Are we all clients of Sony ?

rattymouse wrote:
Thomas An wrote:

Note: It is true that some companies (like Fuji) might be bringing their patented blueprints to a Sony fab for manufacture (using sony only as a machinist), at which point it is not really a Sony sensor. But I still don't see the harm in coming forth with it when asked and say "Our sensors are designed in house but we have outsourced our fabrication to Sony for cost benefit".
Right, just like Apple does. Everyone knows that until recently, someone else besides Apple built their equipment. It doesnt matter because it is designed and spec'ed by Apple.

But watch what happens when someone asks Fujifilm who designed their sensors? It gets really ugly, real fast. This all started with the X100, which had a "customized" sensor.
 
rattymouse wrote:
Alphoid wrote:

Manufacturing requires an IC fab, which costs a few billion dollars. Not many companies have that kind of pocket, so it tends to be pretty centralized. Sony is one of the biggest sensor manufacturers, and we're unlikely to see many companies manufacture sensors. We do and will continue to see many companies design sensors, or modify standard designs.
We will also see many people claim that Company X designed a Sony sensor despite there not being a single piece of evidence to back up that claim.
 
MarkInSF wrote:
rattymouse wrote:
Alphoid wrote:

Manufacturing requires an IC fab, which costs a few billion dollars. Not many companies have that kind of pocket, so it tends to be pretty centralized. Sony is one of the biggest sensor manufacturers, and we're unlikely to see many companies manufacture sensors. We do and will continue to see many companies design sensors, or modify standard designs.
We will also see many people claim that Company X designed a Sony sensor despite there not being a single piece of evidence to back up that claim.
 
MarkInSF wrote:
rattymouse wrote:
Alphoid wrote:

Manufacturing requires an IC fab, which costs a few billion dollars. Not many companies have that kind of pocket, so it tends to be pretty centralized. Sony is one of the biggest sensor manufacturers, and we're unlikely to see many companies manufacture sensors. We do and will continue to see many companies design sensors, or modify standard designs.
We will also see many people claim that Company X designed a Sony sensor despite there not being a single piece of evidence to back up that claim.
 
rattymouse wrote:
MarkInSF wrote:
rattymouse wrote:
Alphoid wrote:

Manufacturing requires an IC fab, which costs a few billion dollars. Not many companies have that kind of pocket, so it tends to be pretty centralized. Sony is one of the biggest sensor manufacturers, and we're unlikely to see many companies manufacture sensors. We do and will continue to see many companies design sensors, or modify standard designs.
We will also see many people claim that Company X designed a Sony sensor despite there not being a single piece of evidence to back up that claim.
 
You are talking about trivial differences between sensors that amount to a little bit of nothing. My point is that some folks make claims that Fujifilm (for example) designed the sensor in the X100 and that it is very much substantially different from an off the shelf model, thus giving it very unique properties unable to be matched by said off the shelf model. Consider Fujifilm's SuperCCD SRII sensor vs an off the shelf Sony one. The SuperCCD sensor was designed and built in-house by Fujifilm, back when they owned and operated a fabrication facility. The architecture of that sensor was fundamentally different from any other sensor (besides the obvious CCD/CMOS difference).

I seriously doubt we will see such a large difference from "customized" sensors.
 
Bobn2 wrote:
MarkInSF wrote:
rattymouse wrote:
Alphoid wrote:

Manufacturing requires an IC fab, which costs a few billion dollars. Not many companies have that kind of pocket, so it tends to be pretty centralized. Sony is one of the biggest sensor manufacturers, and we're unlikely to see many companies manufacture sensors. We do and will continue to see many companies design sensors, or modify standard designs.
We will also see many people claim that Company X designed a Sony sensor despite there not being a single piece of evidence to back up that claim.
 
rattymouse wrote:

You are talking about trivial differences between sensors that amount to a little bit of nothing. My point is that some folks make claims that Fujifilm (for example) designed the sensor in the X100 and that it is very much substantially different from an off the shelf model, thus giving it very unique properties unable to be matched by said off the shelf model. Consider Fujifilm's SuperCCD SRII sensor vs an off the shelf Sony one. The SuperCCD sensor was designed and built in-house by Fujifilm, back when they owned and operated a fabrication facility. The architecture of that sensor was fundamentally different from any other sensor (besides the obvious CCD/CMOS difference).

I seriously doubt we will see such a large difference from "customized" sensors.
 
MarkInSF wrote:

I know that Nikon essentially designs its own sensors, though I was under the mistaken belief that there was a bit of Sony in their, too.
The truth is much stranger and more interesting. What there is in Nikon sensors is more than a bit of Mtsubishi. If you track the names on the patents, then you find that the core team came from Mitsubishi, and Renesas that does the fab is a successor to Mitsubishi Semiconductor. Now if you go back to the origins of canon sensors, and look at the papers they publish you find that Canon collaborated with Mitsubishi Semiconductor to get their sensor operation off the ground (Canon had no semiconductor design expertise at the time), and indeed one individual was in both teams. So the truth is that Nikon sensors have a bit of Canon in them.
 
