Change of the guards… reflection on 100 MP CMOS

[ErikKaffehr]

Hi,

Back in the old times Velvia ruled the world. There were some great CCD devices but at a very high price.

Suddenly,Canon's D30 arrived and changed the world. Many observers, me included believed that 24x36 film corresponded to to 6MP, but the D30 has challenged that view, because it delivered very smooth images at 3 MP. But those images looked really good.

Back in 2006 I was shooting a 6 MP CCD camera with a Sony sensor. Just a year later Sony introduced the Alpha 700. It had twice the resolution, but foremost it had an Exmoor CMOS sensor. That camera combined 12 MP with low noise and excellent DR. The 6MP camera was long behind.

In just a few years, APS-C and 24x36 moved to CMOS, and CCD was history.

I am pretty sure that it is what we see in MFD now. MFD has been a CCD resort. The previous year, Hasselblad, Phase One, Leaf and Pentax released models based on CMOS designs, using the same technology Sony introduced back in 2007. This year, Phase One introduced the 100MP full frame CMOS backs using the technology sony developed in 2007, obviously in a much refined form.

What I would think we see is a change of the guards. MFD jumping on a new technology curve. The new technology obviously makes the old one obsolete, but it does not render the old technology irrelevant. The old CCD based technology is perfectly good at making great pictures. But, the new technology expands usability and flexibility.

So, what I see is that the future of MFD is best technology, and from now on that is based on CMOS.

Best regard
Erik

[voidshatter]

Hi,

Do you have predictions of when the IQ3 100MP will become obsolete? Maybe in 3-5 years? Sony could enlarge the IMX251 BSI sensor used in the A7R-II at any time. If the Sony A6100 comes with an APSC BSI at 36MP then it is pretty safe to bet that 135 format will reach 80MP by around 2018.

[Doug Peterson]

Hi,

Do you have predictions of when the IQ3 100MP will become obsolete? Maybe in 3-5 years? Sony could enlarge the IMX251 BSI sensor used in the A7R-II at any time. If the Sony A6100 comes with an APSC BSI at 36MP then it is pretty safe to bet that 135 format will reach 80MP by around 2018.

"Obsolete" is a strange word to use here, assuming we're talking about actual photography rather than spec-hunting and gearophilia.

We still have a large number of clients (much to our chagrin!) using Phase One H20 backs made in 2001 or H25 backs made in 2003. They are still supported in Capture One (and look better than the day they were released because of improvements in raw processing), are nearly bullet-proof durability wise, and still produce a great image with excellent color and tonality. By no means is it a "cutting edge" electronics device; DXO would be underwhelmed by it's engineering dynamic range. But it's still a great camera if it does what you need it to do*.

Likewise I strongly suspect many of the IQ3 100mp backs we've sold the last few weeks will still be in regular use in a decade. High res, big sensor, great color, great support, five-year warranty, great lenses... these things won't evaporate the moment an IQ# XXXmp (whatever comes next) knocks it off the "flagship" pedestal.

*H20/H25 backs are tether-only, lack an LCD, and are slow to shoot, and only work well at very low ISO, so were purchased almost exclusively for studio still life / product / catalog shooters. But for those shooters it still works like a champ; the only major change in the last 15 years of using it has been you now need to use a Thunderbolt hub in between the computer and back since FireWire has evaporated from most computer lineups.

[torger]

Hi,

Do you have predictions of when the IQ3 100MP will become obsolete? Maybe in 3-5 years? Sony could enlarge the IMX251 BSI sensor used in the A7R-II at any time. If the Sony A6100 comes with an APSC BSI at 36MP then it is pretty safe to bet that 135 format will reach 80MP by around 2018.

I think we need some more CMOS MFD history before one can make good guesses about that. If we look at the CCDs the IQ360 is using about the same sensor as the P65+ (some minor changes), it's 7 years. It could be the case that this 100MP sensor will be used in a similar period of time re-released in several new models with usability upgrades.

