Will Thom's prediction come true? Locked

Started Jun 20, 2013 | Discussions thread
This thread is locked.
bobn2 Forum Pro • Posts: 59,075
Re: Lots of nonsense here

Laurence Matson wrote:

Bobn2 wrote:

Laurence Matson wrote:

Bobn2 wrote:

I doubt whether three layer technologies will ever get any traction.


It is an opinion, so whether it is 'nonsense' or not can only be borne out in the future. Even though you tell me it is 'nonsense', it remains my opinion. To convince me otherwise you would have to show me a three layer technology that has any potential to produce a better price/performance ratio than Bayer. To date I have seen none, including Foveon (and the market has already cast its verdict on Foveon).

Traction the Foveon technology already has. How good the traction is, is another question.

It's difficult to see where that traction is, with the tiny market share it has. Traction would suggest that companies with sensor supply problems, of which there have been several, would come to Foveon. They don't. Olympus went to Sony. Leica went to CMOSIS. Panasonic is dealing with Fujifilm. Nikon went to Toshiba and Aptina.

It is certainly better than CFA technology was 15 years after its emergence versus the then king of the hill: film.

Can't make much sense of that. You suggest that Foveon technology is better with respect to Bayer than Bayer was with respect to film 15 years after Bayer technology's emergence? Well the Bayer CFA was dependent on the development of silicon that could use it, just as was the Foveon system, and they ride on the same silicon process improvements. The point about that is that there is not a means whereby the Foveon technology will suddenly overtake the bayer technology - the relative performance is where they are locked. That wasn't the case with respect to film or digital - film was a mature technology developing slowly, semiconductor image sensors were a young technology developing fast and overtook the film very quickly.

For most purposes, Bayer is a perfectly acceptable design solution and given the accumulated investment that it gathers, it simply will continue to offer better price/performance than any emergent three layer technoloy.

To some extent, true

The extent being about 100%.

Nonsense. If this were absolutely true, then why are at least three of the major players now working on similar technology? If you think it is because they are completely daft, I might agree with you, but for other reasons.

How much real work do you think the major players are doing on their alternative technologies? Sure, they make patents. They are profligate with patents - they take a creative engineer a couple of days to think up and the scattergun approach to patenting them they hoe will ensure that no competitor gets a march on them, but none see the light of day. Mostly they have hugely difficult manufacturing problems. Like the Nikon one with nano scale dichroic mirrors embedded in the toppings - no-one has successfully manufactured anything like that. Or the panasonic one with three layers of different III-V alloys, no-one has made a chip remotely like that (and now they seem keen enough on the alternative single layer graphene to have actually made a press release). There is no real evidence of anyone other than Foveon working on any three layer tech. Why, well Mead and Merrill, both very clever men, came up with about the only workable option, and it still isn't good enough to give a better price/performance point than Bayer's (also a very clever man) idea - which is a lot ore subtle than people give credit for.

The problem is this: to succeed in the market and three layer has to offer clearly better pice/performance than Bayer at the first shot. That is what Foveon failed to do, which is why it didn't succeed.

This is not true at all, re. success.

You think Foveon succeeded? No major design wins. Think of that. Nikon, Olympus, Pentax, Ricoh, Konica Minolta all played the commodity sensor market, none bought Foveon. Of all the phone manufacturers, none chose Foveon. Canon, Nikon, Sony, Panasonic, Toshiba etc all had the resources to buy Foveon lock stock and barrel had they wanted, but none did. The only partnership the could find, the only route to survival was Sigma (a company I admire) with less than 1% of the market. I'm afraid by any commercial standard that is failure.

You have no way of knowing this, but you are wrong.


I will only go so far as to say that one of the major companies you list above was actually the first "adopter." Then cut-throat business politics got in the way. But one of them actually were going to leave the commodity market and launch a flagship dSLR. Nuff said.

But then they didn't. What a surprise. Large corporations have executives nosing about after all sorts of stuff. It only counts when the people with the sign-off commit. They obviously didn't. Spin it how you like. I nearly made a billion. But I didn't. 'If only' doesn't buy anything.

Foveon has one huge advantage over any other proposed three layer technology, which is that it is basically compatible with existing CMOS fabrication processes.

Also not really true.

What are you saying, that Foveon is not compatible with existing CMOS processes (it is, it's fabricated on Dongbu's CMOS image sensor line) or that other three layer proposals are? If it's the latter that you are saying, which one is compatible?

You got the former manufacturer correct. There have been some shifts there, and likely will be more in the face of new challenges. To the extent that it is a CMOS fab and technology, yes. But the equipment is very different.

