5D MKIII specs

squig said:
This one sold me on it. I'm gonna pick one up next week.

[video=youtube;w47y-1EU8bE]http://www.youtube.com/watch?feature=player_embedded&v=w47y-1EU8bE[/video]
Click to expand...

May have sold me too. I don't like the color yet, but there is NO aliasing, and that does look sharp, much more so than the bits they've let our so far.

Amazing what a bad job both Canon and Nikon have done showing off the capabilities of these new machines.... almost feels like a conspiracy!
 
Kholi said:
What Andrew showed is what looked to be applied on in those promo videos, to my eyes. Sharpening. None of the other test videos look like the first Canon narrative examples even at quarter size and better compression.

It's a very distinct and almost obvious look when moving, like everything's been peppered with electricity.
Click to expand...

whadyoutalkinaboutwillis?

I can't see how it could be softer than the MKII. I read somewhere that the OLPF isn't as strong as the one on the MKII which explains the higher res I've noticed in RAW stills. I see a 1-2 stop improvement in overexposure latitude. Check out the shot at 00:15

So far Canon have addressed just about every issue I had with the MKII. The rolling shutter could be better but I understand the 1DX has a sensor read twice as fast as the MKIII so things are looking good for the same improvement with the C-DSLR. I earn some of my living shooting stills so a MKIII B-cam is a no brainer. C-DSLR or C300 A-cam.
 
Didn't say it was softer, said that the videos were sharpened that canon put out (the puppet one) and that's pretty much what Andrews still shows.

Nobody's saying not to buy it, just looks like another very marginal "update" to the video side of a still camera. I dunno about seeing anything near a full extra stop being retained in video... but, okay.
 
squig said:
This one sold me on it. I'm gonna pick one up next week.
Click to expand...

Yeah, that was looking pretty acceptable in the sharpness department to my eyes. I'm primarily a stills photographer so I don't think I can go wrong with the 5DIII for stills as well as my adventure into cinema. I was almost impulsive enough to jump on the C300, but beside taking a bit more gusto out of my recreation account than I would like, it was canon's first foray into the cinema field, and even they admit they wanted to get it to market as soon as possible. Before I drop that kind of coin, I'd like to see what they produce now that they have the development time for the C350. And as far as the CDSLR is concerned, I'm pretty sure it will be $6K+, and I don't think I want to venture into 4K video just yet. Bottom line, the 5DIII will hold me over the next few years as this evolution of the large sensor cinema cam unfolds. Things are going to get better and cheaper rather quickly.
 
Last edited:
Cosimo Bullo said:
Shouldn't say no aliasing, but very minimal compared to Mk2.
Click to expand...

And looks like less aliasing and moire than the Nikons. I can't see how the 36MP D800 could have less moire. The codec looks solid and Canon appears to have caught up to Nikon in the low light dept. The D800 probably has the edge in latitude but the MKIII looks better at high ISOs (in RAW stills) The crop mode would be nice but it's not something I'm going to cry about. I think Canon have delivered a package beyond my expectations which is a welcome surprise. We're still not clear whether the ALL-I codec is 4:2:2 anyone? That would be the icing on the cake for me with no clean HDMI out. The horizontal level will be handy for anamorphic shooting and I'm really happy about manual control of NR.
 
Kholi said:
I dunno about seeing anything near a full extra stop being retained in video... but, okay.
Click to expand...

You serious?

The background in the MKII shot is totally blown out on my screen. The colour profiles don't look exactly the same in the 2 shots though. Big difference in the skin tones.
 
squig said:
You serious?

The background in the MKII shot is totally blown out on my screen. The colour profiles don't look exactly the same in the 2 shots though. Big difference in the skin tones.
Click to expand...

That could be attributed to a ton of factors, the most obvious how or if they changed picture profiles in conjunction with how the new sensor behaves....

Assuming it's straight-up latitude is ... well... I guess optimistic. Which isn't a bad thing.

I see about 9 stops as usual in the videos. Nothing new.
 
