Who make D3x/s's sensors ?

Brad Morris wrote:

It's unclear why Nikon in their launch of EXPEED decided to claim it was a "group" of things. When you disassemble a D3, for instance, the only item labeled EXPEED is the imaging ASIC.

I suspect that Nikon's use of EXPEED was an attempt to create an umbrella name for a "look" not a specific set of parts. Shortly after the D3 they announced Coolpix models with "EXPEED." I haven't disassembled one yet, but based upon what I know about Coolpix design, I'd be a little surprised if the same ASIC were in the Coolpix. I think that EXPEED is actually being used to describe "cameras that use Picture Controls" (as opposed to the former Image Optimization).

--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
I don't think it is even as complicated as that. I'm sure Nikon engineering team don't call any of the ASICs EXPEED and that ASICs are individually tailored to each product (though with as much commonality as possible to reduce development costs). EXPEED is just a fancy name invented by the arty types in marketing because every one else in the electronics industry has got into the habit of giving code names to custom microelectronics.
--
Dave
http://www.rosser.pwp.blueyonder.co.uk
http://www.pbase.com/dgrosser
 
I have already pointed you towards the evidence. Taking a look for yourself is obviously too hard for you it is easier to call me a liar. Facts are pesky when they get in the way of a good story aren't they?
Yes they are, but unfortunately for you, it's not my facts that are pesky.
Copied from the D700 FW V1.01 update opened in Notepad:

UNI-S Encode V1.00 COPYRIGHT (C) NEC CORPORATION 2003,2005 06911?
Note the copyright dates. This code was not written for the D700. As I suggested in my earlier post, it is library code supplied by NEC to drive one of their chips (or ASIC IPR) included in the camera. If you'd hadd any involvement with ASIC development, you'd know that this is how it works. Typically, the IPR blocks come with driver code to speed up development. It's also not unknown for IPR, in the form of complete subsystems to be included in a chip from another vendor. Digital CMOS is pretty generic. There arre companies that do business purely supplying such block of IPR.
And further down the page in the same file

}Pp îÌZ ®¶HI1000/4 V.1.05.00.000 Copyright (C) 2003(2007) Renesas Technology Corp. and Renesas Solutions Corp. All rights reserved. \
All the same comments apply to this. I know that the D3 and d700 have several Renesas chips in them.
BTW the D700 FW update consist of only a single file, not 2 files as in all other Nikon DSLRs that I have seen.
and...?
In any case, none of that tells you who actually fabricates the chip.
And cut and pasted from the D3 FW 2.01 "B" file

à6Softune REALOS/FR is Realtime OS for FR Family, based on micro-ITRON COPYRIGHT(C) FUJITSU LIMITED 1994-1999 Ÿ? JxŸ?h '?0? ƒß ? ???Œÿ?ƒß?0Ÿ?àÿÀ Ÿ?h Ÿ?h
Good, so we know that Fujitsu supplied the RTOS (but not necessarily the chip, although in actual fact, they did)
The "A" file is copyright Nikon. I believe, but i am not sure, that the A file has traditionally been for things like menu format.

It is not me who has taken over the thread with "authoritative" assumptions with no basis in fact -
I post what I find interesting and hope others will find interesting. I think, if you look, my speculation is clearly such.
[Nikon couldn't be using a different chip cause they just signed a JV with Fujitsu]
Which are you saying has no basis in fact? That Nikon has formed a JV with Fujitsu? Because they have - http://www.nikon.com/about/news/2009/0630_01.htm .

