Should/can Canon keep making its own sensors?

[marekjoz]

I've already done it. Works fine for me.

Maybe you got lucky and selected a clean 7D, they do exist.
Run of the mill ones aren't so good.

But also, lets' not see a high contrast B&W shot which can reduce the effect of (primarily) red channel noise.  I use B&W to hide a noisy shot too. 

Lucky you. So what's your answer to the question in topic?

My answer is to buy what works and my advice to Canon is to develop better sensor tech on their own or license it from the competition. And start supplying it to us FAST!
(...)

Many of us shoot in SUNLIGHT.  Harsh, contrasty and uncontrollable for all but intimate settings.  Under these circumstances, the extremes of shadow and light are captured, compressed and presented as a final image far better, IMO, by sensors in recent Nikon cameras than by sensors in any Canon DSLR.

As such, Canon has an option to listen to the complaints of consumers of their products.  Should they try to satisfy these demands for better low ISO performance, their engineering branch is going to have to work very hard to catch up to the competition in this aspect.

Call or email Canon and let them know your opinion on this if it matters enough.
I have.

So the best option for you would be a camera with ISO 50-800. It could be a niche market anyway.

[archangelrichard]

EPIC FAIL

Let's start with the logic: "I wonder if Canon will reach a kind of "Apple Moment", like when Apple quit spending its R&D money trying to beat Intel on a component (CPU's) and started just using Intel chips like everyone else."

Apple has NEVER made the CPU, The Apple I, II, III used Rockwell 6502's as the base, the Mac's used standard 68000 family (68008, 68016 et al) until the Power series used IBM power PC chips (which have so much power that they were illegal to export - remember those commercials with a mac surrounded by tanks because it was too powerful to export>); the switch to Intel chips lobotomized the Mac's to an extent (there are several things you just can't do with an Intel chip because of it's architecture so Apple has to ignore / block those instructions when using the chip in apple os modes

The OP has NO idea what they are talking about so the logic is an epic fail

The question, rather, is how important is the sensor to the camera - is the camera (now that we left film for digital) no more than an extension of the sensor? At that point the sensor becomes key to product development and - yes - it is best (if you can afford it) to keep it in-house

Going back to computers; IBM set a standard in 1981 with the PC - but by 1985 IBM was number 5 in the PC-compatibles market; the problem was that the only thing IBM actually made on the PC was the l;abel; everything else was stardard parts that anyone else could piece together (some better than others) and most cheaper than IBM; Apple had many non-standard parts (for example Steve Wozniak had developed a way to go from 20 chips to run a disk drive as found on standard S-100 computers to 4 chips, one being called the "Integrated Woz Machine" chip, or proprietary and better architecture).

This is what sensor design is to cameras - you can do much better when using your own technologies if they are better technologies; make your own "woz" devices and Canon has, which is why Nikon had to outosurce - they just can'y afford to keep up with Canon on this.

Here is where your analogy falls apart - IBM went to they PS-2 because they lost their market and were trying to get it back not by making better 6technology but by making proprietary technology. Back in the mainframe days IBM had figured out a way to make money by taking their standards and creating paper :Institutes" as they wrote off the costs of development a second time by "charitable donation". The names of these "institutes" are ANSI (American National Standards Institute) and ASCII (American Standards Conference for Information Interchange). A Company called Control Data joined these and began making IBM-Compatible terminals at a lower price undercutting IBM. IBM sued and lost - once CD had joined these tax dodges they had tjhe right to build terminals with these standards -- so IBM set a NEW standard that was thoroughly incompatible (3270 and EBCDIC) which are NOT better hut different - and a large part of the market migrated, when they needed to upgrade, to smaller but compatible manufacturers - Digital Equipment (DEC PDP computers which were not originally called computers in fear of IBM, Control Data moved into computers, HP, and many others

This is where Nikon is - to a certain extent they can't be the Canon alternative without simply being a different lens mount, they need to offer something significantly different -- but they haven't the money for it. The issue with Nikon is that Canon is 60% of the market, Nikon is only 25% so a much smaller company; sony is in many more markets, essentially dominating video, so sensor technology is necessary to their operations; and (here is where outsourcing works) it is a much more trivial matter to incorporate some of the concept that they developed for video into still camera sensors (for their own (minolta - remember) use as well as for Nikon's with Nikon supplying some of their expertise on their versions but manufacturing exclusively by sony)

