5D Mark III Information [CR1]

[Fleetie]

If I had a 46.7mp sensor at my disposal, that would be an image size of 12430x8286 pixels in size.

No, it wouldn't. That would be very close to 103 megapixels. Where did you get those numbers from?

[dr croubie]

If I had a 46.7mp sensor at my disposal, that would be an image size of 12430x8286 pixels in size.

No, it wouldn't. That would be very close to 103 megapixels. Where did you get those numbers from?

FYI: 10328 x 7760 is the Phase One IQ180, at 80MP it's the biggest MF sensor available, just beating Leaf Aptus' 10320 x 7752 (disregarding specialist telescope sensors and scanning backs).

[tt]

To me, "we want to see what the actual sales numbers will be for the 1D X and so determine how many customers we might be losing if we do not introduce a higher resolution camera"
says - you ain't seeing a Higher res camera until
a) The 1DX has launched
b) They have analysed the sales numbers

but they also said  "and if the projected profitability of a higher resolution 24×36 mm format camera will justify development, marketing, and manufacturing investments"

If it was nearer ready, why would they be talking about development costs?

"or if there is a sustainable market for the even greater costs of development, marketing, and manufacturing medium format cameras and lenses – an area where we have no internal expertise."

Ie - we haven't even started much if any development for a medium format camera, we haven't really done much yet on a higher res camera, and we're waiting out the 1DX launch and feedback.


Putting that to one side  - Canon on the intro to the 1DX did say that basically the 1DX was angled more for a niche, and that whilst they had some video options, the trade offs were in order ot make the camera better (eg the ethernet port meant losing some connections useful for video work, I think they mentioned in an interview).

What if they angle the 5D Mark II more to video (through incremental upgrades)? Some side effects will be useful (ISO, grain). Watching a course on making video with the 5D it seems there were needed work arounds for things that Canon could potentially fix, eg
- For your Z-finder - a higher resolution screen would help
- Having an AV jack that could at least take headphones
- Improving Live view (eg the issues of changing output format whilst watching when filming vs playback)

As photographers would like better low light performance, less grain banding, better ISO, better AF, so the video community would presumably want - less rolling shutter, better output connections.

If Canon just bumped the 5D's low light ISO capabilities, and their AF, maybe 3-4 more MP, would people buy? Seems there's a decent enough steady stream of customers.

What'll be more interesting is maybe in a few years time, seeing the low light power of the 5D MkII trickling down to more consumer products.

[jrista]

If I had a 46.7mp sensor at my disposal, that would be an image size of 12430x8286 pixels in size.

No, it wouldn't. That would be very close to 103 megapixels. Where did you get those numbers from?

Apologies, you are indeed correct. I kind of half-mathed/half-guessed the number, and I'm not sure what I did wrong. I think the correct image dimensions would be around 8382x5588 for a 46.7mp FF sensor. So certainly not as good as it sounded from a print standpoint before, but still a lot better than 21.1mp. Thats about a 30"x19" print, which is around 50% of the size I wish to print at (where as the native 21.1mp print is around 30% the size). From a scaling standpoint, scaling up to 2x is usually ok, where as scaling beyond 3x without special tools or better ways to fabricate information tends to produce terrible results.

[Jettatore]

What in god's name are you talking about jrista....

[jrista]

@Jettator: Sorry, correcting a previous post. I was explaining before why I find more MP useful, as I prefer to print at very large multi-foot dimensions. I was stating how scaling post-process holds nothing to having more native resolution, hence the reason I don't think the sensor race should end at 20-some MP.

[Jettatore]

You don't have to scale in post processing to print pictures larger.  You leave the pixel resolution (scale) alone and only tell the software what size in inches/mm's you want your print to be, and it will tell you before you print what DPI that will be.  If you are scaling the image or forcing the DPI before printing without some sort of special up-rezing technique that you have tested well to work, then you might very well be doing something wrong.

[jrista]

If you simply tell the printer to print at a certain size (WxH) and resolution (PPI), it will scale for you, however printer driver scaling is usually the worst kind possible. I always manually scale my images to exactly the correct PPI (not DPI, its a misnomer to use DPI to refer to printed pixels, as that refers to device units not pixel units) and physical size. I also fine tune sharpness for the specific target print size and paper (i.e. I don't care quite as much on a rough paper or canvas as I do on a smooth photo rag). I am pretty meticulous about scaling for my prints, as I prefer to keep as much detail as I can. I use the best scaling technique to eek out the best results as well, be it an iterative bicubic, s-spline or fractal approach. Despite all that, when scaling an image some 3 times or more its original size, its tough to keep it from ending up soft.

