Canon to Continue Using Canon Sensors in DSLRs

[Loswr]

Q: Canon released two cameras at Photokina – the EOS 7D Mark II and PowerShot G7X. One thing we’ve learned is that the sensor in the G7X is not made by Canon. Does this represent a new philosophy at Canon?


A: We select the best sensor, whoever the manufacturer is. That’s our policy.

'Best' can be defined in many ways. In this case, 'best at making a profit for our business' is likely the most apt definition.

 

[AcutancePhotography]

"A note: during a recent interview with DPreview, when asked about the G7 X sensor, a Canon representative Marketing person said:
... A: We select the best sensor, whoever the manufacturer is. That’s our policy."


Fixed.

 

[PhotoCat]

"A note: during a recent interview with DPreview, when asked about the G7 X sensor, a Canon representative Marketing person said:
... A: We select the best sensor, whoever the manufacturer is. That’s our policy."


Fixed.

+1 LOL!

 

[David Hull]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

On the other hand, it's entirely possible that Sony facilities are used to fabricate the sensor but sensor design comes from Canon.

what about industrial espionage?
would that not be a problem?

buying a finished sensor is no problem but handing out all your technical stuff to a competing company so they can produce your latest and greatest?

i mean sony is not like other companys who could produce sensors for canon, they are a direct competitor.

it´s not like apple is handing out chip production to TSMC.
TSMC does not produce phones or tablets.

It is a bit the other way around actually. Sony would have to tip their hat to Canon in that they would have to provide all of their process models and libraries to enable Canon to design on their fab. That said, unless Sony is geared up to support fabless semiconductor companies, they may not really have the support in place to allow Canon to use their fabs.

 

[jrista]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

If this is true, Jrista is probably over the moon now. ;D

On the other hand, it's entirely possible that Sony facilities are used to fabricate the sensor but sensor design comes from Canon.

Not really. It's never been my goal to see Sony own the sensors in the vast majority of cameras. That would be terrible, as it would mean a monopoly, which would just stifle innovation in the long run.


What I want is for Canon to step up their game and compete. I like the idea of Canon's total in-house integration. It just seems like they are focusing on everything but sensors, and have been for some time. I find that sad. Canon's lack of competition in the sensor marketplace is going to bite them in the ass at some point (and I believe that Nikon already gained a good deal of market share in the high end DSLR market thanks to their improved sensor IQ).


Canon losing ground or even losing out entirely in the competitive race against Sony would be a real negative, IMO.


I honestly don't think Canon will be using a Sony-manufactured sensor. Even if it is a "Canon-designed" sensor, it just doesn't seem like either Canon nor Sony would work together as competitors like that. I think Canon is also WAAAY too prideful to step out and even try that...given the last interview with Masaya Maeda where he basically bold-faced lied about the state of their sensor tech.


On a side note, I don't consider Canon's use of Sony sensors in their compact cameras to be the same. Those are mass-market devices, and Canon has been using Sony CCDs in that arena for a very, very long time. It's the high end pro and semipro cameras that I think Canon takes a lot of pride in their "all in house, fully integrated" model.

 

[ahsanford]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

I honestly don't think Canon will be using a Sony-manufactured sensor. Even if it is a "Canon-designed" sensor, it just doesn't seem like either Canon nor Sony would work together as competitors like that. I think Canon is also WAAAY too prideful to step out and even try that...given the last interview with Masaya Maeda where he basically bold-faced lied about the state of their sensor tech.

+1

Pride pride pride. Top end gear will be entirely an in-house affair for Canon until they start losing pros in large enough numbers *expressly for sensor reasons* (DR, resolution, etc.).

I don't put stock in DXO's thoughts about Canon's sensors, but there is a fundamental argument that higher resolution and higher DR (at lower ISO) is a weak point for Canon's most demanding users. But the fact that the 5D3 is still sitting near its original asking price some 2.5 years later says that it's still a very desirable camera (or Canon is losing its shirt to make a point about protecting price).