Which equipment? Actually you undercut your own argument - if Foveon have easily moved away from Dongbu (which is quite likely because the smaller pixel pitch probably requires a smaller process) then it says that the process is more compatible with standard CMOS, not less. This is a plus point for Foveon, not a negative - as the people arguing against it seem to have inferred. Anyway, trying to find who the fab is, I came across this little article on the SD1 sensor in Image Sensor World - here's a picture for anyone interested:

But more interesting are the comments below:

Eric R Fossum May 26, 2012 at 3:32 PM

Albert, I see what you mean by 3-way shared 4T but I guess that the transfer gates act like switches (potential is same on both sides) and there is not complete charge transfer and thus kTC noise. If the the transfer gates are just barriers, then they might have lag plus kTC noise. I guess we won't know until someone publishes the noise specs.

None of this addresses one of the other primary problems with silicon-self-filtered color -- the large amount of color processing and poor YSNR10.

Both still seem a large price to pay for eliminating some color aliasing, esp. considering 1.1 um pixels below the diffraction limit (anti-aliasing by physics), BSI QE, and sub 2- read noise.

yang ni - Eric, why do you think there is no complete charge tranfer please ?

Eric - It is a long way down to those deeper wells and you would need a long neck that would be fully depleted at high potential to reach down there and do complete charge transfer out. Not impossible, just unlikely at these dimensions and operating voltages.

None of the other proposals is, so that they will require enormous R&D investment to get them anywhere close to commercial release.

And not just because of that. They are 15 years behind in solving the issues.

And they will go further behind unless they put in the investment to develop completely new fabrication processes, and the further behind they get

And here you acknowledge to an extent what I just wrote.

So you think Sigma will put in more investment than Sony, Canon, Toshiba, Panasonic?

Sigma bought Foveon for a nominal sum.

Not true.

How much did they buy them for, then? Who did they outbid? Put it this way, Foveon was a venture startup, it had investors that put a significant investment (Glynn Capital Management and New Enterprise Investment). Venture Capitalists when they do that do so to get a return on their investment, so when a company is sold 100% as Foveon was, it would be so on the open market, to the highes bidder. I successful startup would have gone public and been listed. Instead it was sold in a private deal, no competitors, which says Sigma was the only option for the investors to get out, which says the company failed.

This is largely wrong with the exception of a couple of names. That's all you'll get from me for now.

There are only a couple of names, and them being right makes the lot right. So, let's see what is largely wrong.

Foveon was a venture startup, it had investors that put a significant investment (Glynn Capital Management and New Enterprise Investment) - well that is clear and public record.

Venture Capitalists when they do that do so to get a return on their investment - seriously, are you saying this is wrong?

A successful startup would have gone public and been listed. - what else do you think is the mechanism by which the VC's get their money back?

When a company is sold 100% as Foveon was, it would be so on the open market, to the highest bidder - you think the VC's are going to do other than get the best deal that they can? You know, if they did that, their investors would be after them.

it was sold in a private deal - well, if it was public, where is the record?

which says Sigma was the only option for the investors to get out - so you say that Sigma managed to outbid the major corporations for this disruptive technology? Really?

For instance, here you see the results for Omnivision, an image sensor startup that succeeded. It is a public company, quoted on public markets. This is what you get if you search for Foveon.

If any of the major players had wanted the technology, they could have easily outbid Sigma.

Not true

You think that none of the massive corporations active in the imaging business had the resources to outbid Sigma? Sony has just put 50bn yen into Olympus. If it had that money to get into Olympus, why didn't it invest in Foveon? You think the Foveon shareholders were keen to minimise their return by allowing only Sigma to bid? Or do you think Sigma was their only option? Which is more likely?

Perhaps it did not come down to a bidding war. Perhaps decisions were made on who to go with for other reasons.

Why do you think the VC's would not want to get the best return for their investment?

Simply, there is little to no commercial interest in Foveon technology and by implication other three layer technologies.

Not true

OK, you say it is not true. Show me some evidence of significant commercial interest in Foveon. Anything at all.

Not yet ripe for prime time.

And it will be some time in the future? When will that be?

Writing a lot of fiction here, Bob. You might give us a cite or two to support these claims.

Quite a lot above, as opposed to nothing to support your contention that it is 'fiction'. You are the one making the claims that need supporting, so lets have some 'cites' to support the following that you need to refute what I said.

i) There are other three layer proposals compatible with standard CMOS technology.

I have not seen one yet that fully fits this bill, but I will grant you that this might be splitting hairs on my part.

That would be a 'no' then.

ii) Foveon had major design wins other than Sigma and Polaroid.

It was not Polaroid but a license holder. And there was a major before Sigma.

A 'licence holder' doesn't make it any better. You're saying that the monkey was better than the organ grinder? Who then was the 'major'. If it was really a design win, it would be on the record, a product, in the market, using a Foveon sensor. There was none.

iii) Sigma was not the only bidder for Foveon.

This is true, but it was not a true bidding war.

It either was or it wasn't. What you're saying was that someone else put in a derisory bid, which just says they did not seriously want it.

iv) No company other than Sigma had the resources to buy Foveon.

Plenty of them did, but the question might not have been one of resources.

Yes, they didn't want it.

-- hide signature --


Post (hide subjects) Posted by
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow