The next crop of cameras is coming down the chute

[eronald]

Hum... it really depends whether their current equipment enables them to manufacture such large sensors. If they don't, then the cost of machines alone would eat up a significant part of the margin.

I would love this to happen though.

Cheers,
Bernard

Samsung wouldn't have to do anything special to manufacture; the existing MF designs are rolled out on standard fabs, I believe.
Fullframe 35 may be different in some cases, because Canon and Sony may be using their own fabs, and even tweaked processes.

Edmund

[eronald]

Exactly, plus unless they've cracked the low yield issue, the cost per sensor (larger than 24x36) is going to be very high....possibly to the point of not being profitable...and nobody makes camera for charity...

Yair,

 The yield can be approximated by a mathematical curve which depends on process parameters, quality of the design, and device surface area.
 http://www.siliconfareast.com/test-yield-models.htm
 As a results one can infer that if Samsung can already produce APS with the same yields as everybody else, then they have similar parameters as everyone else in the yield curve, and thus when they try a large device (MF) they will again get roughly the same (lower) yield as everyone else. This quick answer was no good when people were getting 1% yields or something and MF sensors cost a fortune, but now we are in calm waters, it may serve.

Edmund

[Radu Arama]

I have no reason to doubt your marketing numbers, and Sony could of course just recycle an existing cell design into a large sensor, thereby giving Pentax what they need. Whether Sony would be happy creating themselves a competitor for their own flagship 35mm full frame is probably the main question to decide whether this happens.

However, Pentax could also commission a sensor from Samsung; I think this is a more probable resolution.

Edmund

Hello Edmund,

Of course Samsung could do such a sensor as well but I guess Japanese companies buy foreign at last resort. Besides, Pentax is (except for the current Kodak KAF40000) an all Sony customer from the humble Q to every single dslr since K-7 (2009) and that means hundreds of thousands APS-C sensors per year (plus they received the 16,2 Mp sensor at the same time as Nikon so they must have some respect fron Sony's part). IMO it makes sense for Sony to cater to their needs and not let another (Japanese or foreign) company get on the line. It is pretty clear that eventually larger than 36x24 mm CMOS sensors will appear so for me it's logical that one brand must take the initiative.

Regarding the competition with Sony's FF cameras all signs are that Sony's sensor division is pretty much independent and has the confessed ambition to be number one in CMOS sensors. I think they could even want to do such large sensors for bragging rights.

Regards,
Radu

[eronald]

Hello Edmund,

Of course Samsung could do such a sensor as well but I guess Japanese companies buy foreign at last resort. Besides, Pentax is (except for the current Kodak KAF40000) an all Sony customer from the humble Q to every single dslr since K-7 (2009) and that means hundreds of thousands APS-C sensors per year (plus they received the 16,2 Mp sensor at the same time as Nikon so they must have some respect fron Sony's part). IMO it makes sense for Sony to cater to their needs and not let another (Japanese or foreign) company get on the line. It is pretty clear that eventually larger than 36x24 mm CMOS sensors will appear so for me it's logical that one brand must take the initiative.

Regarding the competition with Sony's FF cameras all signs are that Sony's sensor division is pretty much independent and has the confessed ambition to be number one in CMOS sensors. I think they could even want to do such large sensors for bragging rights.

Regards,
Radu

Radu,

 I would agree with you; however there is one japanese company with the technology and a historical interest in MF namely Fuji; if a Japanese company ends up fabricating MF chips and relaunching MF domestically, I would expect it to be them, with a view to giving their captive marriage photography business a boost.

Edmund

[Radu Arama]

Hum... it really depends whether their current equipment enables them to manufacture such large sensors. If they don't, then the cost of machines alone would eat up a significant part of the margin.

I would love this to happen though.

Cheers,
Bernard

Hi Bernard!

I don't expect a possible "close to 645 film sized sensor" Pentax camera to cost 10K USD so "margin" is an unknown variable here (but even if the sensor is to cost . If such a camera will come I expect it to cost at least 1.5M Yen (almost 20K USD) but if Pentax could do at least:

- reliable tethering (this is work in progress)
- modern live view
- maybe some form of movie mode  
- make available worldwide all the FA 645 lens range (also on the way to be realized at least in Europe many lenses are in stock and more exotic ones are available to order)
- make available worldwide the new D FA 25 and the new D FA fast portrait lens (maybe with leaf shutter)  
- organize a more convincing support network in Europe and North America

I think they would have great chances to sell such a camera with a proportional success compared with 645D did but in the higher end class.

