5D Mark III Description/Specs

[Alan Stankevitz]

This appears to be legit. Description and specs. taken from Ted's Cameras out of Australia:

Description Unsurpassed Image Quality

• 22.3 Megapixel Full Frame CMOS sensor
• DiG!C 5+ Image Processor
• ISO 100-25600 expandable to 50-102400
• Full HD Movie ISO 100-12800 (H:25600)

High Performance Operation

• 61-point high-density reticular AF (up to 41 crosstype points)
• 6.0 fps for high continuous shooting
• Intelligent viewfinder with approx. 100% coverage
• 3.2-type, approx.1.04m dot (3:2 wide) Clear View LCD II
• iFCL metering with 63-zone dual-layer sensor
• Shutter durability of 150,000 cycles

High end features

• Silent & low vibration modes
• Dual card slots (CF & SD)
• High Dynamic Range (HDR) Mode
• Multiple Exposures
• Comparative Playback function
• Improved durability & water and dust resistance

Specifications

Available Colours	Black
Megapixels	22MP
Sensor Size	36 x 24mm
ISO/Sensitivity	100 - 25600
Image Stabilisation	-
Autofocus Points	61 points
Lens Mount	Canon
LCD Size	3.2"
Touchscreen	-
Liveview	Yes
Viewfinder	Optical TTL
Min Shutter Speed	30 sec
Max Shutter Speed	1/8000 sec
Built-in Flash	-
Continuous Shooting Speed	6 fps
Self Timer	10 sec, 2 sec
Metering	Centre-weighted, Spot, Evaluative, Partial
Video Resolution	Full HD 1080
Memory Type	Compact Flash
Connectivity	USB 2, HDMI, Mic Input, Wireless (optional)
Battery	LP-E6
Battery Type	Lithium-ion
Charger	Includes Li-Ion Charger
File Formats	AVI, RAW, H.264, MOV, MPEG-4
Dimensions	152 x 116 x 76mm
Box Contents	Battery Pack LP-E6 .. Battery Charger LC-E6 .. AV Cable AVC-DC400ST .. Interface Cable IFC-200U .. Eyecup Eg .. Wide Strap EWEOS5DMKIII .. CR1616 Lithium Battery

[arash_hazeghi]

I hope this is not true.
It is really disappointing, the camera is neither high resolution nor fast. The only thing I see that I like is the AF which appears to be the 1DX system, pretty much the 5D2 plus better AF...not a big upgrade after 3 technology years. This camera cannot compete with Nikon D800 IMO unless it is significantly cheaper.

[Kaustubh Deshpande]

interesting. Canon was playing the MP card for all those yrs and Nikon wasn't. Looks like things have turned around completely. If this is true, that is. few more days and we'll know.

[Alan Stankevitz]

Rumor has it, it ain't cheap either. $3,500?

The problem with the D800 is the fps. 4 fps at max resolution, 6 with battery grip and cropped. It's got plenty of resolution, but it is slow fps-wise.

So goes the continuing search for the perfect bird camera...The Mark IV probably comes the closest. My only wish was that it was full-frame.

[arash_hazeghi]

Rumor has it, it ain't cheap either. $3,500?

The problem with the D800 is the fps. 4 fps at max resolution, 6 with battery grip and cropped. It's got plenty of resolution, but it is slow fps-wise.

So goes the continuing search for the perfect bird camera...The Mark IV probably comes the closest. My only wish was that it was full-frame.

Nikon D800 4fps @ 36 Mpixels, 5fps @ 1.2 crop (26 Mpixels w/o grip) 6fps 15Mpixel DX crop with grip, so you sacrifice speed for resolution. with this camera you sacrifice speed for...what?

Nikon is the faster camera at 144 Mpixel/sec. 5D3 is 120Mpixel/sec. The 3-year old 7D is 144Mpixel/sec so this was just a marketing decisions.

Asking $3500 for this camera is non sense, let them sit and hope that people buy it. this might be beginning of the end for Canon in this segment.

[Alan Stankevitz]

On a positive note, it "should" have excellent, low-noise capability considering it is 22mp/FF. The base ISO range of the 5D Mark III is 100-25,600, expanded is 50-102,400 whereas the D800 has a base range of 100-6,400, expanded is 50-25,600. The D800 has the same pixel density as the D7000, so I suspect that comparing raw images from the D7000 to the D800 should not be too different although newer technology might help the D800 a bit.

Until we're able to view raw images from both Canon and Nikon's new cameras, it's all speculation with regards to image quality.

[arash_hazeghi]

On a positive note, it "should" have excellent, low-noise capability considering it is 22mp/FF. The base ISO range of the 5D Mark III is 100-25,600, expanded is 50-102,400 whereas the D800 has a base range of 100-6,400, expanded is 50-25,600. The D800 has the same pixel density as the D7000, so I suspect that comparing raw images from the D7000 to the D800 should not be too different although newer technology might help the D800 a bit.

