A Canon Move into MF: Not in the near future (from NL)

[pedro]

Just read it over at NL. So, the 18 MP 5D3 becomes even more likely 8)
http://www.northlight-images.co.uk/canon_medium_format_2ff.html

 

[wellfedCanuck]

Interesting that they've decided to make the body out of Lego...

 

[dr croubie]

I'd be interested to see how much of a boost in sales of the Pentax 645D and other MF suppliers comes from 1Ds3 owners who had been holding off for a 1Ds4, and now making the jump.

Me too, i'd be guessing a fair few billboard-printing pros. But how many will jump to nikon for the extra 3mp, and/or how many will jump to actually buying into a new system, new lenses, new flashes, be it hassy/p645/phasemiya/s2, or will they keep all their canon glass and just rent the big-boys for certain assignments?

"While we hear the voices, we want to see what the actual sales numbers will be for the 1D X and so determine how many 10,000s of customers we might be losing if we do not introduce a higher resolution camera and if the projected profitability of a higher resolution 24x36 mm format camera will justify development, marketing, and manufacturing investments , 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."

In other words, we're going to see how much money and how many customers we lose with the 1DX gamble before we commit $1 of R&D to beating the already-released D3X, A900, A850, A77. By that time (at least middle of next year, a very very early estimate), if they only start R&D then, we might see a high-mp camera by the end of 2013 if we're very lucky.

Or to put it another way, canon engineers/marketing are one (or more) of the following:
- They have very very good faith that the 1DX will sell well, be it because the 18MP IQ is actually really really good (or they just trust canon fanboys with lots of glass not to switch to nikon of mf).
- They have very very good inside info that the next nikon is a POS worse than a 1Ds3 in every respect (not very likely), or that it will cost a lot more (like p645-price range).
- They are a bunch of absolute idiots who don't know their customers.

I'd vote #1, but #3 is niggling at me a bit (there's already a billboard-printer on these forums who's gone).

 

[wellfedCanuck]

Interesting that they've decided to make the body out of Lego...

5 Smites?! Guys, that was humour. Take a closer look at their photo:

Man, tough crowd...

 

[wickidwombat]

i dont get it how did people manage to print billboards 5 years ago when you could only get 8 or 10 MP?
i'm pretty sure i saw billboards back then

 

[Caps18]

i dont get it how did people manage to print billboards 5 years ago when you could only get 8 or 10 MP?
i'm pretty sure i saw billboards back then

I'm guessing they had to take multiple pictures and stitch together a panorama... Or they took pictures and when you are driving down the road and are over a hundred feet away, the pixel size isn't quite as important.

 

[traveller]

Interesting that they've decided to make the body out of Lego...

5 Smites?! Guys, that was humour. Take a closer look at their photo:

Man, tough crowd...

I wouldn't worry about the smites, they are from people who are either too lazy or too illiterate to post an actual reply.

 

[dr croubie]

i dont get it how did people manage to print billboards 5 years ago when you could only get 8 or 10 MP?
i'm pretty sure i saw billboards back then

Leaf had a 33MP Aptus75 back in 2005, Phase One's 40MP P45 came in 2005, Hasselblad's H3D 39 and 50MP backs came at the end of 2007.
There were lower-MP backs before this, but at 20-30MP you're probably better off scanning (or using a good old-fashioned enlarger) from good-quality 120-film...

 

[funkboy]

This is great news & confirms what I'd already suspected. We've all seen the surveys over the last few years asking what photographers want in their next DSLR. Pretty much after the 50D came out, many more people started ticking the "high ISO performance" rather than the "more MP" box. Canon seems to have listened. The "more fps" crowd is also fairly vocal, & Canon seems to be using fps as their competitive advantage point against Nikon (& others). Also, very high resolution CMOS sensors are not so great for video.

I particularly appreciate the tidbit on the new teleconverters. The line is "if you're an APS-H shooter & like the extended reach of that format, well, the new teleconverters are better than the old ones & will give you back the reach you gave up by switching to FF, and the 1Dx will AF as well or better with an f/4 lens as the old ones did with f/2.8 glass". Unfortunately, the marketing department doesn't really seem to be conveying this particular message, and the no-more-f/8-AF issue doesn't help either (e.g. the 2x TC becomes useless on anything but f/2.8 or faster glass).