Regards,
Radu

P.S. I know that you will watch closely the next CP+ in less than 5 months, there could be lots of answers on what Ricoh thinks about Pentax MF range.

[Radu Arama]

Radu,

 I would agree with you; however there is one japanese company with the technology and a historical interest in MF namely Fuji; if a Japanese company ends up fabricating MF chips and relaunching MF domestically, I would expect it to be them, with a view to giving their captive marriage photography business a boost.

Edmund

Edmund,

IMO Fuji has the same disadvantage as the rest (H, P1) vs. Pentax: they simply can't offset the cost of electronics - parts and R&D - (and most of the software) to hundreds of thousands of dslrs per year. And in the long term I believe that Pentax will be able to update more frequently the electronics from the same reason.

Radu

[hjulenissen]

Yair,

 The yield can be approximated by a mathematical curve which depends on process parameters, quality of the design, and device surface area.
 http://www.siliconfareast.com/test-yield-models.htm
 As a results one can infer that if Samsung can already produce APS with the same yields as everybody else, then they have similar parameters as everyone else in the yield curve, and thus when they try a large device (MF) they will again get roughly the same (lower) yield as everyone else. This quick answer was no good when people were getting 1% yields or something and MF sensors cost a fortune, but now we are in calm waters, it may serve.

Edmund

When a sensor is defective, does this mean that it is "dead", that one column/row is dead, or that a single sensel is dead? Does it mean that the sensor cannot be "cut" into one or more smaller sensor? Generally, my question is if the methology and formulas for baking low-redundancy cpu logic can be transferred directly into imaging sensors.

I imagine that as resolution increases, the practical obstacle of interpolating one sensel or one row of pixels is less apparent.

-h

[ErikKaffehr]

Hi,

Michael Reichmann had a discussion with Phase One engineers on that issue. I got the impression that it was quite frequent for a complete column of cells on the sensor to fail, but that would be taken care of by camera electronics. I got the impression that sensor cost was about half of the cost for the back and that sensor failure rate was about 25%.

The interview was on one of the old LLVJ DVDs.

Best regards
Erik

When a sensor is defective, does this mean that it is "dead", that one column/row is dead, or that a single sensel is dead? Does it mean that the sensor cannot be "cut" into one or more smaller sensor? Generally, my question is if the methology and formulas for baking low-redundancy cpu logic can be transferred directly into imaging sensors.

I imagine that as resolution increases, the practical obstacle of interpolating one sensel or one row of pixels is less apparent.

-h

[eronald]

Those interviews are very old. My guess is a 36x48 sensor *now* costs substantially less than $500, whatever the number of pixels.

Hasselblad could probably sell an entry level body for less than $6000, and I'd be very surprised if at year's end they didn't do exactly that.

Edmund

Hi,

Michael Reichmann had a discussion with Phase One engineers on that issue. I got the impression that it was quite frequent for a complete column of cells on the sensor to fail, but that would be taken care of by camera electronics. I got the impression that sensor cost was about half of the cost for the back and that sensor failure rate was about 25%.

The interview was on one of the old LLVJ DVDs.

Best regards
Erik

[hjulenissen]

Those interviews are very old. My guess is a 36x48 sensor *now* costs substantially less than $500, whatever the number of pixels.

Hasselblad could probably sell an entry level body for less than $6000, and I'd be very surprised if at year's end they didn't do exactly that.

Edmund

Because of competition from Pentax? Nikon/Canon?

Have anyone ever looked into the defects in cameras out there? Could it be that the price-cutting MF cameras use the sensors that have been rejected for the pricier ones? (rather use a sensor in a less critical camera than throwing it in the bin, just like some LCD tvs have more defect pixel tolerance than others).

-h

[yaya]

Because of competition from Pentax? Nikon/Canon?

Have anyone ever looked into the defects in cameras out there? Could it be that the price-cutting MF cameras use the sensors that have been rejected for the pricier ones? (rather use a sensor in a less critical camera than throwing it in the bin, just like some LCD tvs have more defect pixel tolerance than others).