In other words the economics of the situation is that Nikon simply can not afford to make it's own sensors AND keep pace with Canon; this is a Nikon issue and has nothing to do with sensor technology - and that is the OP's epic fail, not understanding the problem. You don't understand the computer industry demonstrating a fanboy mentality and then carry that misunderstanding to Cameras like a bad program gone wild Nikon's adopting a more generic way of getting sensors will be their death - it becomes no more than a (industry standard) sensor in a non-canon lens mount; Nikon is now in a death spiral they may not come out of, necessitated by financials; they simply can not compete on the same turf as Canon. Look for Sony to put money into Nikon, then buy the rest of the company and make Nikon the high end of the Sony cameras

While Canon keeps on in their 60% of the market

[RLPhoto]

EPIC FAIL

Let's start with the logic: "I wonder if Canon will reach a kind of "Apple Moment", like when Apple quit spending its R&D money trying to beat Intel on a component (CPU's) and started just using Intel chips like everyone else."

Apple has NEVER made the CPU, The Apple I, II, III used Rockwell 6502's as the base, the Mac's used standard 68000 family (68008, 68016 et al) until the Power series used IBM power PC chips (which have so much power that they were illegal to export - remember those commercials with a mac surrounded by tanks because it was too powerful to export>); the switch to Intel chips lobotomized the Mac's to an extent (there are several things you just can't do with an Intel chip because of it's architecture so Apple has to ignore / block those instructions when using the chip in apple os modes

The OP has NO idea what they are talking about so the logic is an epic fail

The question, rather, is how important is the sensor to the camera - is the camera (now that we left film for digital) no more than an extension of the sensor? At that point the sensor becomes key to product development and - yes - it is best (if you can afford it) to keep it in-house

Going back to computers; IBM set a standard in 1981 with the PC - but by 1985 IBM was number 5 in the PC-compatibles market; the problem was that the only thing IBM actually made on the PC was the l;abel; everything else was stardard parts that anyone else could piece together (some better than others) and most cheaper than IBM; Apple had many non-standard parts (for example Steve Wozniak had developed a way to go from 20 chips to run a disk drive as found on standard S-100 computers to 4 chips, one being called the "Integrated Woz Machine" chip, or proprietary and better architecture).

This is what sensor design is to cameras - you can do much better when using your own technologies if they are better technologies; make your own "woz" devices and Canon has, which is why Nikon had to outosurce - they just can'y afford to keep up with Canon on this.

Here is where your analogy falls apart - IBM went to they PS-2 because they lost their market and were trying to get it back not by making better 6technology but by making proprietary technology. Back in the mainframe days IBM had figured out a way to make money by taking their standards and creating paper :Institutes" as they wrote off the costs of development a second time by "charitable donation". The names of these "institutes" are ANSI (American National Standards Institute) and ASCII (American Standards Conference for Information Interchange). A Company called Control Data joined these and began making IBM-Compatible terminals at a lower price undercutting IBM. IBM sued and lost - once CD had joined these tax dodges they had tjhe right to build terminals with these standards -- so IBM set a NEW standard that was thoroughly incompatible (3270 and EBCDIC) which are NOT better hut different - and a large part of the market migrated, when they needed to upgrade, to smaller but compatible manufacturers - Digital Equipment (DEC PDP computers which were not originally called computers in fear of IBM, Control Data moved into computers, HP, and many others

This is where Nikon is - to a certain extent they can't be the Canon alternative without simply being a different lens mount, they need to offer something significantly different -- but they haven't the money for it. The issue with Nikon is that Canon is 60% of the market, Nikon is only 25% so a much smaller company; sony is in many more markets, essentially dominating video, so sensor technology is necessary to their operations; and (here is where outsourcing works) it is a much more trivial matter to incorporate some of the concept that they developed for video into still camera sensors (for their own (minolta - remember) use as well as for Nikon's with Nikon supplying some of their expertise on their versions but manufacturing exclusively by sony)

In other words the economics of the situation is that Nikon simply can not afford to make it's own sensors AND keep pace with Canon; this is a Nikon issue and has nothing to do with sensor technology - and that is the OP's epic fail, not understanding the problem. You don't understand the computer industry demonstrating a fanboy mentality and then carry that misunderstanding to Cameras like a bad program gone wild Nikon's adopting a more generic way of getting sensors will be their death - it becomes no more than a (industry standard) sensor in a non-canon lens mount; Nikon is now in a death spiral they may not come out of, necessitated by financials; they simply can not compete on the same turf as Canon. Look for Sony to put money into Nikon, then buy the rest of the company and make Nikon the high end of the Sony cameras

While Canon keeps on in their 60% of the market

Yeah! what this guy said.