Now note from this press release: 'The company will be solely responsible for the firmware development in Nikon digital cameras.' Now, how does this fit with your 'fact' that the firmware in the D700 was developed by NEC and Renesas?
The different types of firmware code is easy evidence.
But not evidence supporting your claim.
Where is yours when you make assumptions about Nikon Business practices?.
Keeps on coming up and biting you in the bum, I'm afraid.
BTW it's 'Renesas' and 'copyright'.
Yes you are right. my keyboard cant spell very well. hehe
The D700 is the first Nikon I have seen that does not use Fujitsu sourced Asic to process images and run the camera
There is every probability the the D700 expeed is the same as the D3, D3x and D300. I don't believe what you say, one bit.
Even Nikon have publicly announced that EXPEED is not a single brand/model/type of chip, why do you suspect Nikon of Lying? Expeed is a "digital processing concept" to quote Nikon, that combines different components to get the best image quality, not a particular chip.
OK. There is every reason to believe that the image acquisition and circuitrt on the D700 is identical to that in the D3. There would be no point in Nikon making it different, they would be paying out additional development costs for no advantage - disadvantage, in fact, since they would be splitting volume across two parts. So, I have no hard evidence which says conclusively that it is the same image processing chip in the D700, but you have o evidence that it is not, and to make it different would be a very strange decision on Nikon's part.
Not sure what you think there is to be learned from that. I could take the use of the term 'DSLR series EXPEED' and the image of one chip with 'EXPEED' written on it to substantiate a statement that the EXPEED in DSLR's is a chip, and it is the same in all DSLR's (I wouldn't actually say that, because I know how slippery marketing speak can be)
 
I have already pointed you towards the evidence.
Yes, but you don't seem to understand the evidence you presented. The claims of Copyright are for modules embedded in the firmware, not the firmware itself. Your first clue is the version numbers (1.01, 1.00, 1.05, etc.). Your second clue is in the names of what is being claimed (firmware 1.01, Uni-S Encode 1.00, etc.). It is not at all unusual for there to be lots of pooled IP in products.

The simple explanation for the difference is that the updates for the different cameras had different embedded IP in them, not that different companies created them. In other words, the firmware for camera X updated things that involved embedded IP X, and the firmware for camera Y updated things that involved embedded IP Y.
Thom,
Mac OSX and Windows can both run on Intel Processors. The same software will not run on top of the different Operating systems. Can you make a program do calculations on both? Yes! 2+2=4 on both system. They are different though.

Firmware for these cameras is effectively the operating system with additional applications on top. A firmware update will replace the existing OS and applications in its entirety whilst retaining the stored data, if any.

Why would a company expend resources developing a completely new branch of FW unless it was essential to release the product for technical requirements?

Differing firmware, by different manufacturers may well be image processing algorithm compatible, but by definition, is not the same base firmware. D70, D200, D300, D3 are all using Fujitsu based firmware - (RealOS) for their cameras processing requirements. The D700 is not. Do D70, D300 and D3 use exactly the same processor? My guess is most likely not, they use new iterations of a family of chips that use the same command set much like Intel 486 uses the same commands as the current Intel i7 chip. are they the same? No way! the I7 is much more practical in today's Operating system world. These cameras are released years apart and of course there is improvement in performance with each iteration of chip but Nikon is still using the same family of processors in each Model, except the D700. Why? I have no idea what their motivation was to do that.

D700 deviates from that path by using a completely different OS sourced from NEC/Renasas, not fujitsu. The distributed firmware update even comes in a different format altogether from the Fujitsu based cameras leading to more proof that they are different processors that are behind the sensor.
Even Nikon have publicly announced that EXPEED is not a single brand/model/type of chip, why do you suspect Nikon of Lying?
It's unclear why Nikon in their launch of EXPEED decided to claim it was a "group" of things. When you disassemble a D3, for instance, the only item labeled EXPEED is the imaging ASIC.

I suspect that Nikon's use of EXPEED was an attempt to create an umbrella name for a "look" not a specific set of parts. Shortly after the D3 they announced Coolpix models with "EXPEED." I haven't disassembled one yet, but based upon what I know about Coolpix design, I'd be a little surprised if the same ASIC were in the Coolpix. I think that EXPEED is actually being used to describe "cameras that use Picture Controls" (as opposed to the former Image Optimization).
That is exactly my point. The term "Expeed" describes a concept and not a physical chip. Nikon have left themselves open to use what ever chips they like as long as they can output images with consistent Picture controls.

This discussion is possibly educational but at the end of the day, it really makes no difference what sensor and CPU each model cameras have, as long as they reliably output Nikon quality NEF and JPG files to the installed storage. Nikon have been doing that with DSLRs and even coolpix models since before the term Expeed was even coined.
--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
--
http://www.flickr.com/photos/k-blad/
 
[snip]
It is not me who has taken over the thread with "authoritative" assumptions with no basis in fact -
I post what I find interesting and hope others will find interesting. I think, if you look, my speculation is clearly such.
[Nikon couldn't be using a different chip cause they just signed a JV with Fujitsu]
Which are you saying has no basis in fact? That Nikon has formed a JV with Fujitsu? Because they have - http://www.nikon.com/about/news/2009/0630_01.htm .

Now note from this press release: 'The company will be solely responsible for the firmware development in Nikon digital cameras.' Now, how does this fit with your 'fact' that the firmware in the D700 was developed by NEC and Renesas?
D700 release predates 2009 does it not? In fact it was announced 1 July 2008. Your are basing an assumption on a press release that came out a year after the D700 was announced. I have never said that Nikon will drop Fujitsu and continue down the D700 path. I have no idea where they will be heading with their internal electronics. I would imagine that they will move back to Fujitsu the fold for future model releases due to the JV agreement. You are assuming cause you saw the announcement it applies top all Nikon cameras both past and future
The different types of firmware code is easy evidence.
But not evidence supporting your claim.
Where is yours when you make assumptions about Nikon Business practices?.
Keeps on coming up and biting you in the bum, I'm afraid.
BTW it's 'Renesas' and 'copyright'.
Yes you are right. my keyboard cant spell very well. hehe
The D700 is the first Nikon I have seen that does not use Fujitsu sourced Asic to process images and run the camera
There is every probability the the D700 expeed is the same as the D3, D3x and D300. I don't believe what you say, one bit.
Even Nikon have publicly announced that EXPEED is not a single brand/model/type of chip, why do you suspect Nikon of Lying? Expeed is a "digital processing concept" to quote Nikon, that combines different components to get the best image quality, not a particular chip.
OK. There is every reason to believe that the image acquisition and circuitrt on the D700 is identical to that in the D3. There would be no point in Nikon making it different, they would be paying out additional development costs for no advantage - disadvantage, in fact, since they would be splitting volume across two parts. So, I have no hard evidence which says conclusively that it is the same image processing chip in the D700, but you have o evidence that it is not, and to make it different would be a very strange decision on Nikon's part.
Why? cause they both produce NEF and JPG files? That is a product of the application software and algorithms that sit on top of the operating system that talks to the CPU. I can produce an Excel spreadsheet or PDF file on a PPC Mac and a Windows box too. Are the CPU, OS and application software identical. Of course they are not but they can write identical files.

I agree that if they used the same hardware family base it doesn't make sense to source different or rewrite the OS and embedded application software...yet they did....I wonder why? Maybe cause it was essential to drive differing hardware???

BTW it is "circuit"
Not sure what you think there is to be learned from that. I could take the use of the term 'DSLR series EXPEED' and the image of one chip with 'EXPEED' written on it to substantiate a statement that the EXPEED in DSLR's is a chip, and it is the same in all DSLR's (I wouldn't actually say that, because I know how slippery marketing speak can be)
That is exactly what the term Expeed is...marketing speak to check a box and keep up with the Canons and the Sonys. The hardware has improved along the way and they have tweaked the way the image settings are presented by changing the software to now call them "Picture Controls" but the process chain is still pretty much the same conceptually as it was in the D1 and D100 days

--
http://www.flickr.com/photos/k-blad/
 
D70, D200, D300, D3 are all using Fujitsu based firmware - (RealOS) for their cameras processing requirements. The D700 is not.
No, you're still not getting it. You're examining firmware updates , not the full firmware listings. Let's say the full firmware contains IP from A, B, and C. You're comparing a firmware update that updates sections that contain A and B IP with one that updates sections with C IP. Yes, in that case I'd expect to see different Copyright notices in the firmware update.
D700 deviates from that path by using a completely different OS sourced from NEC/Renasas, not fujitsu.
Sorry, you've provided no evidence of that. I haven't disassembled my D700 lately, but I'm pretty sure its running the same family of processor as everything else.
That is exactly my point. The term "Expeed" describes a concept and not a physical chip.
Yet on every Nikon DSLR that's been produced since then, there is exactly one chip on the digital boards clearly labeled "EXPEED." I think you're reading too much into Nikon's marketing, and you're doing it from translations, to boot.
Nikon have left themselves open to use what ever chips they like as long as they can output images with consistent Picture controls.
And the advantage to that would be?