MarkInSF wrote:
rattymouse wrote:

You are talking about trivial differences between sensors that amount to a little bit of nothing. My point is that some folks make claims that Fujifilm (for example) designed the sensor in the X100 and that it is very much substantially different from an off the shelf model, thus giving it very unique properties unable to be matched by said off the shelf model. Consider Fujifilm's SuperCCD SRII sensor vs an off the shelf Sony one. The SuperCCD sensor was designed and built in-house by Fujifilm, back when they owned and operated a fabrication facility. The architecture of that sensor was fundamentally different from any other sensor (besides the obvious CCD/CMOS difference).

I seriously doubt we will see such a large difference from "customized" sensors.
 
tecnoworld wrote:

Another one who talks without having even tried the nx300. It's much, much better than anything sony has in the mirrorless market. And better than any m4/3, since it has aps-c sensor.
I'll stick with my Sigmas. :)
 
Bobn2 wrote:
Nikon - Well, they're 16MP APS-C cameras and the D600 do. Do any others? I thought the new 24MP cameras used a toshiba sensor. No idea who makes the D4 or D800 sensor.
Nikon 1 - Aptina sensor

D3200 - Nikon sensor

D5200 - Toshiba sensor

D7100 - Toshiba sensor

D600 - Sony sensor

D800 - Sony sensor

D4 - Nikon sensor

Nikon sensors are truly Nikon sensors, fabrication of the wafers is outsourced to foundries, but that is also true of some of Sony's production, part of Canon's process and all of Foveon, Aptina's and Omnivision's production.
 
Bobn2 wrote:
MarkInSF wrote:
rattymouse wrote:

You are talking about trivial differences between sensors that amount to a little bit of nothing. My point is that some folks make claims that Fujifilm (for example) designed the sensor in the X100 and that it is very much substantially different from an off the shelf model, thus giving it very unique properties unable to be matched by said off the shelf model. Consider Fujifilm's SuperCCD SRII sensor vs an off the shelf Sony one. The SuperCCD sensor was designed and built in-house by Fujifilm, back when they owned and operated a fabrication facility. The architecture of that sensor was fundamentally different from any other sensor (besides the obvious CCD/CMOS difference).

I seriously doubt we will see such a large difference from "customized" sensors.
 
Bobn2 wrote:
MarkInSF wrote:

I know that Nikon essentially designs its own sensors, though I was under the mistaken belief that there was a bit of Sony in their, too.
The truth is much stranger and more interesting. What there is in Nikon sensors is more than a bit of Mtsubishi. If you track the names on the patents, then you find that the core team came from Mitsubishi, and Renesas that does the fab is a successor to Mitsubishi Semiconductor. Now if you go back to the origins of canon sensors, and look at the papers they publish you find that Canon collaborated with Mitsubishi Semiconductor to get their sensor operation off the ground (Canon had no semiconductor design expertise at the time), and indeed one individual was in both teams. So the truth is that Nikon sensors have a bit of Canon in them.

--
Bob
Oy. I bet they don't like that. Doesn't surprise me, though. My partner used to work for a company spun off from NEC in an attempt to commercialize some work done at their research lab in Princeton. NEC, like most big companies, especially the Japanese giants, had a lot of cool stuff coming out of their labs but the companies moved so slowly they were making no money off them. Some nimble startup, usually American, always got established first. So they decided to give the IP toban American startup andclet thembrun with it. Lots of agreements were made and key people hired and granted stock options (it was about 15 years ago). Then NEC discovered they would have to pay a small fee for every product they made using this technology. That was a necessary part of the agreement to spin off this company. The problem was that several NEC competitors were going to get to use the IP for free due to cross-licensing agreements. Paying for 'their' technology and 'giving' it to Sony and Panasonic was too much, especially since they had decided to push this as an essential piece of a major standard. So they sneakily pulled the plug on the spin-off, in violation of the agreement, all the shareholders sued and got very nice settlements. The technology wasn't adopted as part of the standard and never made a penny for anyone except the core employees/shareholders who sued and settled (including my partner). I got a free trip to Tokyo (business class) out of NEC. It was all very revealing of how slow and inbred Japanese companies can be. Tokyo was lovely.
 
rusticus wrote:

No - there is also an Aptina sensor: awesome, will represent a revolutionary development in the future V3 milestone!
It is indeed awesome and I hope it makes it into the V3. The only question is whether the iq will actually improve. The speed and video capabilities will be superb. I suspect they're aiming it more at the video world, but have my fingers crossed for improved stills iq. I figure to be getting a new camera about the time the V3 is on clearance (maybe two years).
 
