[jeffkrol]

Re: Hoya/Pentax - answer to questions about Pentax future: Pentax SLR Talk Forum: Digital Photography Review
I trust this interpretation..... sometimes you just have to go with your gut How can you not trust a Swedish researcher.
http://www.blogger.com/profile/06834641501438709866
Two things that stick out to me:
1)False. Misreading. It's the article author that speculated that since Pentax/Hoya is not involved, perhaps it's Sigma. But he asks the Sigma head (who's at the show), who says he's not the first to ask, and categorically denies it.
If the OP had actually read the article he would not have made a statement completely opposite to what the article actually states.

2)Again (and some other poster made the same point), it's not about anything incompatible or anything. They also said/hinted that the 30mm and so on are not canceled but will be part of the new lens lineup - though I'm not good enough at Japanese to really be sure how strong the implication was here.
In short, the OP is rather misrepresenting the article (it's not really an interview as most content is the writer's speculation, not statements from anyone at Pentax) and a lot of the speculation in the thread are based on those misrepresentations.

[falconeye]

How can you not trust a Swedish researcher.

Yes, and this is him:

It became pretty obvious that the Sigma part was misinterpreted. RH didn't say anything wrong here (as others did) but he also didn't translate the denial of Sigma involvement. So, it isn't that easy to read Japanese text for him either.

So, to summarize, nobody (so far) has offered a native translation of the sentence above mentioning FF.

[jeffkrol]

Yes, and this is him:

It became pretty obvious that the Sigma part was misinterpreted. RH didn't say anything wrong here (as others did) but he also didn't translate the denial of Sigma involvement. So, it isn't that easy to read Japanese text for him either.

So, to summarize, nobody (so far) has offered a native translation of the sentence above mentioning FF.

Seee.......
As to to RH translations, yes it is not always "on the money"....

[*isteve]

http://www.robgalbraith.com/public_files/Canon_Full-Frame_CMOS_White_Paper.pdf
Will you take Canon's word for it???????
$450 to $500, $1,000 or even $5,000. After several hundred process steps, perhaps
between 400 and 600 (including, for
example, thin film deposition,
lithography, photoresist coating and
alignment, exposure, developing,
etching and cleaning), one has a wafer
covered with sensors. If the sensors are
APS-C size, there are about 200 of them
on the wafer, depending on layout and
the design of the periphery of each
sensor. For APS-H, there are about 46
or so. Full-frame sensors? Just 20.
Consider, too, that an 8" silicon
wafer usually yields 1000 to 2000 LSI
(Large-Scale Integrated) circuits. If, say, 20 areas have defects, such as dust or
scratches, up to 1980 usable chips remain. With 20 large sensors on a wafer, each
sensor is an easy “target.” Damage anywhere ruins the whole sensor. 20 randomly
distributed dust and scratch marks could ruin the whole batch. This means that the
handling of full-frame sensors during manufacture needs to be obsessively precise,
and therefore they are more expensive.
Of course, there is more to this topic. For example, the circuit pattern of a fullframe
sensor is too large to be projected on the silicon wafer all at once; it requires
three separate exposures (See page 53). This means that the number of masks and
exposure processes is tripled. For now, appreciate that a full-frame sensor costs not
three or four times, but ten, twenty or more times as much as an APS-C sensor. Here,

Your numbers (and Canon's) tally with information I have from contacts in the microprocessor industry.

For large scale production, the "guideline" cost per chip I had suggested to me was $50 $500 and $5000 for APSC, FF and MF sensors but for initial small scale batches on a new design that could be a lot more. That is the cost to the sensor fabricator, not the price they charge to the buyer which is more still.

Furthermore, the chip cost/area equation has not changed for some years. The use of larger wafers only makes a small difference because the cost per wafer is a lot higher. For large chips, there is not much difference between 8" and 12" processes but if you are making LSI's, its a significant saving.

[jeffkrol]

Thank you for that info.

[falconeye]

We will probably be unable to settle this in a satisfactory manner.

Let me summarize the situation:

From the own words of the industry, the costs for CMOS chips are rather high and the word is spreading that those are like $30/300/3000 or $50/500/5000 for APS-C/FF/MF. But all sources are vague or by spoken word.

From the only source which we can probe and which is public (the one I provided above):
IC Knowledge - Wafer Costs
we learn, however, that an entire 300mm wafer is only $2500 to make!

Of course, you have to add for yield, testing, cutting, packaging, development depreciation, reticle depreciation and seller's margins. What I tried to do in my above posts. And what would still lead to a much lower cost estimate.

