 looks cleanest at ISO6400 and then the 5D2 and then the 7D.

But note that this was a quick test, the raw converter profiles and on didn't apply the same color balance even after hitting white point tool on same point and so on at all and the exposures for same setting varied for each camera. At ISO6400 5D3 exposed the most, 7D the least. Didn't normalize for that at all.

[birdman]

I probably would if I decide to stick with Canon. sold my 5d2 for $1825, so feel like i got a great return on investment. Paid $700 to use it for two years.

I am sitting and waiting, but not so much on this new canon full-frame that we all damn well know WILL NOT be released this year. Absurd of anyone to think differently. No, I'm more waiting to see what the D600 offers and how this may further drive Canon's market share (and price of 5d3) down even more. I really want to stay Canon, having only sold my 5d2 (but I'm without camera for the time being). My lenses are sitting around doing zilch, but my Nikon glass hasn't been used in nearly a year either. I wish the 5d3 would drop further to $2,799. It'd be here sleeping in the bed beside me for that much. I did buy a D800 and resold immediately for nice little profit.

Even though I didn't take any pictures, I took it out and held it. Amazing camera, nice nice quality. My 5d2 felt like a cheap date comparatively speaking. Two worlds apart. If the 5d3 is more d800 and less 5d2, it may be the one for me. Yes, the low ISO file "banding" inherent in Canon sensors bothers the hell out of me. But I love their system, ease of use, and consistent metering. Also, the 24-105 & 70-200 4.0 IS are reasons by themselves to stay. See attached for IQ of 5d2 + 70-200 IS

[dilbert]

Comparing ACR Process 2012 for 5D2 and 5D3 though I'm not seeing the detail advantage for the 5D3. With Neutral +4 sharpended, 0 NR in cam jpgs the 5D3 ones at ISO100 showed definitely more and crisper grabbed detail actually. Using ACR with all settings to 0, I can't see much, maybe 5D2 is a hair crisper per pixel but 5D3 has a hair more pixels, seems pretty similar detail and all between them.

Yeah at ISO100 honestly can't see much between the 5D2 and 5D3 for captured detail using ACR. So I don't see anything to match those saying it pulls in noticeably more detail than the 5D2. OTOH I definitely don't see anything to match those saying the 5D3 is soft and no good either. Honestly they seem pretty much the same using ACR, in that regard. Using in cam jpg neutral profil at +4 sharpening and 0 NR the 5D3 jpgs have less moire/artifacting and pull more detail than 5D2, at least for areas of very strong contrast (it might be the 5D3 fuzzes over low contrast detail more with in cam jpgs, at least with certain profile, didn't check that here though).

That's just what you would expect when you look at the results of the DxO lab tests.

The comparison between JPEGs does not really reflect a difference in sensors, rather a difference in the JPEG rendering algorithms implemented in each camera.

[Cgdillan]

I will  buy the $1999 FF body. It's like a 5d mkiii and 7d fused into a 60d/7d body style. Will take amazing photos and will perform greatly. Not a fan of plastic but i happen to not be going into extreme weather and i take very good care of my gear. So it sounds perfect as a second body to my mkiii =-)