But there must come a point where everything else that Canon does well -- that epic stable of glass, CPS, the ergonomics, reliability, access to massive third-party ecosystem of products, etc. -- could eventually be overpowered by a concern over Canon's sensors. Canon's sky is not falling by any means, but if I'm an executive at Canon, the "Threats of losing pros" on my SWOT list would be as follows:

[list type=decimal]
[*]The sensor.
[*]The sensor.
[*]The sensor.
[*]That silly mirrorless thing we're supposed to be working on.
[*]The sensor.
[/list]



- A

 

[unfocused]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

Top end gear will be entirely an in-house affair for Canon until they start losing pros in large enough numbers *expressly for sensor reasons* (DR, resolution, etc.).

Most pros could not care less about the slight dynamic range differences between sensors.

...But the fact that the 5D3 is still sitting near its original asking price some 2.5 years later says that it's still a very desirable camera (or Canon is losing its shirt to make a point about protecting price).

Which shows that high end enthusiasts and pros aren't concerned about this small issue. Obviously Canon is not losing its shirt. The 5DIII continues to do very well because it remains equal to or better than it's competitors in overall quality, regardless of how it may score on one minor data point that comes into play rarely if ever.

But there must come a point where everything else that Canon does well -- that epic stable of glass, CPS, the ergonomics, reliability, access to massive third-party ecosystem of products, etc. -- could eventually be overpowered by a concern over Canon's sensors.

If the concern were significant that would be true. But, keep in mind we are talking about tiny, tiny differences that have little to no impact in 99.9% of cases. Plus, it isn't even an overall weakness. It is again, just one small data point in sensor attributes.

Most reviewers are declaring the 7DII the best APS-C camera made and noting that no one else has a camera that can compete against it. Overall, the new sensor is as good or better than other APS-C sensors.

I personally hope Canon does not expend too many resources on making tiny improvements to sensors. There are much bigger threats out there (light field cameras, for example, represent a much bigger threat because they are truly disruptive technology).

 

[ahsanford]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

Top end gear will be entirely an in-house affair for Canon until they start losing pros in large enough numbers *expressly for sensor reasons* (DR, resolution, etc.).

Most pros could not care less about the slight dynamic range differences between sensors.

Absolutely, you are correct, but landscape shooters would love more low-ISO DR. It seems the D800/800E/810 + that epic 14-24 lens is a great landscape combination for Nikon in that regard. Don't get me wrong, I love the new 16-35 F/4L IS on my 5D3, but the majority of landscape stories I read / videos I see tout that Nikon combination as the one to beat.

I personally hope Canon does not expend too many resources on making tiny improvements to sensors. There are much bigger threats out there (light field cameras, for example, represent a much bigger threat because they are truly disruptive technology).

Agree that lightfield tech could become powerfully disruptive someday, but I think most people would argue that mirrorless will be far more disruptive in the mid-term.

- A

 

[jebrady03]

While DR limitations don't really bother me, low ISO noise does (and obviously the two are related). When I read the headline I thought to myself... "Canon gonna Canon". For those who don't know, that's a take off on "haters gonna hate" which is a way for people to say "it doesn't really matter what others [haters] say that's negative, because all they do is spew negativity".

It doesn't really matter what everyone says about Canon and their low ISO noise, Canon is going to keep doing what they're doing...

 

[vscd]

Most pros could not care less about the slight dynamic range differences between sensors.

That's right. I even would go further... they know, that the difference changes in higher ISO-Ranges. For sports- or Indoorphotography you usually don't use anything below ISO800. On sportevents because of the longer lenses and inside rooms because of the available light.

So let's check the differences of two nearly equally released and categoritzed cams:

So, the (quite huge) difference on ISO 100 flips to an advantage of Canon-Sensors from ISO 800 and upwards. No one tells you this, but it's the opposite of what the forums are telling on and on and on (Hi Drones!). I even wonder why DXO makes no intersection between lo and high of the (unpushed) ISO-Modes.

Canon is traditionally an autofocus/sports aranged brand, I think. So, from their side the sentence "our sensors are the best" could be true, really. For portraits you may use the Nikon instead--- and hell, a real pro doesn't care about brands. If you shoot portraits you can get far better results with a Pentax 645, a cam not much more expensive than a D4s.

 

[BLFPhoto]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?



given the last interview with Masaya Maeda where he basically bold-faced lied about the state of their sensor tech.