...or we see a speedup in technology cycles so we get new sensors more often, and I think that is the more likely scenario but I know too little about the challenges to scale up a CMOS sensor to be any other than just a layman's guess. If there's a lot of cost and custom manufacturing gear associated to scaling up I guess the sensor need to be sold over more years to generate proper profit for Sony.

It shall also be interesting to see the megapixel race in the 135 format, will it settle a while at around 50MP, or will it push further real soon? I see both scenarios as likely so I don't even dare to guess.

I agree with Doug that "obsolete" is a pretty strong word to use, as many of us use old gear and are happy with their performance. But at some point it will be replaced with a newer sensor in new products. If CMOS sensors will be re-used in new products over as many years as the CCDs remains to be seen...

[voidshatter]

"Obsolete" is a strange word to use here, assuming we're talking about actual photography rather than spec-hunting and gearophilia.

We still have a large number of clients (much to our chagrin!) using Phase One H20 backs made in 2001 or H25 backs made in 2003. They are still supported in Capture One (and look better than the day they were released because of improvements in raw processing), are nearly bullet-proof durability wise, and still produce a great image with excellent color and tonality. By no means is it a "cutting edge" electronics device; DXO would be underwhelmed by it's engineering dynamic range. But it's still a great camera if it does what you need it to do*.

Likewise I strongly suspect many of the IQ3 100mp backs we've sold the last few weeks will still be in regular use in a decade. High res, big sensor, great color, great support, five-year warranty, great lenses... these things won't evaporate the moment an IQ# XXXmp (whatever comes next) knocks it off the "flagship" pedestal.

*H20/H25 backs are tether-only, lack an LCD, and are slow to shoot, and only work well at very low ISO, so were purchased almost exclusively for studio still life / product / catalog shooters. But for those shooters it still works like a champ; the only major change in the last 15 years of using it has been you now need to use a Thunderbolt hub in between the computer and back since FireWire has evaporated from most computer lineups.

Could you please remind me the inductive pricing of the PowerPhase FX+ (130MP large format scan back) which was released 13-14 years ago? I think saw it being listed for hundreds of dollars last year. In my opinion depreciation is the most accurate term to define the degree of obsolete. You could still make stunning images with it but I wouldn't bother with it at all.

I am not speaking negatively against the IQ3 100MP. I am ordering my upgrade and it is truly the beast to succeed the CCD backs. I am just talking about the natural rules of semiconductors - thanks to Moore's Law or whatever pace of development, it is pretty safe to bet that anything contains electronics can become obsolete and become e-waste in a decade.

Can we still make stunning images with the CCD backs? Yes. But with the CMOS backs it could be way more easier, and with the CMOS backs we could make images that cannot be made by the CCD backs.

I am keen to see technology advancement. On the other hand, technology advancement means emptying out wallet faster.

If no one can challenge Sony's sensors then it's good that the IQ3 100MP can last a long time, just like no one can challenge Intel's CPUs and we can continue using a 5-year old i7 2600K in 2016.

[voidshatter]

I think we need some more CMOS MFD history before one can make good guesses about that. If we look at the CCDs the IQ360 is using about the same sensor as the P65+ (some minor changes), it's 7 years. It could be the case that this 100MP sensor will be used in a similar period of time re-released in several new models with usability upgrades.

...or we see a speedup in technology cycles so we get new sensors more often, and I think that is the more likely scenario but I know too little about the challenges to scale up a CMOS sensor to be any other than just a layman's guess. If there's a lot of cost and custom manufacturing gear associated to scaling up I guess the sensor need to be sold over more years to generate proper profit for Sony.

It shall also be interesting to see the megapixel race in the 135 format, will it settle a while at around 50MP, or will it push further real soon? I see both scenarios as likely so I don't even dare to guess.