Capture NX2 has at its base the same software engine that's in the ASIC in hardware. Indeed, you may remember that the Nik folk were told that they can't change it, and that it is maintained by Nikon corporate.

--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
 
D700 release predates 2009 does it not?
You keep trying to come up with arguments to support your theory (note that I said theory, not "facts"). The problem is that pixelpeeper is indeed correct. You're seeing bits and pieces of the IP that's in the firmware, but you're not seeing all the firmware. Thus, your conclusions are incorrect. The Nikon bodies use REALOS/FR and have for some time. The use the same family of Fujitsu CPUs (but different ones for different models). Nikon uses additional intellectual property from NEC, Renesas, Fujitsu, NuCore, and quite a few others. Some of that is referenced in the firmware (kind of hard to drive a hardware part like, say, a USB chip, without doing so). I'm sure if we could reverse engineer the EXPEED ASIC (yes, the one marked EXPEED) we'd find more.
OK. There is every reason to believe that the image acquisition and circuitrt on the D700 is identical to that in the D3. There would be no point in Nikon making it different, they would be paying out additional development costs for no advantage - disadvantage, in fact, since they would be splitting volume across two parts. So, I have no hard evidence which says conclusively that it is the same image processing chip in the D700, but you have o evidence that it is not, and to make it different would be a very strange decision on Nikon's part.
Why? cause they both produce NEF and JPG files?
Actually: because they produce identical NEF files under the same test conditions, and the parts on the D700's digital board appear identical to the ones on the D3's. What pixelpeeper is saying is that it would be wasted development money to produce exactly the same output using exactly the same parts by having someone else rewrite everything from scratch. Didn't happen. Would be mind-bogglingly insane if it did.
That is a product of the application software and algorithms that sit on top of the operating system that talks to the CPU.
I can see that you've never worked with a real-time OS before. Think more at the driver level and you'd be closer. But still wrong.
I can produce an Excel spreadsheet or PDF file on a PPC Mac and a Windows box too. Are the CPU, OS and application software identical. Of course they are not but they can write identical files.
Using your logic, Microsoft hired Apple to write the Mac version. If you don't think there is considerable shared code between the Mac and Windows versions of Excel, you'd be wrong. And the reason why it isn't 100% shared code is that it can't be: the engine underneath the application is different. Now your contention is that the engine underneath of the D700 is different than the engine of the D3. You are wrong. Disassemble the two and look at the parts. Moreover, you're being pigheadingly stubborn that your "evidence" points to something different than it actually does.
I agree that if they used the same hardware family base it doesn't make sense to source different or rewrite the OS and embedded application software...yet they did....I wonder why?
"Yet they did." That's presented as a statement of fact. What you really mean is "from my brief sleuthing in the firmware update code I think they did." That's a theory. The problem with your theory is that the actual facts are in contradiction to it.
That is exactly what the term Expeed is...marketing speak to check a box and keep up with the Canons and the Sonys.
Actually, I was hammering on Nikon for some time prior to the D3 introduction about having an imaging ASIC (ala Digic) but not managing to bring that to consumers attention via marketing. That they've broadened the marketing presentation to mean "everything we do in creating an image" actually brings further criticism from me. EXPEED as a term is vague and meaningless. If I were in Nikon's product marketing I'd be writing a line like "The EXPEED chip at the heart of every Nikon DSLR has some of the most sophisticated image rendering capabilities available, including things like..." and then I'd go into the source/benefit list, ala "Distortion elimination: removes all linear distortions from even the best of lenses, resulting in straight lines that are straight." But the way Nikon's gone about using the term, well, they might not have bothered.
The hardware has improved along the way and they have tweaked the way the image settings are presented by changing the software to now call them "Picture Controls" but the process chain is still pretty much the same conceptually as it was in the D1 and D100 days
Yep. And those cameras had an imaging ASIC chip marked "Nikon," while the new ones have one marked "EXPEED."