[marekjoz]

EPIC FAIL

Let's start with the logic: "I wonder if Canon will reach a kind of "Apple Moment", like when Apple quit spending its R&D money trying to beat Intel on a component (CPU's) and started just using Intel chips like everyone else."
(...)

Eric - I think i understand what OP tried to say - let them focus on what they do best like Apple did. You would admit, that Apple was involved with IBM and Motorola in designing CPUs. Apple didn't design it from scratch, nor did it alone, but was involved and some of their resources were spent on it. Later they decided to adapt to "ready to use" Intel processors. I think that was what OP wanted to say - maybe also Canon could abandon it's sensors production as there are others who might do it better. One of the questions is whether this would be smart, as from strategic as well from operational points of view. As far this is one of many differentiating factors between Canon and others. The other question is if Canon could make a better camera using ie Sony sensors than Nikon itself. The other question is what in a longer time horizon would it mean for Canon itself: it's future market share and what their publicity would be after this.
True is, that Apple focused on setting a product vision, using fully outsourced hardware assembly and delivering "apple-centric" software solution. The advantage there is not a CPU itself, but let's say Retina in example. Apple doesn't care how the specification set by them, would be achieved in production - they just order it to be designed, manufactured and delivered for use in a final product. Putting this idea on Canon DSLRs it would be a camera visioned by Canon, with specification set by Canon. If we say that sensor is essential here, then Canon sets the specification of the sensor, orders it and prepares for use in their final product. And I would not be so sure, that this is not possible. There are a lot of electronic circuits in a DSLR and I could agree, that it's build around the sensor, but I'm not sure it's not possible to achieve a better product the same way as Apple successfuly did. There is of course an assumption, that Canon sensors are bad, what a lot of people don't agree. There is also another assumption, that there is a company in the world, that Canon would trust as a sensor deliver.
There are similarities but also differences in what the Apple's product is in comparison to Canon's product. Here we focused on Apple PCs and Canon DSLRs. But there are more hardware differences between DSLRs of different manufacturers, as it is in PCs world. Maybe one of the reasons Canon holds it's own sensors production is that they see it rather as an advantage and not a problem.
Many assumptions here, many "ifs" and "whens" but I don't see OP's idea as an epic fail, as you did.

[briansquibb]

I wonder what you did back in the 2001-2002 era when these sensors and cameras were unheard of.  Photography I suppose was just all crap?

They moaned that the D60 was not much of an improvement over the D30 and that the 1Ds was too expensive

[bdunbar79]

I wonder what you did back in the 2001-2002 era when these sensors and cameras were unheard of.  Photography I suppose was just all crap?

They moaned that the D60 was not much of an improvement over the D30 and that the 1Ds was too expensive

And that the 1V was crap because it wasn't digital?   

[bdunbar79]

I would just like to say, that regarding the posts of D800 vs. 5D Mark III in this thread, I printed 3 posters out tonight on my Canon printer with images taken with a 5D Mark III and they looked absolutely fabulous, and sharp down to the last detail.  What would 36mp gain me?  That's all  .

[wickidwombat]

I wonder what you did back in the 2001-2002 era when these sensors and cameras were unheard of.  Photography I suppose was just all crap?

They moaned that the D60 was not much of an improvement over the D30 and that the 1Ds was too expensive

And that the 1V was crap because it wasn't digital?   

and the 1Dmk1 used a 4MP kodak sensor and had too much shadow pattern noise

[Woody]

First you need to get your facts straight.
1. Canon uses sensors from Sony, and they make sensors as well

Canon buys Sony CCDs for their compact cameras, but all Canon DSLRs use Canon designed and manufactured CMOS sensors.

I too believe competition is good. But if they cannot match the performance of Sony sensors, they may as well buy Sony CMOS sensors until they are able to come up with more competitive sensor designs. The poor low ISO dynamic range of Canon sensors is hurting their reputation.

[jrista]

EPIC FAIL

Let's start with the logic: "I wonder if Canon will reach a kind of "Apple Moment", like when Apple quit spending its R&D money trying to beat Intel on a component (CPU's) and started just using Intel chips like everyone else."