Mahmoud Mousef wrote:

Patent-wise ...we probably are all clients of Sony, along with many other manufacturers. Paying for tech that isn't developed in-house is something all manufacturers do, even if they manufacture their own sensors.
Sometimes they don't even have to pay for it, if there are cross-licensing agreements.
 
MarkInSF wrote:
Bobn2 wrote:
MarkInSF wrote:
rattymouse wrote:

You are talking about trivial differences between sensors that amount to a little bit of nothing. My point is that some folks make claims that Fujifilm (for example) designed the sensor in the X100 and that it is very much substantially different from an off the shelf model, thus giving it very unique properties unable to be matched by said off the shelf model. Consider Fujifilm's SuperCCD SRII sensor vs an off the shelf Sony one. The SuperCCD sensor was designed and built in-house by Fujifilm, back when they owned and operated a fabrication facility. The architecture of that sensor was fundamentally different from any other sensor (besides the obvious CCD/CMOS difference).

I seriously doubt we will see such a large difference from "customized" sensors.
 
MarkInSF wrote:
Bobn2 wrote:
Nikon - Well, they're 16MP APS-C cameras and the D600 do. Do any others? I thought the new 24MP cameras used a toshiba sensor. No idea who makes the D4 or D800 sensor.
Nikon 1 - Aptina sensor

D3200 - Nikon sensor

D5200 - Toshiba sensor

D7100 - Toshiba sensor

D600 - Sony sensor

D800 - Sony sensor

D4 - Nikon sensor

Nikon sensors are truly Nikon sensors, fabrication of the wafers is outsourced to foundries, but that is also true of some of Sony's production, part of Canon's process and all of Foveon, Aptina's and Omnivision's production.

--
Bob
Just for curiosity, since you apparently know about the ins and outs, do you have any take on why Nikon isn't designing and having made their recent cameras? I know the Sony and Toshiba sensors are excellent, but Nikon has a lot of experience at this, too, and must have a substantial patent portfolio. Or do licensing agreements mean Sony and Toshiba have been able to use Nikon-owned IP in their designs? It would seem a bit risky dropping out of designing sensors completely, but that seems to be happening (the D3200 not being the newest design.)

I have a V1 and did a bit of reading on the Aptina sensor when I bought it. Nikon seems to have been heavily involved in the design of those sensors, though using Aptina technology fundamentally. Which made sense for an unusual new sensor from a company that hadn't made anything so ambitious. I wish its dynamic range was better, but in general the sensor is quite good. It certainly enabled some fancy features still unequaled by others in the still camera world.
I don't have inside information, but I have a take on it. I think Nikon takes competitive tenders on each new camera, with their in-house sensor operation being one of the bidders. Whoever gives them the best solution gets that design win. It's a good strategy, keeps all their suppliers honest, including their in-house operation, which doesn't have the opportunity to get lazy. This is their second go at having in-house sensor capability after the LBCAST debacle - and I think they learned a lot from that. The point is that their sensor team doesn't cost so much, it's is a few individuals with the capacity for maybe a couple of designs a year but its presence is very important to their overall sensor supply strategy - it means they are beholden to no other company but still use the best of what other companies offer, and if they have a requirement no-one else can fill, they can do it themselves. Compare it with Canon's strategy, where the development of their whole DSLR line is limited by the capability of their fab and because they have tied so much capital up in that fab, they can't easily go outside.

As for Nikon being involved in the Aptina design, I'm not sure they were too much - it mostly looks like tech present on other Aptina products. Of course, where Nikon engineers will get involved in every design is process development - they are the suppliers of the photolithography equipment for most of the sensor suppliers, and they also supply much of the process related software that goes with that - as the suppliers develop their processes, Nikon support engineers will be called in to advise - which allows Nikon to say with some truth that they were engaged in the development process. Where the product is for Nikon, one can imagine that support might be part of the whole deal.

--
Bob
 
Last edited:
Sigmas are very good, at least at low iso. They are particular and probably not comparable with anything on the market. But if you consider nex, m4/3 or even fuji (the latter at least for af performance and overall speed) then nx300 is better.
 
Bobn2 wrote:
MarkInSF wrote:
Bobn2 wrote:
MarkInSF wrote:
rattymouse wrote:

You are talking about trivial differences between sensors that amount to a little bit of nothing. My point is that some folks make claims that Fujifilm (for example) designed the sensor in the X100 and that it is very much substantially different from an off the shelf model, thus giving it very unique properties unable to be matched by said off the shelf model. Consider Fujifilm's SuperCCD SRII sensor vs an off the shelf Sony one. The SuperCCD sensor was designed and built in-house by Fujifilm, back when they owned and operated a fabrication facility. The architecture of that sensor was fundamentally different from any other sensor (besides the obvious CCD/CMOS difference).

I seriously doubt we will see such a large difference from "customized" sensors.
 

Keyboard shortcuts

Back
Top