The 1Dx, then, is a calculated risk for Canon. They are betting that they'll make more money from the faithful that move up to a very high IQ 12/14fps FF body that costs $2500 more than the 1DIV (plus the few working at f/8 that will upgrade their lenses so that they still have supported AF, & those that buy TCs for the reach), and that the higher unit volume of the now single-model 1D line will drive costs down over time & keep them competitive for years. In this crazy economy, they'll have to get as much mileage out of the 1Dx as they did with the 1DsIII (and the 5DII, which continues to sell like hotcakes 3 years after its introduction).

Can't afford it & need the fps? Well, then I guess the 7D's your camera, but at least the 7D's there. I'll go so far as to say that the 7D (& certainly the 7D II) is a calculated part of their APS-h phase-out strategy in the 1-series. If they incorporate the mirror & processing developments of the 1Dx in the 7D II then we're looking at a mighty fine sportsman's crop camera in 2013.

It's a similar gamble on the resolution side of the house for the traditional 1Ds customer: "we're giving you 18 very high quality gapless megapixels & a completely redesigned metering system rather than 25+ MP". They've obviously figured that the few pros that they loose to Nikon (or medium format) will cost them less than maintaining a two-model 1D lineup.

Personally, I think it's the right decision. I'll add another reason that hasn't gotten a lot of airtime in the discussions revolving around the 1Dx:

Pretty much anything but the best L lenses would be massively outresolved by a ~30MP FF sensor, especially in the corners where higher resolution APS-c (& even APS-h to a lesser degree) sensors don't have to worry about it. They're well aware that wide-angles are their weak spot & have updated several of them to compensate, but at 30MP anything with less resolution than five-star-rated L glass just wouldn't cut the mustard. Such a camera would expose every single flaw in their aging lens lineup that they use as a major marketing point for its breadth & versatility (which unfortunately also makes it harder to maintain over time). We'd just continue to see more & more reviews like this one...

A very high IQ 18 MP FF camera has the opposite effect. That's the same pixel density as a 7mp APS-C sensor. Not much to worry about there; if a lens looked fine on a 1DsII then it should look fine on this (& the same assurance cannot be made for the same lens on the 1DsIII and 5DII).

Unfortunately that leaves me with even less hope that we'll ever see lenses like a 28mm f/1.8 USM II, but lenses are a lot harder to predict than bodies so who knows...

 

[Edwin Herdman]

Not really surprising in light of the 1D X news, which is pushing image quality over pixel counts and being (perhaps optimistically) touted as appropriate for studio work.

In any case, not really sure what they'd gain by slogging after the ever-diminishing MF market. It's also nice not to have to worry about their support for TS lenses on EOS...

 

[moreorless]

It's a similar gamble on the resolution side of the house for the traditional 1Ds customer: "we're giving you 18 very high quality gapless megapixels & a completely redesigned metering system rather than 25+ MP". They've obviously figured that the few pros that they loose to Nikon (or medium format) will cost them less than maintaining a two-model 1D lineup.

I'd guess that Canon may also view that market as far more uncertain, the high end ISO,FPS, AF market is very much a known quantity where as an ultra high megapixel FF body is much more unknown. Will the pro market look to save money rather than buyign superior MF? is there a strong market for amature users?

They might I spose also believe that on a 5D the 1DX sensor could provide value rather than top end performance. 6 fps without dual processors, less need to buy expensive lenses for both resolution and appature with superior ISO performance.

 

[Edwin Herdman]

I'd guess that Canon may also view that market as far more uncertain, the high end ISO,FPS, AF market is very much a known quantity where as an ultra high megapixel FF body is much more unknown.

I agree, but replace "unknown" with "declining." That said, the weeding out of the marketplace probably indeed makes it iffy for Canon to try to enter that segment, and it wouldn't realistically do much for their volume leading trend. At this point, I think the companies that are sticking with servicing MF users are doing it out of love.

A MF offering with glass capable of resolution numbers like Canon's SLR lenses would be fantastic, but also hugely expensive, probably even by current MF standards (look at some Schneider lens prices!), as Canon almost certainly wouldn't be able to produce in enough volume to really ameliorate the costs (at least in my off-the-cuff judgement).

The people that are still around for MF are pretty well served these days, although it has taken a severe shake-down of the industry (this article optimistically describes the last two years as "busy ones in the medium format industry," but read what comes immediately after that: "There are no exact numbers, but it's likely that sales volume has been reduced by some 50% over what they were in 2007." !) to get to the point where anybody can come close to leading the market. It is basically a small market and R&D is not getting committed to the point you'd expect from DSLRs which are driven by volume. The relative competitiveness of DSLRs (35mm and APS-C-based) in IQ (i.e. they aren't completely blown away by cameras with much bigger sensors) through most of the years you can do comparisons is a sign of this (granted, DSLRs and MF aren't close in IQ for any year, but if current tech were always used in MF sensors that gap could be much larger).