I agree with Doug that "obsolete" is a pretty strong word to use, as many of us use old gear and are happy with their performance. But at some point it will be replaced with a newer sensor in new products. If CMOS sensors will be re-used in new products over as many years as the CCDs remains to be seen...

Suppose we see a new full frame digital back in 2019, with BSI design, at 200MP (i.e. same pixel density of rumoured 36MP APSC BSI) with great angular response (maybe even compatible with the discontinued Schneider 28XL), and a 135 format Nikon/Sony at 80MP (with 8k video @ 120fps), what would you think about the current IQ3 100MP?

[ErikKaffehr]

Hi,

Difficult to predict the future is.

One thing I have noticed in Doug's tabletop test shot is that the Schneider lens used in the test keeps up very well with the IQ3-100MP. So, I think it could do well with higher resolutions.

On the other hand, I also feel it controlled aliasing pretty well, so I feel that those 100 MP are quite adequate for that lens.

What I may also have seen in Doug's test is that some of the other lenses did not keep up with Phase One/Schneider. The Pentax 55 was quite horrible in the top left corner and neither the Sony 55/1.8 or the Canon 58/1.2 L was even close. For that reason I don't think we will see a lot of competition from 24x36 going to 80 MP.

It is quite obvious that now that MFD has essentially the same sensor technology it can fully enjoy the benefit of size.

It seems that the 60 MP CCD sits on a sweet spot regarding shifts on technical cameras, but I would think that for DSLRs, the only disadvantage of the new CMOS is cost. My guess is that the CMOS backs are cheaper to make, so I guess the IQ3-100MP is quite profitable for Phase One.

Another interesting question is if Phase One has any competition and if so what that competition does. Leica seems to be happy with it's small 37.5 MP sensor. Will Hasselblad release a 100 MP CMOS at a competitive price? Will Pentax enter the game? Are the lenses good enough? Some talk about Fuji jumping into MFD, would they be the first to make a mirrorless MFD camera? Perhaps Phase One has an exclusive contract with Sony?

Anyway, this development is very interesting, and it seems that Phase One has been successful in implementing it.

Best regards
Erik

Hi,

Do you have predictions of when the IQ3 100MP will become obsolete? Maybe in 3-5 years? Sony could enlarge the IMX251 BSI sensor used in the A7R-II at any time. If the Sony A6100 comes with an APSC BSI at 36MP then it is pretty safe to bet that 135 format will reach 80MP by around 2018.

[voidshatter]

It seems that the 60 MP CCD sits on a sweet spot regarding shifts on technical cameras, but I would think that for DSLRs, the only disadvantage of the new CMOS is cost.

For pixel-peeping purposes, it might be even better to shoot with the 100MP CMOS unshifted with keystone correction in post-processing, than to shoot with the 60MP Dalsa CCD shifted. I find it difficult to justify the 60MP now, other than cost efficiency.

[ErikKaffehr]

Hi,

Good point, I have not tested.

What I would also add that I feel that a sensor with live view may make a lot of sense on a technical camera.

I have a Hasselblad Flexbody for my P45+, but I found it to cumbersome to use in the field. I live view back would be far more functional. I may have bought a Hasselblad VF 50c back if it was not for the crop factor and lack of wide angles.

Best regards
Erik

For pixel-peeping purposes, it might be even better to shoot with the 100MP CMOS unshifted with keystone correction in post-processing, than to shoot with the 60MP Dalsa CCD shifted. I find it difficult to justify the 60MP now, other than cost efficiency.

[torger]

For pixel-peeping purposes, it might be even better to shoot with the 100MP CMOS unshifted with keystone correction in post-processing, than to shoot with the 60MP Dalsa CCD shifted. I find it difficult to justify the 60MP now, other than cost efficiency.

I did a quick surfing and to my surprise I didn't find any good tests/comparisons on keystone vs lens shift. It should be possible to make a formula that states if you shift X mm with focal length Y you need focal length Z and you will loose Q in resolution.