-h

You can't take a "defective" 50MP 49x37mm sensor and put it in a price cutting 40MP 44x33mm camera...it just doesn't work like that

In general it is easier to achieve higher yields with CCD chips than with CMOS chips (that are suitable for stills photography). This is only One of the reason why we do not see larger-than-35mm FF CMOS sensors

(probably worth a reminder here of who made the first FF 35mm CMOS sensor)

Yair

[hjulenissen]

Higher yields of CCDs at lower volumes and higher prices isn't what the market wants though. I want lower priced, higher volume CMOS but with the option of larger physical size and higher MP that fulfill the original master designs of 6x6 and 6x7 cameras, even 4x5 or higher. I simply don't want an overpriced "high yield", low volume, high price digital back that won't go larger than 645. Not sure why that destiny is so oblivious to the medium format digital manufacturers.

I want a Ferrari, but probably will never have one :-)

The fact that 4x5 digital sensors are nowhere to be found in regular photography suggests to me that they are/would be difficult and expensive to produce compared to the perceived benefit (sellability) they would have over FF and MF sensors. If there are 5 manufacturers easily could make a really large digital sensor that was suitable for a $4000 camera that would blow FF and MF away in terms of quality and enough buyers out there appreiciated that quality high enough to pay that money, surely we would have seen one?

I think that the way digital sensors scale in terms of quality/size/price is very different from how film scaled. Applying old-world norms may not make much sense if you want to maximize "IQ" per dollar.

-h

[yaya]

I want lower priced, higher volume CMOS but with the option of larger physical size and higher MP that fulfill the original master designs of 6x6 and 6x7 cameras, even 4x5 or higher.

That's the challenge right there....

[ced]

I think that after 20+ years the sensor technology, other than getting larger have not really advanced too much.
I would have imagined that by now with the possibilities that we have of high resolution masking techniques they could produce sensors that would eliminate moire by building some random patterned array with irregular shaped sensels...

[ondebanks]

I'm sure the ones at Sony and Nikon know, and also the guys at Dalsa and Kodak who do the military sensor chips.

That's the bit that puzzles me. If the entire MF digital world is built around piggybacking on Kodak and Dalsa's large designed-for-military sensors, why aren't they CMOS yet? Who needs low light sensing more than the military? Who needs fast frame rates (live view, video) more than the military? Large CCDs just can't cover those applications, but CMOS can, so why would the military not insist on CMOS?

Ray

[uaiomex]

I've said before in another photo forum. "24X36 is the holy grail of digital photography".
Everything goes perfectly logical with sensors up to this size. Suddenly, just a few millimeters above, things become strangely more complicated, extremely expensive and reliability drops significantly. Why? It totally beats me. I ended up taking this true as one more of those difficult to believe, strange, awkward or unpleasant things of life that people learn to accept as dogmas, because no matter how you reason it, the math doesn't add up.
Once I came to this personal conclusion about FF, my lust for DMF terminated. That's was a couple of years ago and made a happier photographer after I started on concentrating my efforts on building a yet better and stronger system for it.
I couldn't be happier. I'm sure that with the exception of those pixel-peepers that inhabit the medium format forums of the world, the IQ from current FF cameras is found amazing. Day-in and day-out. Everyday. No issues, just beautiful pictures. 
Eduardo


 

The MFDB companies are simply protecting their interest and and have no interest in actually selling what ALL medium format shooters past and present want, more reasonable priced backs/sensors that actually accomplish more. And since hobbyists with the cash seem to make the over 51% of the buyers now, what pros want or need is taken less and less into consideration. The technology is simply now too expensive for how slow the improvements are from model to model every few years. And honestly the hassle factor makes them even more overpriced. If I can have a $2500 5DII which has NEVER had a single issue EVER in 50K-100K exposures, you wonder what the hell are MFDB makers, well making? My experience with digital backs was averaging an issue maybe every 500 frames, didn't matter what camera body or lens, or temperature, or software, or cable. Nothing but general annoyance with using medium format digital. Funny my Bronica ETRSi was hassle free, Rollei 6001 bodies, Mamiya 645 AFD and RZ's all hassle free with film. My personal user error with film was a whopping 2 times in the last 17 years. Find someone who doesn't have 2 issues a week or more with the back, software, cables, connections and more with MFDBs.