Will the pro market look to save money rather than buyign superior MF? is there a strong market for amature users?

Depends on the pro. Some need to print billboards. If money grew on trees, I wouldn't mind having a MF system to play with, but it doesn't, and I don't. Amateurs are already limited by the economy to buying $1000 lenses, instead of $2500+ lenses (prime lenses) that just have a bigger image circle.

They might I spose also believe that on a 5D the 1DX sensor could provide value rather than top end performance. 6 fps without dual processors, less need to buy expensive lenses for both resolution and appature with superior ISO performance.

Yeah, that's the assumption I have made.

Your last point here also reminds me of something that follows from my second to last paragraph: The fact that you can freely switch (well, more or less - ignoring crop factors and the f/5.6 AF limit for the 1D X) camera bodies with the EF mount is a big deal. On MF, sometimes you're even stuck with anachronistic lens boards (somehow I don't think they're weather sealed) and other often-clumsy setups. Even if you can assume the lenses screw on well, for MF there are still competing systems, and they're not guaranteed to be interchangeable.

 

[torger]

I think it will be much a matter of competition. Say if Nikon provides a 36 megapixel fullframe in D800, and a D4 similar to 1DX, which does seem quite likely. It will be more attractive to invest in a system that gives you more options.

Sure 36 megapixel fullframe will not be as good as a 80 megapixel medium format tech camera with rodenstock lenses, but perhaps it brings you far enough into medium format territory to make it a viable option.

For fine art and photo books (300 - 400 ppi capable high quality prints) 36 megapixel can provide a visible quality improvement. You can still argue that 18 megapixel is good enough, which it kind of is, but if you can get visible improvement why not use it? Ever dreamt about medium format but it is just too pricey? Why not invest in a 135 system that can bring you close to that.

It seems odd not maxing out what 135 fullframe format can do in studio and still life when the technology exists. Problem with Canon is that they are so large and have so many customers locked in so they don't really need to lead the market. They can let Nikon or someone else do the high res experiment, and if it is a hit they can follow on later. Let the competition take the risk.

 

[funkboy]

It seems odd not maxing out what 135 fullframe format can do in studio and still life when the technology exists <...> Let the competition take the risk.

Yep, that's their gamble with the 1Dx.

I'm not sure how well the D3X sold, but given its price point & the economic situation I'm guessing "not as well as Nikon thought it would" is the answer. I've seen plenty of tourists, pros, etc using D700 and D3S over the last few years, but I've never seen a D3X in the wild. That's not to say that they're not out there, but given how good cameras like the D3s and 5DII are, you've really got to *need* the extra resolution D3X to justify the extra cash, even for pros. I imagine that some Nikon folks that really did need the extra resolution & couldn't afford it jumped ship to the 5DII to get the extra resolution for less than half the cash.

A Canon D3X resolution beater would likely cost even more than the D3X, which is pushing dangerously close to 645D territory. & as I mentioned before, without lenses like Nikon's 14-24 to take full advantage of the extra resolution, they're really not in a position to play the megapixel king anyway.

One other argument I've read is "the 1DX's pixels are so good that if you need to print larger the images upres much better than 5DII or 1DsIII". We'll see how that works out in practice, but I wouldn't be surprised to see a feature in the new version of DPP that comes with the 1DX specifically tailored to upresing.

If it's true then that's another great play in Canon's playbook, as it allows finished output at final resolutions as good or better their previous cameras while not pushing the lens stable as hard.

 

[Picsfor]

i dont get it how did people manage to print billboards 5 years ago when you could only get 8 or 10 MP?
i'm pretty sure i saw billboards back then

There is a nice show in the UK called the Gadget Show.
Not so very long ago - they tested the capabilities of 2 cameras, of which 1 of the tests involved creating the largest size print they could.

For the record, the print covered a quarter of the long wall of the Tate Modern (i've never seen any billboards that big) and the pics done themselves proud.

The cameras used were the 7D and D700 (if memory serves) - but the trick was the printers used, who had software designed to scale a 12mp and above up to that size for printing.

So, billboards do not require MF - though i'm sure it makes life a little easier.