--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
 
It is not me who has taken over the thread with "authoritative" assumptions with no basis in fact -
I post what I find interesting and hope others will find interesting. I think, if you look, my speculation is clearly such.
[Nikon couldn't be using a different chip cause they just signed a JV with Fujitsu]
Which are you saying has no basis in fact? That Nikon has formed a JV with Fujitsu? Because they have - http://www.nikon.com/about/news/2009/0630_01.htm .

Now note from this press release: 'The company will be solely responsible for the firmware development in Nikon digital cameras.' Now, how does this fit with your 'fact' that the firmware in the D700 was developed by NEC and Renesas?
D700 release predates 2009 does it not? In fact it was announced 1 July 2008. Your are basing an assumption on a press release that came out a year after the D700 was announced. I have never said that Nikon will drop Fujitsu and continue down the D700 path. I have no idea where they will be heading with their internal electronics. I would imagine that they will move back to Fujitsu the fold for future model releases due to the JV agreement. You are assuming cause you saw the announcement it applies top all Nikon cameras both past and future
That is a safe assumption. The story you now want us to believe is
  • Prior to the D700, Nikon sourced their image processor from Fujitsu, and co-developed the firmware with them
  • After the D700, Nikon will source their image processor from Fujitsu, and develop the firmware using the JV.
  • For the D700 only the processor and firmware were differently sourced.
Thom has already explained to you why your evidence doesn't stand up. Now, you're trying to use this flaky evidence to support what is, on first sight, a completely ludicrous proposition.
The different types of firmware code is easy evidence.
But not evidence supporting your claim.
Where is yours when you make assumptions about Nikon Business practices?.
Keeps on coming up and biting you in the bum, I'm afraid.
BTW it's 'Renesas' and 'copyright'.
Yes you are right. my keyboard cant spell very well. hehe
The D700 is the first Nikon I have seen that does not use Fujitsu sourced Asic to process images and run the camera
There is every probability the the D700 expeed is the same as the D3, D3x and D300. I don't believe what you say, one bit.
Even Nikon have publicly announced that EXPEED is not a single brand/model/type of chip, why do you suspect Nikon of Lying? Expeed is a "digital processing concept" to quote Nikon, that combines different components to get the best image quality, not a particular chip.
OK. There is every reason to believe that the image acquisition and circuitrt on the D700 is identical to that in the D3. There would be no point in Nikon making it different, they would be paying out additional development costs for no advantage - disadvantage, in fact, since they would be splitting volume across two parts. So, I have no hard evidence which says conclusively that it is the same image processing chip in the D700, but you have o evidence that it is not, and to make it different would be a very strange decision on Nikon's part.
Why? cause they both produce NEF and JPG files? That is a product of the application software and algorithms that sit on top of the operating system that talks to the CPU. I can produce an Excel spreadsheet or PDF file on a PPC Mac and a Windows box too. Are the CPU, OS and application software identical. Of course they are not but they can write identical files.
No, because they do precisely the same job, from precisely the same sensor. To invest money in reworking something they already had would be absurd.
I agree that if they used the same hardware family base it doesn't make sense to source different or rewrite the OS and embedded application software...yet they did....I wonder why? Maybe cause it was essential to drive differing hardware???
The hardware is not differing , it is identical. Again, you have to ask, why would they spend good money devloping differing hardware when they had proven hardware that did the job?
That is exactly what the term Expeed is...marketing speak to check a box and keep up with the Canons and the Sonys. The hardware has improved along the way and they have tweaked the way the image settings are presented by changing the software to now call them "Picture Controls" but the process chain is still pretty much the same conceptually as it was in the D1 and D100 days
What Nikon brand it as makes not difference to the fact that all the Nikon D3's, D700's and D300's that people have looked inside have the same image processing chip, labelled 'Expeed'.
 