Until we're able to view raw images from both Canon and Nikon's new cameras, it's all speculation with regards to image quality.

of course SNR will be higher for 5D3 at pixel level however that is a naive way to compare them, both sensors collect the same number of photons so for a given output size (you need to down sample D800 files) the SNR will be identical and perhaps even better for Nikon because they usually have higher QE and lower FPN. Canon is really poor in controlling FPN and optical cross talk in the CFA.

[Alan Stankevitz]

It's all speculation until we can compare raw images from both cameras.

[Clemens Vanderwerf]

Canon press release expected tomorrow!

[Roger Clark]

On a positive note, it "should" have excellent, low-noise capability considering it is 22mp/FF. The base ISO range of the 5D Mark III is 100-25,600, expanded is 50-102,400 whereas the D800 has a base range of 100-6,400, expanded is 50-25,600. The D800 has the same pixel density as the D7000, so I suspect that comparing raw images from the D7000 to the D800 should not be too different although newer technology might help the D800 a bit.

Until we're able to view raw images from both Canon and Nikon's new cameras, it's all speculation with regards to image quality.

Remember, ISO does not change sensitivity, it only tells the range to digitize. For example, lets say a pixel is a 1-liter measuring cup. Here is what we get:

ISO 100: 0 to 1.0 liter
ISO 200: 0 to 0.5 liter
ISO 400: 0 to 0.25 liter
ISO 800: 0 to 0.125 liter
ISO 1600: 0 to 0.0625 liter
ISO 3200: 0 to 0.031 liter
ISO 6400: 0 to 0.0156 liter
ISO 12800: 0 to 0.078 liter
ISO 25600: 0 to 0.0039 liter
ISO 51200; 0 to 0.0019 liter
ISO 102400: 0 to 0.00098 liter

That last one is less then 1 mm, or only 20 drops.

Now a pixel in a camera may be full at around 60,000 electrons (thus recorded 60,000 photons) at ISO 100. Than at ISO 102400, one would be recording 0 to 60,000*0.00098 = 59 photons. And that is for the highlights. 18% gray would be about 59*.18 = 10.6 photons. Noise is square root the number of photons. so 18% gray would have a S/N ~ 10.6/sqrt(10.6) = 3.3. Yuk. Really Yuk! And that assumes no electronic noise contributes.

ISO only boosts smaller and smaller signals. High ISO, like above 6000 or so, are just marketing gimicks.

Roger

[Roger Clark]

That last one is less then 1 mm, or only 20 drops.

oops. That 1 mm should have been 1 ml (mililiter).

[arash_hazeghi]

it seems the specs are correct...bummer.

[Michael Gerald-Yamasaki]

ISO only boosts smaller and smaller signals. High ISO, like above 6000 or so, are just marketing gimicks.

Roger, you should take a look at a D3s 6400 image or the new D4 at 6400 (full size raws are now available). Really quite good.

Cheers,

-Michael-

[arash_hazeghi]

official specs

http://www.dpreview.com/news/2012/03...on-5D-Mark-III

[Roger Clark]

Canon says "A new photodiode structure with an increased photoelectric conversion rate increases the sensor's sensitivity by approximately 2 stops over previous models, meaning higher ISOs with the lowest noise of any EOS digital camera."

CMOS was already in the 30 to 40% quantum efficiency, so I don't see how they can get another 4x improvement in QE.

Roger

[arash_hazeghi]

Canon says "A new photodiode structure with an increased photoelectric conversion rate increases the sensor's sensitivity by approximately 2 stops over previous models, meaning higher ISOs with the lowest noise of any EOS digital camera."

CMOS was already in the 30 to 40% quantum efficiency, so I don't see how they can get another 4x improvement in QE.

Roger

I don't understand this either because the photodiode efficiency is already pretty high, most of the loss is from the color filter not from the pin diode. 2 stops is physically impossible because they need a QE of 120% (even taking the lower 30% number)

They said the same thing fro 1DX but, I think the real answer is this (From Doug's Brown interview with Chuck Westfall) ;)

DB: One of the advantages of moving to a full-frame sensor with a relatively low megapixel count is an improvement in high ISO noise levels. Canon is claiming a two stop gain in high ISO performance when compared to the 1D Mark IV. Will we see the full two stops when shooting RAW files, or is this limited to JPEG files only?
CW: To get the full two stops of improved performance you’ll need to shoot in the JPEG format. That’s not to say that high ISO performance isn’t significantly better when shooting in RAW, because it absolutely is.

Canon measure noise subjectively using JPEG output ;)

[arash_hazeghi]

here is a comparison of features between the D800 and 5D3



And here are Canon sample images for what it's worth, they are heavily compressed and lack fine detail thanks to the heavy noise reduction applied by Digic 5. In particular the high ISO images look like water color paintings and something that comes out of my iphone...I'm sure the RAW files will look much better but how much better compared to 5D2...It is just ridiculous that they even put these images online

http://web.canon.jp/imaging/eosd/samples/eos5dmk3/

[Alan Stankevitz]

I find it hard to believe that Canon posted some of those images. Did you guys check out the ISO 800 pic of the aurora? The brighter stars have a black halo. I'm not sure if this is due to their NR or is Canon repeating history? (When the Mark II came on the market, it had a problem with black halos around bright, pin-point objects. It was eventually corrected with firmware.) What was Canon thinking with some of those samples?