Apple has NEVER made the CPU, The Apple I, II, III used Rockwell 6502's as the base, the Mac's used standard 68000 family (68008, 68016 et al) until the Power series used IBM power PC chips (which have so much power that they were illegal to export - remember those commercials with a mac surrounded by tanks because it was too powerful to export>); the switch to Intel chips lobotomized the Mac's to an extent (there are several things you just can't do with an Intel chip because of it's architecture so Apple has to ignore / block those instructions when using the chip in apple os modes

The OP has NO idea what they are talking about so the logic is an epic fail

The question, rather, is how important is the sensor to the camera - is the camera (now that we left film for digital) no more than an extension of the sensor? At that point the sensor becomes key to product development and - yes - it is best (if you can afford it) to keep it in-house

Going back to computers; IBM set a standard in 1981 with the PC - but by 1985 IBM was number 5 in the PC-compatibles market; the problem was that the only thing IBM actually made on the PC was the l;abel; everything else was stardard parts that anyone else could piece together (some better than others) and most cheaper than IBM; Apple had many non-standard parts (for example Steve Wozniak had developed a way to go from 20 chips to run a disk drive as found on standard S-100 computers to 4 chips, one being called the "Integrated Woz Machine" chip, or proprietary and better architecture).

This is what sensor design is to cameras - you can do much better when using your own technologies if they are better technologies; make your own "woz" devices and Canon has, which is why Nikon had to outosurce - they just can'y afford to keep up with Canon on this.

Here is where your analogy falls apart - IBM went to they PS-2 because they lost their market and were trying to get it back not by making better 6technology but by making proprietary technology. Back in the mainframe days IBM had figured out a way to make money by taking their standards and creating paper :Institutes" as they wrote off the costs of development a second time by "charitable donation". The names of these "institutes" are ANSI (American National Standards Institute) and ASCII (American Standards Conference for Information Interchange). A Company called Control Data joined these and began making IBM-Compatible terminals at a lower price undercutting IBM. IBM sued and lost - once CD had joined these tax dodges they had tjhe right to build terminals with these standards -- so IBM set a NEW standard that was thoroughly incompatible (3270 and EBCDIC) which are NOT better hut different - and a large part of the market migrated, when they needed to upgrade, to smaller but compatible manufacturers - Digital Equipment (DEC PDP computers which were not originally called computers in fear of IBM, Control Data moved into computers, HP, and many others

This is where Nikon is - to a certain extent they can't be the Canon alternative without simply being a different lens mount, they need to offer something significantly different -- but they haven't the money for it. The issue with Nikon is that Canon is 60% of the market, Nikon is only 25% so a much smaller company; sony is in many more markets, essentially dominating video, so sensor technology is necessary to their operations; and (here is where outsourcing works) it is a much more trivial matter to incorporate some of the concept that they developed for video into still camera sensors (for their own (minolta - remember) use as well as for Nikon's with Nikon supplying some of their expertise on their versions but manufacturing exclusively by sony)

In other words the economics of the situation is that Nikon simply can not afford to make it's own sensors AND keep pace with Canon; this is a Nikon issue and has nothing to do with sensor technology - and that is the OP's epic fail, not understanding the problem. You don't understand the computer industry demonstrating a fanboy mentality and then carry that misunderstanding to Cameras like a bad program gone wild Nikon's adopting a more generic way of getting sensors will be their death - it becomes no more than a (industry standard) sensor in a non-canon lens mount; Nikon is now in a death spiral they may not come out of, necessitated by financials; they simply can not compete on the same turf as Canon. Look for Sony to put money into Nikon, then buy the rest of the company and make Nikon the high end of the Sony cameras

While Canon keeps on in their 60% of the market

+1 Awesome.

Not sure I entirely agree with the last paragraph...I don't think Nikon is in a "death spiral"...they have had a smaller market share for some time and seem to fit and survive quite well within that "niche" (its not really a niche...the market for DSLR's is monstrous, so 25% is an ungodly amount of "market" on a per-customer basis). I would say that this is Nikon's bid to expand their market share beyond 25% rather than simply sit happy and live with it. If they fail, well they remain at that comfortable but less than ideal 25%. They certainly seem to be succeeding from a marketing and minds standpoint as well...if I were to put some off-the-cuff ratios out there regarding overall customer sentiment and satisfaction (regardless of their current adherence to any given brand), I'd say 60% were more than happy to ecstatic about what Nikon is producing these days, and excited about what the future holds for new Nikon gear. On the flip side, I would say closer to 25% feel the same about what Canon is producing these days, or excited about what the future holds for new Canon gear. I could see the future becoming more balanced...35%+ market share for Nikon, and closer to 50% for Canon, and possibly a long-term shift in market share if Canon can't find ways to compete with Nikon on a level footing, and sustain that competition.