D70, D200, D300, D3 are all using Fujitsu based firmware - (RealOS) for their cameras processing requirements. The D700 is not.
No, you're still not getting it. You're examining firmware updates , not the full firmware listings. Let's say the full firmware contains IP from A, B, and C. You're comparing a firmware update that updates sections that contain A and B IP with one that updates sections with C IP. Yes, in that case I'd expect to see different Copyright notices in the firmware update.
Thom,

You are not getting it. Flashing firmware deletes the existing FW from camera flash memory and replaces it with a whole new image contained in the FW update file. That is why you can brick your camera or iPhone if you turn it off mid FW update. While it is true that some of the new Bin file/s potentially contains some code that is identical to the preceeding fw version, it still replaces the lot as an image like the block level disk images you can make with Norton Ghost on a PC. It does not do a file by file replacement like your PC does with a software update.
D700 deviates from that path by using a completely different OS sourced from NEC/Renasas, not fujitsu.
Sorry, you've provided no evidence of that. I haven't disassembled my D700 lately, but I'm pretty sure its running the same family of processor as everything else.
Yes i have provided evidence. A Firmware update is an image files that replaces the existing firmware on the camera flash memory. Every time you install new firmware, you are installing a new copy of the cameras operating system. D700 compared to all other Nikon DSLRs I have seen are intrinsically different.
That is exactly my point. The term "Expeed" describes a concept and not a physical chip.
Yet on every Nikon DSLR that's been produced since then, there is exactly one chip on the digital boards clearly labeled "EXPEED." I think you're reading too much into Nikon's marketing, and you're doing it from translations, to boot.
All new P&S cams from Nikon also have "Expeed" CPUs as well. Do you really think that a slow P&S uses the same processor as a D3 or D3x? If they did, we should be able to expect DSLR like write times clearing the buffer to storage, multitasking for continuous shooting etc. No Nikon P&S that i know of has particularly strong credentials in that area.
Nikon have left themselves open to use what ever chips they like as long as they can output images with consistent Picture controls.
And the advantage to that would be?
An example i can think of is D3 to D3s and D300 to D300s. Now, for better or worse, can process video that the previous versions couldn't. If the guts were the same, they could indeed provide video functionality with a FW update but they have not done that, cause i think, the older models are missing new hardware that assists with video encoding.

Using a catch all term like Expeed also means that they can update their hardware and use best of breed for new models while still benefiting from the previous investment on marketing setting up the Expeed brand. That happens all the time in tech gear. Many home network Routers use Linux with open source firewall software and wrap it up in a brand. During teh life of teh product the name does not change but the software and CPU inside changes cause new tech comes along that is better of cheaper to produce. Dlink is a master of that type of marketing cause as long as the router does its routing job and it is relatively painless to use, nobody cares how it does it. It is an appliance.

The only manufacturer that is doing processor version numbering so far is canon and they are talking about a particular chip, not a processing concept. No-one, not even Nikon, would put marketing spin on something to muddy the waters if there is a simple direct answer like "we use a CPU that is called Expeed". They never talk in those terms. They do talk in terms of we use LBCAST, CMOS or CCD sensors then get mysterious or vague on where the sensor comes from, that is, with the exception of LBCAST that they shouted from the rooftops they produced it.
Capture NX2 has at its base the same software engine that's in the ASIC in hardware. Indeed, you may remember that the Nik folk were told that they can't change it, and that it is maintained by Nikon corporate.
That is true and it makes my point exactly. Mac NX2 does not need Microsoft .Net as the framework to run as the Windows version does. The Mac version has a completely different codebase to run on OSX. The calculations and algorithms the application uses are defined by Nikon, That in simple terms, sit on top of a code foundation that talks to the operating system. That code foundation is unique to each OS version and provides the appropriate system calls to run the Nikon defined algorithms in a manner the OS can deal with

The operating system protects the overall system by preventing the NX2 software from directly accessing the hardware. That has been the case with windows since the release of NT3.51. In the Mac world that has been true since the beginning, it is also the reason why Macs can run Rosetta and have the non Universal binaries running on a PPC cpu or an Intel CPU.
--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
--
http://www.flickr.com/photos/k-blad/
 
You aren't understanding Thom's claim.

I would bet that EXPEED uses a common API across all platforms. Different ASICS might be used in different cameras, and different features might be present, but I'd bet anything that there is a common interface.

And the RT OS is going to be largely the same, just as the Windows API remains mostly constant across updates. Different drivers from different manufacturers will also be updated. But the whole thing rewritten from scratch? Creating a mature and stable firmware package isn't cheap or simple. The backbone will be retained wherever possible. To do otherwise would be suicidal and impossibly expensive.
 