You lose more the wider field of view you need to capture. It would be interesting to compare say the 40HR 12mm shifted on 60MP with a wider angle keystone corrected center frame from the 100MP. I think the 100MP would lose with a fair bit, but I haven't done the math or seen any test...

[eronald]

For pixel-peeping purposes, it might be even better to shoot with the 100MP CMOS unshifted with keystone correction in post-processing, than to shoot with the 60MP Dalsa CCD shifted. I find it difficult to justify the 60MP now, other than cost efficiency.

Alpa told me at Photokina that they thought software would at some point replace movements.
That point in time might have arrived.

Edmunds

[ErikKaffehr]

Hi,

Something on my mind is mirrorless MFD. In essence, technical cameras are mirrorless MFD. They sort of need better integration and live view. Now, with the new Sony 50 and 100 MP sensors live view is here.

Hartblei HCam B1 and later Alpa FPS has demonstrated the concept of fully integrated digital camera, but both devices use a modified Mamiya focal plane shutter.

The MFD market seems to be small, but quite profitable. I would not be to much surprised if some company introduced mirrorless MFD. Fujifilm may be a good candidate. They make the lenses for Hasselblad, and I would bet they make a lot of other parts.

Best regards
Erik

[torger]

Can't guarantee this is correct, but with the help of an angle of view calculator (http://www.radical.org/aov/);

645 digital fullframe 54x41mm landscape orientation on 40mm 12mm shifted up requires a 25mm lens center frame to cover the same area with keystone correction. On the 25mm lens 50% of the image is cropped away leaving 50 megapixels.

So it's not too bad, assuming you find a 25mm lens that is very sharp.

Really you won't be doing "keystone correction", that's just a waste as far as I understand, you gain nothing from pointing the camera up as you need to stretch and crop the sides anyway. You just shoot with the camera level and have enough angle of view to fit your subject and then crop to what you want. Oh well, you can point the camera up when you don't have enough angle of view, but then when you keystone correct you lose width, for example your 4:3 image ends up as a 1:1 aspect ratio.

I can just imagine how boring the shooting process becomes just running around with an supersharp ultrawide pointing it in the general direction and do all composition at home... but that may be where we're headed.

[Miyata610]

"Obsolete" is a strange word to use here, assuming we're talking about actual photography rather than spec-hunting and gearophilia.

We still have a large number of clients (much to our chagrin!) using Phase One H20 backs made in 2001 or H25 backs made in 2003. They are still supported in Capture One (and look better than the day they were released because of improvements in raw processing), are nearly bullet-proof durability wise, and still produce a great image with excellent color and tonality. By no means is it a "cutting edge" electronics device; DXO would be underwhelmed by it's engineering dynamic range. But it's still a great camera if it does what you need it to do*.

Likewise I strongly suspect many of the IQ3 100mp backs we've sold the last few weeks will still be in regular use in a decade. High res, big sensor, great color, great support, five-year warranty, great lenses... these things won't evaporate the moment an IQ# XXXmp (whatever comes next) knocks it off the "flagship" pedestal.

*H20/H25 backs are tether-only, lack an LCD, and are slow to shoot, and only work well at very low ISO, so were purchased almost exclusively for studio still life / product / catalog shooters. But for those shooters it still works like a champ; the only major change in the last 15 years of using it has been you now need to use a Thunderbolt hub in between the computer and back since FireWire has evaporated from most computer lineups.

I'm one of those annoying H25 users.  In conjunction with my ancient 503CX and/or 553ELX.  Yep, it's better than ever now with C1 v9.  I use it in the field quite happily with the last of the firewire MBPs, yes it will power it quite happily and I cary a couple of Lemar external batteries it will last all day.