[/quote]

Attend many Canon executive meetings? Work in their manufacturing and engineering teams?

That's a highly inflammatory, bold statement for someone on the outside looking in. I deeply respect your knowledge of general sensor and IC technology but, respectfully, you don't have the position, information, or perspective to call someone inside Canon a liar. I see all sorts of statements like these by pundits whose tone comes across as if they have some inside knowledge calling out all of these companies for alleged motives and/or improprieties. It happens here about Canon, over on Nikonians about Nikon, and elsewhere about whatever company is on topic. I especially love the discussions where people make marionettes out of the engineers for the mythical Marketing monster to move around at will. That's a complete lack of perspective on how an entity like Canon, Nikon, or otherwise operate.

It is good enough for you to write intelligently about what is known about current products and how you feel about their level relative to what you believe is state-of-the-art. You do yourself and your considerable insight there a great discredit when you impugn the integrity and motives of those whom you do not know and have no reasonable knowledge of.

Can you stick to just discussing the aspects of sensor design and a discussion of their merits?

 

[jrista]

The notion that more DR is only for low ISO is now an old and archaic notion. The A7s changed the game again, as it brings more than two stops additional DR at very high ISO (1DX=6.6 sops, A7s=8.8 stops @ ISO 51200). The A7s has 8.8 stops of DR at ISO 51200! That is on top of having 12.9 stops of DR at ISO 100 (vs. the 1D X 11.2 stops).


Granted, the 1D X is a superior camera for it's intended use cases, which are primarily fast action shooting. However, I'd happily take a 1D XI that had 13 stops of DR at ISO 100, and even more happily take a 1D XI that had ~9 stops of DR at ISO 51200! That would be amazing. Not for shadow pushing, no...an additional two stops of DR at ultra high ISO makes it that much more usable, and printable at large sizes.


More dynamic range is more dynamic range. It allows for shadow pushing when you have more DR than a computer screen can render, but that is not it's only benefit. More DR means less noise, throughout the entire signal. More DR at high ISO is just as valuable, and arguably even more valuable, than more DR at low ISO because it means less noise globally. I would say that ISO 3200 is pretty usable on my 5D III pretty much all the time...however two more stops of DR would mean ISO 12800 becomes the new ISO 3200. I'd pay...a lot, to have that kind of high ISO performance in the 5D IV. It means cleaner, crisper photos with better color and sharper detail, with more flexibility to use motion-freezing shutter speeds in increasingly difficult lighting (such as the dim post-sunset light where ungulates and many other wildlife, hawks, etc. are often most active.)

 

[ahsanford]

Most pros could not care less about the slight dynamic range differences between sensors.

That's right. I even would go further... they know, that the difference changes in higher ISO-Ranges. For sports- or Indoorphotography you usually don't use anything below ISO800. On sportevents because of the longer lenses and inside rooms because of the available light.

So let's check the differences of two nearly equally released and categoritzed cams:

Sure. It's exactly those sort of charts that have people say portraiture and landscape does well with Nikon and events, photojournalism and sports are Canon's territory. But every type of shooter has both Canon and Nikon users and competent folks net great shots. Talent trumps the tech unless you are shooting yetis in darkness at ISO 2 Million.

- A

 

[Loswr]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

Most reviewers are declaring the 7DII the best APS-C camera made and noting that no one else has a camera that can compete against it.

Yeah, but that's only because the DxOMark Scores haven't come out yet. Once that happens, all those 7DII reviewers will realize how foolish they look for touting a camera that scores no better than the 70D, and nearly the same as the original (now so dated as to be decrepit) 7D. Heck, the D7100 gets a better Sports Score than the 7D(inosaur), I'm sure it'll beat the 7DII (no need to consider AF, fps or buffer for Sports, after all).

Reviewers will have egg on their face, and no one will buy the 7DII.

: : :

 

[jrista]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

given the last interview with Masaya Maeda where he basically bold-faced lied about the state of their sensor tech.

Attend many Canon executive meetings? Work in their manufacturing and engineering teams?