Kholi said:
That could be attributed to a ton of factors, the most obvious how or if they changed picture profiles in conjunction with how the new sensor behaves....

Assuming it's straight-up latitude is ... well... I guess optimistic. Which isn't a bad thing.

I see about 9 stops as usual in the videos. Nothing new.
Click to expand...

The MKII had 11.2 in the Zacuto shootout. You know something I don't? The Sony/Nikon sensors have improved by 2-3 stops since the MKII came out so it's conceivable that Canon have kept pace.
 
squig said:
The MKII had 11.2 in the Zacuto shootout. You know something I don't? The Sony/Nikon sensors have improved by 2-3 stops since the MKII came out so it's conceivable that Canon have kept pace.
Click to expand...

I'd like to see 11.2 stops of usable DR out of an MKII video or MKIII video. More reason to go back to full frame, no?

I don't, though. Ever. Don't shoot charts or wedges. I know if I shoot Epic I can capture about 12~13 of usable data reliably. That's RAW 5K data.

HDRx is a different story of course. What you're saying is that MKII has 11.2... then 5D MKIII in video has 12.2~13.2 of usable DR (1 to 2 stops that you see), almost as much or equal to RAW 5K.
 
Last edited:
squig said:
The MKII had 11.2 in the Zacuto shootout. You know something I don't? The Sony/Nikon sensors have improved by 2-3 stops since the MKII came out so it's conceivable that Canon have kept pace.
Click to expand...

2-3 of those 11.2 stops were ruined by noise though and not useable at all..
 
Kholi said:
I'd like to see 11.2 stops of usable DR out of an MKII video or MKIII video. More reason to go back to full frame, no?

I don't, though. Ever. Don't shoot charts or wedges. I know if I shoot Epic I can capture about 12~13 of usable data reliably. That's RAW 5K data.

HDRx is a different story of course. What you're saying is that MKII has 11.2... then 5D MKIII in video has 12.2~13.2 of usable DR (1 to 2 stops that you see), almost as much or equal to RAW 5K.
Click to expand...

Well we haven't seen any video files out of the camera but going on the jpeg stills and the fact that the ALL-I codec is putting out frames at a comparable size to a high quality jpeg and NR is applied to video, I see and Canon have confirmed a 2 stop improvement in high ISO and I would expect given murphys law that there's a comparable improvement in latitude too. Even the 7D has another stop over the MKII in the highlights. It's not RAW but just like the driftwood hack made a leap in IQ on the GH2 I think the new sensor and ALL-I codec will give the MKIII a healthy IQ boost too. I'm not suggesting that it will be a match for the GH2 at 800 ISO but at 3200 or 6400 it should smoke it and there are lots of occasions when 1250 ISO hasn't been enough for me because the shallow DOF of full frame cancels out some of the low light advantage. How does the Epics overexposure latitude compare to the RED 1?
 
squig said:
The MKII had 11.2 in the Zacuto shootout.
Click to expand...

yeah, but watch it again. They differentiate between technical 11.2 stops and USABLE stops. And the 5D had much less usable.
I have a C300 which technically only has .8 stops more and I can tell you 5DMK2 vs. C300 is like night and day, literally.
So 12 stops or 11.2 stops technically mean not much in the real life, because sensitivity, noise, codec etc are a huge part of the equation.
 
None of the hacking business increased the threshold for usable DR to me. What it changed was the amount of detail present within any given frame, and how it's retained when motion comes into play.

Again, I don't see any of the those claims in a real shooting scenario applied to video. I didn't see the 7D with a stop over the MKII in any way, I don't see 11.5 or more stops of usable DR in any of the MKIII footage. 11.5 looks like RED MX's capabilities. None of these cameras are even there yet. It surely is not messing around in Epic or Alexa territory.

But, either which way.

It looks like to me that Canon's saying if you want a video oriented device and you're okay with a crippled one, we're gonna price our EOS C camera at 4.5K (That's my guess for EOS C price, but maybe 6) and you can buy this one, or you can spend more, get more, with the C300.