You are not getting it. Flashing firmware deletes the existing FW from camera flash memory and replaces it with a whole new image contained in the FW update file.
Nikon cameras now have as many as four different "firmware" modules that can updated independently. Sorry, but I'll stick with my contention.
All new P&S cams from Nikon also have "Expeed" CPUs as well. Do you really think that a slow P&S uses the same processor as a D3 or D3x?
Never said it did. And do you think a Canon G10 has the same DIGIC as a 1DIV? What these imaging ASICs provide is an API. I'd guess that the API is pretty much the same across the products.
If they did, we should be able to expect DSLR like write times clearing the buffer to storage, multitasking for continuous shooting etc.
If the ASIC were the only thing dealing with those things, sure. But unfortunately, it's not.
No Nikon P&S that i know of has particularly strong credentials in that area.
Nope. But that's not because of the imaging ASIC.
An example i can think of is D3 to D3s and D300 to D300s. Now, for better or worse, can process video that the previous versions couldn't.
Nikon's video so far is Motion-JPEG. That means that the imaging ASIC is still outputting JPEGs, but now something is stringing them together into a movie file. But I think the simple explanation is that the D300s EXPEED chip (yes, still one chip labeled EXPEED) is a newer variant. It's not unusual to take your old ASIC and add things to it for the next generation of products. Given that what I'm seeing on a D300s is exactly the same output as from a D300 under the same test condition, I'd guess that all the still shooting aspects are exactly the same in the two generations of chip.
If the guts were the same, they could indeed provide video functionality with a FW update but they have not done that, cause i think, the older models are missing new hardware that assists with video encoding.
Actually, you're missing a key aspect of using an ASIC. I'm fairly certain that Nikon could create firmware that strings JPEGs together into a Motion-JPEG. You may remember the D100 (I do ;~). It used the camera's CPU to do compression of NEFs. That's why NEF buffer clearing was so slow when you set compression: the camera's CPU was being taxed. In the D70, Nikon moved the software compression routine into the ASIC, where it became optimized hardware. Poof. Fast compression. Same thing would be true with stitching Motion-JPEG: slow if done in firmware, fast in the ASIC. So if you're iterating the ASIC, you put it in there. People wonder why it might take two years to add video to a D300: well one answer is that developing, changing, and testing the ASIC isn't exactly a "you'll have it next week" process.
Using a catch all term like Expeed also means that they can update their hardware and use best of breed for new models while still benefiting from the previous investment on marketing setting up the Expeed brand.
Right, so Canon is wrong. They specifically point to their ASIC as DIGIC. Across all models. It's unclear to me why Nikon would want to widen the claim as to what EXPEED is and then provide no substance for that.
The only manufacturer that is doing processor version numbering so far is canon and they are talking about a particular chip, not a processing concept.
What the heck is a "processing concept"?
No-one, not even Nikon, would put marketing spin on something to muddy the waters if there is a simple direct answer like "we use a CPU that is called Expeed".
Ah, you don't know Nikon like I know Nikon ;~). They're very good at fuzzy claims in marketing. Unfortunately, fuzzy claims aren't good marketing.
That is true and it makes my point exactly. Mac NX2 does not need Microsoft .Net as the framework to run as the Windows version does. The Mac version has a completely different codebase to run on OSX. The calculations and algorithms the application uses are defined by Nikon, That in simple terms, sit on top of a code foundation that talks to the operating system. That code foundation is unique to each OS version
You're mixing application foundations with library code. Yes, the Mac and Windows versions use different application foundations, sourced from a third party. What I'm referring to is a very specific walled off library that does the raw conversion. If you write an application on either platform in any application foundation, you can do the same thing Capture NX2 does by licensing the SDK. That provides you with that special module that is callable by your application. That module is the core that is also coded in hardware in the EXPEED chip in the camera. That module does not "talk to the operating system." It talks to your application.

--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
 
Brad,

I was preparing a long reply, showing how much you have to learn about firmware updates, real time software construction and the like Then I found this, which settles the matter simply and easily.