As far as 2-stop noise advantage, I'll bet dollars to donuts it's with jpg only. I find this really frustrating from both Canon and Nikon that when they taut their new cameras it's always with samples that have been touched up by NR. It isn't until it gets in the hands of the user that we see raw samples.

Alan

[Jay Gould]

5D2 at $2400 seems like a bargain if you do not have a full frame camera.

I was hoping for more camera for the cost; $700 more in Australia than in the USA after conversion taking into account shipping and taxes.

In Australia you need two 5Dx to qualify for CPS; the 7D does not count as it does in the USA.

The jury is still out whether I would buy a 5D3 to back up my 5D2.

[Colin Knight]

Your opinion of the Mark III is pretty harsh, Arash. Jeff Ascough loved it. I think it will have a huge place in wedding photography and production video, just like the Mark II. Looks like it is much improved in overall responsiveness, noise, FPS, and certainly AF - apparently the same as the 1Dx. Anyway, you might find his review interesting: http://blog.jeffascough.com/photogra...ii-review.html

[Jay Gould]

One major change is AE Bracketing which has been increased from +-2/3 frames to +-3/7 frames; big HDR impact.

[Colin Knight]

Yes, Jay. Also, it has in camera HDR processing. Definitely don't know what to think of that in the Mark III or the 1Dx. But it does look good on paper.

[Jay Gould]

Why would you want in camera processing? I do not want the camera to make any decisions other than to give me the best RAW file possible.

7 frames at 1/2 stops instantly is a big change from 3 frames.

dpreview just posted a series of low noise images that is rather impressive: http://www.dpreview.com/news/2012/03...kiii-isoseries

[Colin Knight]

That's what I'm saying- I don't know if I would want it or not. But I'm definitely interested in seeing how it works.

[arash_hazeghi]

Your opinion of the Mark III is pretty harsh, Arash. Jeff Ascough loved it. I think it will have a huge place in wedding photography and production video, just like the Mark II. Looks like it is much improved in overall responsiveness, noise, FPS, and certainly AF - apparently the same as the 1Dx. Anyway, you might find his review interesting: http://blog.jeffascough.com/photogra...ii-review.html

I am not sure, Jeff is paid to write these for Canon. I don't find his review objective as it is his opinion rather than any systematic measurements. Plus he is not an action photographer, wedding isn't a demanding subject in terms of AF, speed and ISO. I am not sure he understands/is familiar with the requirements of bird photographers.

Maybe it is a good camera for wedding, but the D800 is even better for cheaper.

D800 has better video capabilities too, if you want that, for $500 cheaper, namely uncompressed video output.

[Alan Stankevitz]

Jay...thanks for the link. This is more like what I expected to see. The sample images from Canon Europe and Japan are so heavily laced with NR it's not doing any justice to the camera.

Alan

[arash_hazeghi]

Why would you want in camera processing? I do not want the camera to make any decisions other than to give me the best RAW file possible.

7 frames at 1/2 stops instantly is a big change from 3 frames.

dpreview just posted a series of low noise images that is rather impressive: http://www.dpreview.com/news/2012/03...kiii-isoseries

In-camera processing is just a gimmick...

dpreview samples look much better than Canon's but they still have noise processing, plus it's a studio scene with no shadows. I can make the 5D2 files look like or better than those...One of the major reasons for lacking IQ in some of the images posted in the avain forum is too much NR. we all know what it does to bird feather details...

we have to wait to examine RAW files.

[Colin Knight]

Arash, heads up if you ever shoot a wedding- they are very demanding wrt ISO and AF due to low lighting, contrary to what you said above. I have never shot a well light reception, and only slightly more frequently a well lit ceremony.

Back on topic though, yes. Raw files are necessary to make the final determination on Mark III IQ.

[arash_hazeghi]

Arash, heads up if you ever shoot a wedding- they are very demanding wrt ISO and AF due to low lighting, contrary to what you said above. I have never shot a well light reception, and only slightly more frequently a well lit ceremony.

Back on topic though, yes. Raw files are necessary to make the final determination on Mark III IQ.

I have shot 3 weddings, so I disagree with you. I am not claiming I am as good as Jeff, but compared to what I do with AF wedding is a joke, period. Since when you hand hold a 500 for wedding and track an incoming avocet? I suggest you try flight shots and then compare

This is what I consider demanding AF, 700mm hand hold.



incoming avocet




I also have to say I was quite successful with 5D2 AF as long as I was using the center point with Cf.n-6 enabled. The new system is sure more configurable and faster but the old one was not as bad as people claim.

[Colin Knight]

Yes, I'll have to give this bird photography thing a try.

[Alan Stankevitz]

Has anyone found this curious? There are 5DM3 images showing up at Canon's websites as well as review websites, yet we still only have seen a handful of 1DX images.

[arash_hazeghi]

Has anyone found this curious? There are 5DM3 images showing up at Canon's websites as well as review websites, yet we still only have seen a handful of 1DX images.

I think 5D3 is actually ready to ship, you can order one at B&H now. The 1DX is still not ready.