And even then... well... C300 seems like a mistake more than ever now... mark my words in this post.
 
Kholi said:
And even then... well... C300 seems like a mistake more than ever now... mark my words in this post.
Click to expand...

haha, well if you look strictly at IQ and work only in controlled environments maybe (and even then it is arguable IMO).

But I can give you the basic difference between the MK3 and the C300.

First time you take a 5DmkIII you will say: "Good! About time! Finally!". When you watch he footage you say: "Ok! Phew."

First time you grab a C300 you say: "Wow! What?! No way! Ahhh, clever. Incredible. Haha! Wootwoot. Hmmmmm....that is sweet." When you watch the footage you will say "Incredible! I need to dial that sharpness down. Whoa, that codec is awesome!"

So when you need the soft, over saturated FF look and you are happy with all the bitching and rigging the DSLR shooting gives you, fine. I see not many big steps in the MK3 so far especially with the steeper price, but at least it seems safer to use it with fewer dealbreakers in the image.
The C300 has so many features that are helping you as a cameraman and producer you need to consider those features in the price as much as the IQ alone.
 
Here's the DR that my Flaat picture styles get on my 550D (T2i):
* Flaat_1 is about 9 stops (same as the Canon picture styles)
* Flaat_2 is about 10 stops (same as Technicolor CineStyle and Marvels Cine), and I'd say it's all usable DR
* Flaat_3 is about 11 stops, and shows quite some noise in the pushed-up shadows (like shooting with 1 stop higher ISO)
* Flaat_4 is clearly above 11 stops, and shows a hell of a lot of noise in the heavily-pushed-up shadows (like shooting with 2 stops higher ISO)

these are the wedge charts:
http://www.similaar.com/foto/flaat-picture-styles/technical-tests-1.html

these are some stills tests where you can see how this extra DR looks in the real world:
http://www.similaar.com/foto/flaat-picture-styles/real-world-tests-2.html

and this is a video test where you can see the amount of noise this expanded DR gets you:
http://vimeo.com/34943260


notice that "clearly above 11 stops" is what dxomark measured in RAW stills from this camera http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/676|0/(brand)/Pentax/(appareil2)/680|0/(brand2)/Nikon/(appareil3)/645|0/(brand3)/Canon
so with these picture styles I'm getting in video mode just as much maybe-usable-maybe-not DR as you get in RAW stills mode

with the extremely low-noise that the 5D3 is showing, and the improved codec (expanding so much DR across an 8-bit codec is always going to be problematic, but the higher-bitrate all-I mode *could* make this better) I would expect to get all its RAW-stills DR, and without these noise issues - if that's 14, I'd be a VERY happy camper


Lucas Pfaff (@camerarick) is shooting a short with my extended-DR picture styles this weekend - he said "I need any stop I can, it's a subway on the outside with BRIGHT windows :) I'll keep you posted if it worked good!" - let's see how it goes...
http://twitter.com/#!/CameraRick/status/180778879047307267
 
Last edited:
Kholi said:
What Andrew showed is what looked to be applied on in those promo videos, to my eyes. Sharpening. None of the other test videos look like the first Canon narrative examples even at quarter size and better compression.

It's a very distinct and almost obvious look when moving, like everything's been peppered with electricity.
Click to expand...
Right. But would Canon do that? I don't think they need. I think the proof is in the pudding. It is also a matter how the workflow and all the variables such as glass' choice as for instance, are done. And it is not hard to tell the difference from distinct shots samples also available there.
 
Emanuel said:
Right. But would Canon do that? I don't think they need. I think the proof is in the pudding. It is also a matter how the workflow and all the variables such as glass' choice as for instance, are done. And it is not hard to tell the difference from distinct shots samples also available there.
Click to expand...

Yeah I think you'll find almost all of the examples so far are shot with EOS glass. EOS vs Leica glass is night and day.
 
You must log in or register to reply here.
Back
Top