Hence the desire for more native resolution at an affordable cost. I do have to commend Pentax for their 645D medium format camera. I just hope someday that kind of technology is more within the realm of a prosumer budget.

[AprilForever]

Regardless of all the pro/con discussion of MP, image quality, sensitivty, noise, etc:

I shoot stock for a living. It is all I do.

I started with a 1Ds2. I switched to a 5D because it gave me better IQ and made me more profitable. I thought the 5D was good enough.

I resisted the 5D2 for over a year because it made me mad. Who needs more than 13MP for stock. Max size needed is double-truck, for which a 5D does just fine. (well, not really, because now I have customers wanting to do large display prints and even small murals suitable for close viewing distances)

But I continued to inspect the 5D2 files from time to time. Finally I jumped. I wish I had not waited a minute.
The 5D2, along with today's software, gives me files that are just plain better than the 5D files. Even when I reprocess my old 5D files with new software. And the improved quailty is something I need to keep my customers happy. The improved quality is visible in the sizes used in my business. Resolution, noise, smooth tonal transitions, croppability, etc.

21MP is enough to meet the current demands of my job and customers. But just barely enough. Almost everything I shoot is with controlled lighting, so I don't need great high ISO performance except when shooting aerials from a helicopter at dusk. And for that I can rent a 1DX.

What I really need now for my work is 25-30MP. A little more resolution and a little better noise and image quality performance. Plus better lenses. Lenses that are truly good enough. I'd like Canon to re-do the 24-105, because the existing version is so marginal (I've tried 5 different copies so far).

The quality of my work would benefit by moving to MF. But the cost of the total kit (bodies, plus backs, plus lenses) would make the move NOT cost-effective. It would hurt my bottom line significantly. So MF is out.

Therefore my vote is:

• 5D3 with 25-30MP, even if it costs up to $3k. But PLEASE give me better focusing with wider spread on the focus points.
  
• Better lenses (24-105 IS and 100-400 IS) that can easily more than handle 25-30MP. Even if they have to cost $1500-$3000 to be good enough. A $1000 lens that isn't quite good enough is NO bargain.

I like this too, except I would like to see it in a 7D...

[Dave Sucsy]

I like this too, except I would like to see it in a 7D...

April,
Why do you want this in a 7D instead of a 5D3 (or 1Ds4)?
Features of the 7D? Smaller size?

The only thing I don't like about the 7D is the small (area) sensor.
As a full-time pro, I'd really like:

• 2 axis leveling like in the 7D
  
• Better focusing like in the 7D
• A-Dep function. (the camera's computer can handle this much better than a pro can wing it)
  (this used to be included in the 1Ds series, for the photographers who need it most, but now it is only in low-end DSLRs for amateurs who use it least)
• built-in GPS (very inexpensive to implement, and essential for any traveling photography, available on many point-n-shoots)
  (Why??? must I take notes on where each shot is made, when my camera can and should do it for me??)
• built-in atomic clock (very inexpensive, and essential for accuracy for many pros)
  (Why??? must I spend half my life changing and synchronizing my camera's clock every time I change time zones , when my camera can and should do it for me??)
• voice memo, essential to quickly annotate shots to send back to the main office, available on almost all point-n-shoots
  (all photojournalists and most pro landscape photogs need to get this info to the office and the folks who keyword and caption the photos!!!!)
• built-in flash: essential for almost all pros for simple fill-flash, available on almost all point-n-shoots
• speech recognition shot annotation to embed identifying metadata and keywords into each shot, as necessary
• and other stuff too that I can't remember now.

Whoever will implement these essential productivity features first (Canon or Nikon) will win me over to their product line.


Dave

[mathino]

I want to see 18-21 Mpx sensor in 5D-isch body with 7D AF - Ill buy it. I really dont care much about video features as I am a stills guy. But sadly, there is no camera without video and I think there will be no camera without video in the future. I think many of us need/want better DR and noise in trade off for less video features. If you wants to shoot video you need to invest even more in good rigs for camera and lenses to make it at least a bit professional (not look of you shooting but look of captured video).

[Jettatore]

That makes a bit more sense than what you were talking about before jrista.  You were really making it sound like certain dimensions of printing were permanently linked to specific (edit:) pixel dimensions.

edit:  My language in describing this is also very bad.  I'm using DPI and PPI interchangeably even though I am always referring to PPI.  And I also often use the word resolution when I really mean pixel dimensions.

[catz]

I want to see 18-21 Mpx sensor in 5D-isch body with 7D AF - Ill buy it. I really dont care much about video features as I am a stills guy. But sadly, there is no camera without video and I think there will be no camera without video in the future. I think many of us need/want better DR and noise in trade off for less video features. If you wants to shoot video you need to invest even more in good rigs for camera and lenses to make it at least a bit professional (not look of you shooting but look of captured video).