That's a highly inflammatory, bold statement for someone on the outside looking in. I deeply respect your knowledge of general sensor and IC technology but, respectfully, you don't have the position, information, or perspective to call someone inside Canon a liar. I see all sorts of statements like these by pundits whose tone comes across as if they have some inside knowledge calling out all of these companies for alleged motives and/or improprieties. It happens here about Canon, over on Nikonians about Nikon, and elsewhere about whatever company is on topic. I especially love the discussions where people make marionettes out of the engineers for the mythical Marketing monster to move around at will. That's a complete lack of perspective on how an entity like Canon, Nikon, or otherwise operate.

It is good enough for you to write intelligently about what is known about current products and how you feel about their level relative to what you believe is state-of-the-art. You do yourself and your considerable insight there a great discredit when you impugn the integrity and motives of those whom you do not know and have no reasonable knowledge of.

Can you stick to just discussing the aspects of sensor design and a discussion of their merits?

I'm sorry, but I'm far from the only person who was rather ticked off by Maeda's comments regarding their sensor technology. Even the interviewer was surprised by the response. Look at the comments on DPR's page for that interview...the very vast majority of the commentators were blown away by Maeda's comment about their sensor tech, the way he feigned surprise at the notion that there were any sensors on the market that performed better than Canon's.


Sorry, but I could really care less what you guys find "inflammatory" these days. Every word I write is inflammatory, so I am done caring about writing inflammatory things. You want to take offense at what I write, go right ahead, it doesn't matter to me. It doesn't take sitting on the Canon board to read between the lines of a set of interview responses like that. Canon is covering their ass, playing games, and pretending nothing is wrong (which is honestly not surprising, given they are a Japanese company...this is just a matter of "saving face"...however it really isn't saving them anything, as the vast majority of people who read that interview saw right through the facade, and a lot were quite unhappy with it all.)


I'm done having this community dictate to me what I can or cannot say. Be offended, if that's how you choose to respond to what I write. I no longer give a damn.

 

[Dylan777]

The notion that more DR is only for low ISO is now an old and archaic notion. The A7s changed the game again, as it brings more than two stops additional DR at very high ISO (1DX=6.6 sops, A7s=8.8 stops @ ISO 51200). The A7s has 8.8 stops of DR at ISO 51200! That is on top of having 12.9 stops of DR at ISO 100 (vs. the 1D X 11.2 stops).


Granted, the 1D X is a superior camera for it's intended use cases, which are primarily fast action shooting. However, I'd happily take a 1D XI that had 13 stops of DR at ISO 100, and even more happily take a 1D XI that had ~9 stops of DR at ISO 51200! That would be amazing. Not for shadow pushing, no...an additional two stops of DR at ultra high ISO makes it that much more usable, and printable at large sizes.


More dynamic range is more dynamic range. It allows for shadow pushing when you have more DR than a computer screen can render, but that is not it's only benefit. More DR means less noise, throughout the entire signal. More DR at high ISO is just as valuable, and arguably even more valuable, than more DR at low ISO because it means less noise globally. I would say that ISO 3200 is pretty usable on my 5D III pretty much all the time...however two more stops of DR would mean ISO 12800 becomes the new ISO 3200. I'd pay...a lot, to have that kind of high ISO performance in the 5D IV. It means cleaner, crisper photos with better color and sharper detail, with more flexibility to use motion-freezing shutter speeds in increasingly difficult lighting (such as the dim post-sunset light where ungulates and many other wildlife, hawks, etc. are often most active.)

I'm no expert in DR stuff. I did get a chance to play with A7s(rental) last Friday-Halloween. I was shooting with A7s + FE 55mm f1.8 and 1DX + 85L II. I set both cameras to @ f1.8 shutter speed around 200ish. Shooting condition was in school parking lot. At higher ISO, 25000ish, the a7s does have the upper hand. I'm not sure it has to do with Sony sensor or lower MP, compared to 1DX. All raw files were converted through DxO Prime same setting.

To me, most Pros would use flash(s) in shooting condition I was in.

 

[ahsanford]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

Most reviewers are declaring the 7DII the best APS-C camera made and noting that no one else has a camera that can compete against it.

Yeah, but that's only because the DxOMark Scores haven't come out yet. Once that happens, all those 7DII reviewers will realize how foolish they look for touting a camera that scores no better than the 70D, and nearly the same as the original (now so dated as to be decrepit) 7D. Heck, the D7100 gets a better Sports Score than the 7D(inosaur), I'm sure it'll beat the 7DII (no need to consider AF, fps or buffer for Sports, after all).

Reviewers will have egg on their face, and no one will buy the 7DII.

: : :

Neuro's tongue-in-cheek notwithstanding, I am curious to see the brouhaha that comes of DXO stating that Canon has (hypothetically) only improved one point over the 7D in their absurd rating system after 5 years of development.

There's already a thunderous din about how the 7D2 preliminarily appears to offer the same high ISO performance as the 70D, and only offers a stop better performance than the 7D (again: after five years). I'm still waiting for the Carnathans of the world to demonstrate this more thoroughly, but if substantiated, I can only imagine the ruckus that will ensue across the world's photo forums.

There's so much more to a camera than a sensor, but it is our favorite thing to obsess about, isn't it?

- A

 

[jebrady03]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

Most reviewers are declaring the 7DII the best APS-C camera made and noting that no one else has a camera that can compete against it.

Yeah, but that's only because the DxOMark Scores haven't come out yet. Once that happens, all those 7DII reviewers will realize how foolish they look for touting a camera that scores no better than the 70D, and nearly the same as the original (now so dated as to be decrepit) 7D. Heck, the D7100 gets a better Sports Score than the 7D(inosaur), I'm sure it'll beat the 7DII (no need to consider AF, fps or buffer for Sports, after all).

Reviewers will have egg on their face, and no one will buy the 7DII.

: : :

Neuro's tongue-in-cheek notwithstanding, I am curious to see the brouhaha that comes of DXO stating that Canon has (hypothetically) only improved one point over the 7D in their absurd rating system after 5 years of development.

There's already a thunderous din about how the 7D2 preliminarily appears to offer the same high ISO performance as the 70D, and only offers a stop better performance than the 7D (again: after five years). I'm still waiting for the Carnathans of the world to demonstrate this more thoroughly, but if substantiated, I can only imagine the ruckus that will ensue across the world's photo forums.

There's so much more to a camera than a sensor, but it is our favorite thing to obsess about, isn't it?

- A

I'm sure DxO is looking forward to it as well. MORE TRAFFIC!!!! (to their already abysmally slow website)

 

[jrista]

Re: Canon’s Upcoming Multi-Layer, Foveon-Like Sensor Made By Sony?

I'm sure DxO is looking forward to it as well. MORE TRAFFIC!!!! (to their already abysmally slow website)

It really is. Which is why I don't bother anymore...I just use Sensorgen.info (same core data, left up to us to interpret, without DXO's inane black box data massaging.)

 

[jrista]

I'm no expert in DR stuff. I did get a chance to play with A7s(rental) last Friday-Halloween. I was shooting with A7s + FE 55mm f1.8 and 1DX + 85L II. I set both cameras to @ f1.8 shutter speed around 200ish. Shooting condition was in school parking lot. At higher ISO, 25000ish, the a7s does have the upper hand. I'm not sure it has to do with Sony sensor or lower MP, compared to 1DX. All raw files were converted through DxO Prime same setting.

It would be the sensor. At that high of an ISO, it would be a combination of Q.E. and pixel size (although in the absence of pixel size, you can always average down in post and get similar, although possibly not quite as good, results.) The A7s also uses the BionzX 16-bit processing pipeline, which applies NR per pixel in the camera.


The big difference with the 1D X is the Q.E. and processing...Q.E. alone is quite important, 48% vs. 65%...that's a significant difference. If the 1D XI gets a DIGIC 6 or better yet a DIGIC 7 with improved NR capabilities, we might see an improvement in total DR, meaning the part of the signal not affected by read noise should improve. DIGIC 6 seems to have helped the 7D II a bit on that front...it may have the same read noise, but the rest of it's signal definitely seems improved over the original 7D.


Canon's read noise is probably always going to be the prime limiting factor on DR, so until they get their read noise issues resolved, I don't see Canon breaking 12 stops native DR, regardless of whether they jump to ~60% Q.E., or increase pixel size, or anything else.