Here is a picture of the D700 image processor chip, (labelled 'EXPEED by Nikon'), taken from Nikon's press resources:



Here is a picture of the D3 main PCB, including the image processor chip (labelled EXPEED by Nikon)



Notice that the part identification numbers on the chip are identical .

Here is an image from the D700 flyer showing the EXPEED labelled chip on the D700 main circuit card



(a bit blurry here, but anyone with a real copy can confirm that the markings are identical)

Incidentally, we can see how things move on - here's the D3s PCB, it has a new version of the EXPEED labelled chip



As does the d300s



I would assume the upgrade has something to do with video.
 
I was preparing a long reply, showing how much you have to learn about firmware updates, real time software construction and the like Then I found this, which settles the matter simply and easily.
And if I'm correct, E1-142 is the part identifier, YYWW is the date of manufacturer, and Z## is the actual iteration of the part (the D3, D300, and D700 used Z01, the D300s uses Z02, and the D3s use Z03).

--
Thom Hogan
author, Complete Guides to Nikon bodies (21 and counting)
http://www.bythom.com
 
I was preparing a long reply, showing how much you have to learn about firmware updates, real time software construction and the like Then I found this, which settles the matter simply and easily.
And if I'm correct, E1-142 is the part identifier, YYWW is the date of manufacturer, and Z## is the actual iteration of the part (the D3, D300, and D700 used Z01, the D300s uses Z02, and the D3s use Z03).
More complex yet. Using the images from the Nikon press center:
D300 was Z00



D3 and D700 were Z01



D3x was Z40 ! (I suspect the D3x has a lot of clever firmware!)



D90 was Z00, just like the D300



D300s was Z02



D3s was Z03

 
The sensors:
D3



D3x



D3s (note subtle differences from D3)



D300 (note family similarity to D3x)



D90 (note different packaging from D300)



D300s



And just for a reference, A900



(I know I've posted these before, but these versions are all direct from the manufacturers, so less arguable)
 
The Canon 5D Mk II sensor



And just to repeat, for comparison, the D3 sensor



Draw your own conclusions.
 
Thank you for proving my point. There is no such thing as a single Expeed chip that is used in each of the cameras as you have been arguing. Nikon selects appropriate CPUs for each model. Your photos show different Part numbers for different CPUs.

D3
E1-142 0708 Z01
D3s
E1-142 0852 Z03
D300s
E1-142 0819 Z02

I cannot see your D700 image, the only image I could find from Nikon USA on D700 expeed is a 12 page PDF brochure and that image is so pixelated, it could be anything and i cant read any detail.

You even concede that the differing model number is to support the additional Video functionality.

Silkscreen labels mean nothing, other than it is falling into a range of chips that Nikon has chosen to install in various camera models that they market with the expeed processing concept. The chip producers could silkscreen "Canon Suks" in the similar font on the chips if Nikon wanted them to. The Black square plastic that is labeled is a UV proof protective layer that protects the actual silicon wafer.

Model numbering for Nikon branded EXPEED CPUs says more about Nikon's inventory management to ensure the right CPU with the right specs is installed in the right camera body design than is says about the chips manufacturer which , by the way is nothing.

The only thing that you can say from the various images is that the top of the IC is of similar shape and possibly size. There is no pin out detail and no specifics to exactly what is on the silicon wafer inside.

--
http://www.flickr.com/photos/k-blad/
 
Thank you for proving my point. There is no such thing as a single Expeed chip that is used in each of the cameras as you have been arguing.
It does not prove your point at all. Your point , if you follow this thread, was that the D700, uniquely, has a image processor made by NEC rather than Fujitsu, something for which you have supplied zero evidence, so far as I can see. The evidence to the contrary is:
  1. The image of the chip labelled 'Expeed' given in Nikon's publicity for the D700 is iteration Z01, just like the D3.
  2. The firmware uploads you refer to for the D700 contain the RTOS copyright strings referring to Fujitsu and the FR processor series, indicating that this firmware runs on a Fujitsu FR series processor.
  3. The copyright strings for 'NEC' and 'Renesas' which are you sole evidence, also appear in D3 and D300 firmware updates.
So, your point is one of the most disproven there is. Talk about beating a dead horse.
 

Keyboard shortcuts

Back
Top