I'm a Canon user myself, but I'm NOT particularly excited about what Canon's been producing these days as far as DSLR tech goes, or what seems to be over the horizon. I AM actually intrigued and even excited about what Nikon is and has been doing. They are innovating, and even on the backs of Sony, innovation is particularly exciting for consumers, because it increases competition. Canon has sustained a mediocre level of competition with their latest cameras, however on all critical hardware aspects, including image sensor, AF sensor, etc. Canon is innovating less and "cheating" more, just to get by. They know the state they are in, they know they have increasingly inferior technology upon which to build their products. They know that bugs a growing percentage of their customers (however a small enough percentage that they haven't actually done anything about it yet.)

A bigger, better, more competitive Nikon, again even if it is on the backs of Sony, is a really good thing for this market. A useful shift in market share, even if it ultimately ends up being to a Sony-owned Nikon brand a few years down the road (still not sure thats going to happen), is probably the only thing that will really light the fire under the collective Canon arse and force them to REALLY compete in the areas where Nikon is currently excelling. Canon turned the tides back in the 80's with autofocus, and ushered in a new DSLR era. Perhaps now Nikon is turning the tides again with advanced image and f/8 AF sensor design, and will usher in the next DSLR era.

[Aglet]

So the best option for you would be a camera with ISO 50-800. It could be a niche market anyway.

yes, a real ISO 50 to 800 would be just ducky!

Not so niche a market tho, mainstream product in the last year has MET my requirements in this area. I merely lament it's not by the mfr that I've supported almost exclusively with my purchases for the last 30-some years.

Whether Canon can catch up or surpass with their own in-house sensor tech, or whether they make like recent Apple and package leading edge tech from other component mfrs, wrapped in their own overall design and interface, does not matter to me.

I just want a camera that's as easy to use as Canon's have always been to me, and also supply the kind of image quality I've been craving for years and have only just now experienced by using competitors' products.

I'd rather sell my Nik gear and use Can equipment than the other way around, i still feel some brand loyalty.  But it's fading fast; I'm getting too old to be patient enough to wait.  I only wish I'd have payed more attention back when the D90 came out... The first DSLR I know of with a significant improvement in dark noise pattern reduction.

[LetTheRightLensIn]

Going back to computers; IBM set a standard in 1981 with the PC -

a very poor simple-minded standard

they didn't believe in the PC and dumped it off to some little team they didn't care about much and yeah they got junk

Apple had many non-standard parts (for example Steve Wozniak had developed a way to go from 20 chips to run a disk drive as found on standard S-100 computers to 4 chips, one being called the "Integrated Woz Machine" chip, or proprietary and better architecture).

You really need to look, in that era, back to Atari and CBM if you want to talk about tons of proprietary custom chips.  People like Jay Miner, not Woz, were the real hardware wizards in the home computing realm back then.

While Canon keeps on in their 60% of the market

not necessarily if their sensors remain back of the pack for low ISO and they end up charging 30% more for everything

[LetTheRightLensIn]

I wonder what you did back in the 2001-2002 era when these sensors and cameras were unheard of.  Photography I suppose was just all crap?

They moaned that the D60 was not much of an improvement over the D30 and that the 1Ds was too expensive

Most of shot film is what we did.

[LetTheRightLensIn]

First you need to get your facts straight.
1. Canon uses sensors from Sony, and they make sensors as well

Canon buys Sony CCDs for their compact cameras, but all Canon DSLRs use Canon designed and manufactured CMOS sensors.

I too believe competition is good. But if they cannot match the performance of Sony sensors, they may as well buy Sony CMOS sensors until they are able to come up with more competitive sensor designs. The poor low ISO dynamic range of Canon sensors is hurting their reputation.

The only tricky thing is that they are then potential stuck at the whims of whether sony wants to go low or high mp and with whether movie modes and liveview will work well or not.  I'd rather see if they couldn't pay sony to be able to use exmor patents but stick with designing stuff themselves otherwise (although even here they do seem a bit behind since the non-exmor D4 has a lot more DR than 1DX/5D3, but exmor patents would give them enough). I don't know that sony would particularly be willing to sell canon rights to use exmor patents or maybe canon would not want to pay (eventually market may force them?). And perhaps Nikon has some deal where sony can't let anyone else use exmor??

[NormanBates]

Canon's business plan, as of late, seems to be: add a "I", charge $1000 more, then just do as if it's actually worth that much

but more seriously, check this:
http://www.lensrentals.com/blog/2012/07/is-sony-going-to-be-the-digital-kodak