The purpose of ICKnowledge is to make the mysterious costs in semiconductor industry transparent for competitors and prospective buyers (purchase of their cost model is $2000 per copy!). Which means that it must be standard practice to obfuscate those costs.


The only logical conclusion seems to be this: semiconductor industry has obfuscated costs of (large) CMOS chips to the point where user's of (large) CMOS chips can ask high market prices. And nobody on the vendor side of things can possibly have an interest to make the true costs transparent. Including semiconductor or camera makers.


Which is why it is our (the consumer's) task to make these costs as transparent as possible and insist to get the products which we actually already should get for the money we pay anyway.

Any help in this mission will be appreciated.
As is providing a source which can be probed that the true costs have been underestimated by me. But we need hard facts.


To say it again: The word as it is spreading means that an MF chip is twice as expensive as an entire 300mm wafer!! Something goes wrong here...

[jeffkrol]

We will probably be unable to settle this in a satisfactory manner.








To say it again: The word as it is spreading means that an MF chip is twice as expensive as an entire 300mm wafer!! Something goes wrong here...

I work in retail.. there is nothing mysterious here. Don't forget all those outrageous salaries.... and in real terms miniscule markets..Anyways we shall see... and I am partially in agreement with your calculations.... but still think your low.. these companies still have to compete with each other (unless they are monopolizing or forming cartels) . Only takes one to break it for short term gains. Since this hasn't happened (yet) I assume the true costs are still high.....
Mystery of free market economics...

[Sean Nelson]

The word as it is spreading means that an MF chip is twice as expensive as an entire 300mm wafer!! Something goes wrong here...

My figure of $3500 for a Kodak 36 x 48mm sensor comes from Kodak's press release which is reproduced here: EEProductCenter.com :: Kodak delivers next-generation 50-MP CCD image sensor. Rail against it as much as you want, but to the best of my knowledge those are the simple facts.

I see no point in speculating on the manufacturing costs. The fact is that Kodak has priced their chips based on what they feel the market will bear. There's no reason why raw manufacturing cost needs to be anywhere close to the selling price, as evidenced by movies/software/cosmetics/precious gems/etc. etc.

[*isteve]

We will probably be unable to settle this in a satisfactory manner.

Let me summarize the situation:

From the own words of the industry, the costs for CMOS chips are rather high and the word is spreading that those are like $30/300/3000 or $50/500/5000 for APS-C/FF/MF. But all sources are vague or by spoken word.

From the only source which we can probe and which is public (the one I provided above):
IC Knowledge - Wafer Costs
we learn, however, that an entire 300mm wafer is only $2500 to make!

Of course, you have to add for yield, testing, cutting, packaging, development depreciation, reticle depreciation and seller's margins. What I tried to do in my above posts. And what would still lead to a much lower cost estimate.

The purpose of ICKnowledge is to make the mysterious costs in semiconductor industry transparent for competitors and prospective buyers (purchase of their cost model is $2000 per copy!). Which means that it must be standard practice to obfuscate those costs.


The only logical conclusion seems to be this: semiconductor industry has obfuscated costs of (large) CMOS chips to the point where user's of (large) CMOS chips can ask high market prices. And nobody on the vendor side of things can possibly have an interest to make the true costs transparent. Including semiconductor or camera makers.


Which is why it is our (the consumer's) task to make these costs as transparent as possible and insist to get the products which we actually already should get for the money we pay anyway.

Any help in this mission will be appreciated.
As is providing a source which can be probed that the true costs have been underestimated by me. But we need hard facts.


To say it again: The word as it is spreading means that an MF chip is twice as expensive as an entire 300mm wafer!! Something goes wrong here...

1. The specific process you are talking about is not necessarily relevant, but lets assume it is.

2. The yield per wafer may be as low as 3-4 chips for FF or less than 1 for MF.

3. You are forgetting to add the cost of the post-processing (bayer filter and microlenses) which add considerably to the post production cost.

I dont see this as very inconsistent when you add margins etc.

Also, the cost of chips made in the far east in US $ terms is probably 20-30% higher than it was.

[thibs]

To say it again: The word as it is spreading means that an MF chip is twice as expensive as an entire 300mm wafer!! Something goes wrong here...

I might want (or not) that MF chips are CCD chips and CCd chips or quite different in manufacturing than CMOS (not necessarily more expensive btw). Why do you think MF chips are CCDs?