WE have to remember, it is the photographer they buy in for the shoot - not there kit, unless the manufacturer are sponsoring the shoot!

 

[Bob Howland]

A Canon D3X resolution beater would likely cost even more than the D3X, which is pushing dangerously close to 645D territory.

Only if you use a high end body. That was the mistake that both Canon and Nikon made with their 1Ds3 and D3x. Wedding and event photographers are routinely using D700's and 5D2's. Does anybody think that most studio and landscape photographers routinely need any more ruggedness than wedding photographers? If a 21MP 5D2 could be sold 3 years ago for $3000, then there is no reason why a 36-40MP studio/landscape replacement can't be sold for the same price today (although I'd pay an extra $500 to get state of the art focusing).

 

[torger]

I don't think a high res of 1DX (1DXs?) needs to cost more, that is in terms of manufacturing costs. Perhaps development costs of the sensor would be high though and lower sales volumes and therefore would justify a higher price, but on the whole I'd think development cost would not be a problem even if a relatively low volume is sold. A 1DXs would indeed surely sell much lower numbers than a 1DX, just as D3x compared to D3s.

I think the talk about 1DX good pixels for upsizing is just marketing speak, clearly not from anyone that makes prints. If a sensor is not sufficiently close/past lens/diffraction resolving power there will be jaggies on the pixel level, which needs smearing to look good if upsized, and it is still hard to make it look good at close inspection. AA filter helps some though. A file good for upsizing to low PPIs should be relatively soft at pixel level so even if pixels get large (low PPI) you won't see individual pixels or artifacts from upsizing algorithms trying to hide them. If you're instead printing at high PPIs, being exactly at what the printer can do, sharp pixels are great (assuming printer does some smearing), if printer res is exceeded soft or sharp pixels does not matter of course. To summarize, a 36 megapixel somewhat soft file is more flexible concerning print sizes than a pixelsharp 18 megapixel image.

With 18 megapixel you're in a good range for magazines, and newspapers can do with lower than that. High quality photo books and fine art prints is the main use of higher resolution as I see it.

 

[seacritter]

I've dreamt about MF, like everyone else. But the main reason for my thought of moving up is the jump from 14bit resolution to 16bit. You gain almost 2 stops of dynamic range. As a landscape photographer, this is huge. Image the extra dynamic range in a super high ISO shot. You would be seeing so much more detail in the shadows in a low light shot. I'd rather see them step up to 16bit before they jump up to more megapixels.

Maybe noone remembers the jump from 12 bit to 14 bit, but it was a technological break-through...

 

[torger]

I've dreamt about MF, like everyone else. But the main reason for my thought of moving up is the jump from 14bit resolution to 16bit. You gain almost 2 stops of dynamic range. As a landscape photographer, this is huge. Image the extra dynamic range in a super high ISO shot. You would be seeing so much more detail in the shadows in a low light shot. I'd rather see them step up to 16bit before they jump up to more megapixels.

Maybe noone remembers the jump from 12 bit to 14 bit, but it was a technological break-through...

Unfortunately the 16 bit sampling in medium format sensors are not really effective, since the noise levels are too high, real dynamic range is around 12-13 stops and would do with 12-13 bits. The jump in quality seen from 12 to 14 bit was probably more due to reduced noise levels than increased bit depth in sampling.

Per pixel, an APS-C Nikon D7000 (14 bit) actually has slighthly better dynamic range than an Phase One IQ180 (16 bit), but since the latter has 80 megapixels when averaging them all over a print the IQ180 wins. Possibly it is advantageous to sample the noise with those extra two bits to get a better result in the averaging process, don't know, but the effect is probably very very small if so. The larger area also allows for gathering more photons, reducing photon shot noise (not relevant in bright areas, but can be in dark), about a 2/3 stop gain from 135 fullframe to medium format. So these days, the medium format advantage comes only from the larger sensor area.

 

[seacritter]

Per pixel, an APS-C Nikon D7000 (14 bit) actually has slighthly better dynamic range than an Phase One IQ180 (16 bit), but since the latter has 80 megapixels when averaging them all over a print the IQ180 wins. The larger area also allows for gathering more photons, reducing photon shot noise, about a 2/3 stop gain from 135 fullframe to medium format. So these days, the medium format advantage comes only from the larger sensor area.

Exactly... Bumping up the Canon or Nikon sensor to 16bit would be a greater leap towards MF than just adding MPs. If Canon is truly heading in the direction of quality over quantity, this might be next...