Why are all (well not all, but many canonrumors readers) stills people expecting that better DR and noise is not needed for video? Because these are specifically needed for the video. And also inclusion of a video feature will not be away from any still feature because with the computer chips the cameras have nowadays in them it comes for almost free. In fact, it would be really stupid to not include the video because there would be absolutely no benefits on not including it.

Ideally a video-DSLR (actually it needs to be mirrorless to do that) will produce a still image sequence at speed of 25 frames per second at full sensor resolution like it produces stills today.  That would be truly revolutionary. And this still sequence could be directly ingested to Blender3D and it could be composited with CGI without any transcoding in between. That would be awesome workflow that most of the video guys do not have any slightest idea about and still guys don't know where they would need 25 fps still shooting. But I can tell you that would change the fundamentals of photography as there would be no division between stills and motion, motion would be just a sequence of stills in a steady rapid pace. I am quite sure this is the way of the future if this is not ruined by ignorant product managers who have no idea what they are going to do and what customers might want tomorrow but they only base on their assumptions what is there today and what the consumers are used to today. Video by definition (separated from stills) is so 1980s thing. For stills guys this high fps would produce huge advantage to take exactly the right moment and judge that afterwards the capture has been done instead of being lucky capturing exactly the right moment as it would be captured anyway in the frames in the between. If you don't need all the frames, you could just discard those which you don't need and keep the still that you want while the video people can keep the motion as it is. Photography and videography suddently will be the very same thing and it is not a bad thing, but a good thing. This would be very ideal for myself if I shoot for example airplanes. I can keep it rolling motion all the time and then I can later find the perfect shot from the frames of motion and it will be as perfect as a still.

25 fps is achievable with foreseeable technology. It will require next generation processors to do 60-120 fps at that resolution with low battery consumption but that will be the way of the future, no matter if Canon or someone else does it first. The super high capture rate would come with an advantage for not needing to use motion blur for the captured images to keep the motion smooth (in other words fast shutter speed can be used) and that would enable better still captures from the motion frames. The large amount of frames could be also alternatively combined in camera or in post production to reduce noise and create better dynamic range and it could be reduced to either still images or to lower fps video.

[dilbert]

Ideally a video-DSLR (actually it needs to be mirrorless to do that) will produce a still image sequence at speed of 25 frames per second at full sensor resolution like it produces stills today.  That would be truly revolutionary. And this still sequence could be directly ingested to Blender3D and it could be composited with CGI without any transcoding in between. That would be awesome workflow that most of the video guys do not have any slightest idea about and still guys don't know where they would need 25 fps still shooting. But I can tell you that would change the fundamentals of photography as there would be no division between stills and motion, motion would be just a sequence of stills in a steady rapid pace. I am quite sure this is the way of the future if this is not ruined by ignorant product managers who have no idea what they are going to do and what customers might want tomorrow but they only base on their assumptions what is there today and what the consumers are used to today. Video by definition (separated from stills) is so 1980s thing. For stills guys this high fps would produce huge advantage to take exactly the right moment and judge that afterwards the capture has been done instead of being lucky capturing exactly the right moment as it would be captured anyway in the frames in the between. If you don't need all the frames, you could just discard those which you don't need and keep the still that you want while the video people can keep the motion as it is. Photography and videography suddently will be the very same thing and it is not a bad thing, but a good thing. This would be very ideal for myself if I shoot for example airplanes. I can keep it rolling motion all the time and then I can later find the perfect shot from the frames of motion and it will be as perfect as a still.

Video shot at 1/500 and at a frame rate of 24fps will not look that great because objects moving fast will not have motion blur like they do in real life.

Video cameras do not capture video. They capture a series of stills and convert it into video. MPEG (or M-JPEG) is "Motion JPEG" - a means of compressing image data from frame to frame rather than needing to store complete stills images. AVI, etc, all do the same thing. From any digital video file, you can extract any exact frame that you wish.

btw, you've obviously not seen how professional sports photographers work today. I wouldn't be surprised if some go through 25,000 frames in a single day. 10 days of that and your camera needs to be back in the shop for shutter service. Well, at 400,000 maybe it becomes once a month.

[infared]

When is Canon going to make a 21MP (approx) camera that is JUST for still photographers and priced that way?
I am all happy for them..getting into the cine business...but what about us...the still photographers? Two completely different disciplines in my world. We need two completely different cameras. AND...the still camera should be substantially less expensive.
Does anyone agree with this...or am I just whistling Dixie?