Just for the fun of it? Of course not. It is either less expensive or easier to make or have a better yield or... probably all of that.

You also forget the fact that CMOS need more than one pass manufacturing. If they need more than a wafer, well double the cost.

All in all, it seems to me you really do not understand what the CMOS/CCD manufacturing implies.

You might want to read a couple articles on the subject. CPUS/GPUS manufactring should be interesting.

[tigrebleu]

I hope it is true but I will only believe it when I see a camera I can purchase.
Douglas Film

Agreed.

With so much "645D is comming", "No it's not" statements in the past, I've turned sceptic when it comes to the Pentax digital medium format.

If it does come true, I hope Pentax can make it affordable enough to compete with pro FF DSLRs. Spending US$ 10 000 on a 645D isn't much different than spending US$ 8000 on a Nikon D3X, for instance.

But if it's in the leagues of other high-end MF digital systems (Blad, POne, Leaf, Mamiya), it could prove too expensive. I don't know many pro photographers who would spend US$ 20 000 on a Pentax MF DSLR when they can get full pro support from Blad, POne, etc. for the same price.

The 645D could be the first successful MF DSLR aimed at the (wealthy) amateur.

After all, there is certainly market for that. If people can spend US$ 20 000 on sport motorcycles, vintage cars or home theatre systems, why wouldn't there be a market for a US$ 10 000 medium format DSLR?

[ghelary]

If it does come true, I hope Pentax can make it affordable enough to compete with pro FF DSLRs. Spending US$ 10 000 on a 645D isn't much different than spending US$ 8000 on a Nikon D3X, for instance.

But if it's in the leagues of other high-end MF digital systems (Blad, POne, Leaf, Mamiya), it could prove too expensive. I don't know many pro photographers who would spend US$ 20 000 on a Pentax MF DSLR when they can get full pro support from Blad, POne, etc. for the same price.

Most pros rent their equipment, especially in this price range. Given that, the problem of Pentax is to convince the "pro" brick and mortar shops to hold a lineup of lenses and bodies on 645 system.

The current MF specialized companies are far from being wealthy at the moment (including Hasselblad), so Hoya's financial power may be a good argument for long term support and future of this system.

As FF becomes affordable for amateurs, a lot of pros will want to upgrade from FF to MF (landscape, fashion, art photography, architecture...) to keep a technical gap between pros and amateurs.

At the moment, we have seen a lot of reactions from amateurs, but still I haven't seen anything from the few pros who still shoot Pentax (and some of them still shoot 6x7 fim)

Regards,
Guillaume

[gazonk]

But if it's in the leagues of other high-end MF digital systems (Blad, POne, Leaf, Mamiya), it could prove too expensive. I don't know many pro photographers who would spend US$ 20 000 on a Pentax MF DSLR when they can get full pro support from Blad, POne, etc. for the same price.

The 645D could be the first successful MF DSLR aimed at the (wealthy) amateur.

I think that's not only a question of the price, but the usability. Apparently (I have never tried these beasts myself) the current MF DSLRs have a very different approach to RAW data than even the pro "FF" DSLRs have, the RAW files being "rawer" and requiring more post processing.

If the 645D is as lightweight and compact as the 645 film bodies, it already scores som plus points in usability. If it's also as easy to use as a (pro) small-format DSLR, it might look very attractive to wealthy amateurs, and I won't be surprised if there are quite a few japanese photo amateurs which are willing to spend $20000 for such a body.

One open question is what crop the camera gets. If it's a 48x36 sensor, the current 75 mm is no longer a "perfect normal" (corresponds to 43mm on 35mm), but it's still within the "normal range" (it then corresponds to 54mm on 35mm). Maybe more of a concern is the fact that the 35mm wide will bli reduced from 20 to 25 mm in 35mm-equivalents.

[Tommy]

Back when we were all speculating about the successor to K10D benjikan posted that Pentax may not use a Sony sensor. Someone suggested the new sensor would come from Samsung and a lot of people said that this would be impossible since "even Samsung" would never be able to compete with the Sony sensors since they had no history in making sensors as large as APS-C.

If someone suggested that Samsung would make the MF sensor for the 645D, I guess the same statements would apply: "there is no way they can compete with Kodak!". Which is why I won't suggest it...

[creampuff]

Going by Kodak's checkered history with digital sensors, I hope Pentax never uses a Kodak sensor in the digital 645. Seems like everything Kodak has dabbled in digital is a resounding flop. Here's hoping the MF sensor is a CMOS and not CCD sensor.