[Doug Peterson]

I'm one of those annoying H25 users.  In conjunction with my ancient 503CX and/or 553ELX.  Yep, it's better than ever now with C1 v9.  I use it in the field quite happily with the last of the firewire MBPs, yes it will power it quite happily and I cary a couple of Lemar external batteries it will last all day.

That's the only downside of selling great stuff that lasts a long time!

Of course I'm mostly joking. It is deeply satisfying to see such gear last such a long time in use. Of course if you ever feel the desire to upgrade we're glad to help you upgrade to the system that will last the next decade .

[Miyata610]

That's the only downside of selling great stuff that lasts a long time!

Of course I'm mostly joking. It is deeply satisfying to see such gear last such a long time in use. Of course if you ever feel the desire to upgrade we're glad to help you upgrade to the system that will last the next decade .

Although Doug, it would be great if you could heavily discount the motor cable that you still have in stock for us oldies 

[torger]

Some people shoot today with film to get a more traditional shooting experience. In a way using an old CCD back gives a very similar experience, no particular features in the back just capturing what the lens projects onto it.

I think film will survive perhaps indefinitely into the future in some form, because it's just "real photography" and it's low tech enough to keep some sort of production going with relatively small means. I wonder however if there in the future will be a place for the "low tech" digital back which as the first digital backs just functioned as a drop in replacement to a film magazine. My guess is - probably not - but if it would continue to exist it's probably there I'd place my money.

When the choice becomes between a very high tech digital camera, perhaps with a fixed ultra-high resolution wide angle lens and the rest achieved through digital cropping and automatic focus stacking, and shooting traditional large format film, it will be a more difficult choice than it is today. I started out with photography late and I haven't shot film seriously at all so far, but the more high tech cameras become, the more I feel attracted by film.

[Torbjörn Tapani]

Moore's law don't apply to camera sensors. Miniaturization doesn't get you much beyond what is already possible with current CMOS tech. And the size of the chip determines cost for the most part. This new generation of MF sensors have mature CMOS technology developed elsewhere. I don't see it being updated quite so frequently as smaller sensors. It will not be obsolete any time soon.

[ErikKaffehr]

Hi,

Moore's law is actually a set of laws. The shrinking of component sizes is one of those laws, but it is also about increasing chip sizes. One of the rules says that it is economically feasible to make larger chips.

The other side is that there is a lot of shrinking of pixel sizes for cell phones and other small format technologies like surveillance and machine vision. Jim Kasson has pointed to a moderately technical paper at Chipworks describing some new developments: http://www.chipworks.com/sites/default/files/Fontaine_0.1_The%20State-of-the-Art%20of%20Mainstream%20CMOS%20Image%20Sensors.pdf

Sensor makers seem to address issues like crosstalk with separation walls and shallow pixels, like 1.5 micron deep. That kind of technology may solve a lot of issues with beam angle sensitivity of todays pixel designs.

It seems that APS-C is moving to 36 MP and 4/3 to 24 MP. Scaled up to 24x36 that would correspond to 81 resp 96 MP.

Modern pixels are built with an extra capacitor connected to the photodiode, thus increasing well capacity. That may eliminate/reduce of the disadvantages of small pixels, name reduction of dynamic range.

So, many interesting things going on.

Best regards
Erik

Moore's law don't apply to camera sensors. Miniaturization doesn't get you much beyond what is already possible with current CMOS tech. And the size of the chip determines cost for the most part. This new generation of MF sensors have mature CMOS technology developed elsewhere. I don't see it being updated quite so frequently as smaller sensors. It will not be obsolete any time soon.

[vjbelle]

For pixel-peeping purposes, it might be even better to shoot with the 100MP CMOS unshifted with keystone correction in post-processing, than to shoot with the 60MP Dalsa CCD shifted. I find it difficult to justify the 60MP now, other than cost efficiency.

Simple keystone correction is fairly easy in software but complex vertical and horizontal lines that would exist in a nodal pan stitch are a nightmare and best shot with movements...... at least that has been my experience.

Victor