Small Digital Sensors

Small Digital Sensors

  • Technical Reasons

    Votes: 4 13.8%
  • Economic Reasons

    Votes: 17 58.6%
  • Commercial Strategy

    Votes: 8 27.6%

  • Total voters
    29

ernesto

Well-known
Local time
3:29 PM
Joined
Nov 5, 2007
Messages
254
Lots of words have been written regarding the reasons for the not 35mm compatible sensors.
From the camera makers side: Technical and Economical reasons.
"The technology is not ready" "It would be too expensive"
Some users say they are Commercial reasons.
"Camera makers want to sell cameras, and changing the format, is forcing people to buy new equipment"

Which is your point of view?

Ernesto
 
I have read that the economics of chip fabrication is that the cost of fabrication hasn't fallen, but that each generation of electronics gets cheaper because of shrinking microchips able to do the same job (or better) as the previous generation enable higher yields per wafer.

Digital sensors don't benefit from this because, as we all know, we'd like them to be bigger, not smaller. We can get smaller and smaller sensors, but they aren't delivering the same quality, they're delivering less because in their case size does matter.

Until I see an ordered and logical rebuttal of this, I tend to believe it. I'm hard-pressed to think of any other application that, like digital imaging sensors, demands greater size to increase performance.
 
The cost of larger-area sensors is the dominant issue. This is for technical reasons so I am not sure that options (1) and (2) are really separate.

The "commercial strategy" argument doesn't hold, in my opinion. A lot of people happily use full-frame lenses on reduced frame sensors. Look around and you will find some test shots I recently did on a 500 mm 645 lens using an APS-frame Nikon d40x, and believe it or not, the results weren't half bad! The smaller sensor samples the highest MTF portion of the image circle.

In the early days there was a lot of talk about APS and smaller 2x crop factor sensors making lighter lenses possible- this hasn't proven true,, generally. The most reputable and valuable lenses are still the full frame ones. With *possible* exception of some certain new Zuikos.

P.S. Let me suggest rephrasing your question; it is a bit confusing, as I read it. I would say, "reduced frame sensors" or "cropped sensors" or similar.
 
Technical reasons lead to economic reasons. It's a lot harder to make, therefore it costs more.
 
We are not the market. We're a very small subsidiary of it. They make cameras (and thus sensors) for the masses. The masses do not care about image quality past a very basic point. The masses don't know DoF from Dunkin' Donuts.

It is generally cheaper per unit to make 10,000,000 of a thing than 10,000.
 
Stop for a moment and think about the progression of computing chips and how as they have gotten more powerful their price has fallen. Here is a quote from Wikipedia to remind you:

"Moore's Law describes an important trend in the history of computer hardware: that the number of transistors that can be inexpensively placed on an integrated circuit is increasing exponentially, doubling approximately every two years. The observation was first made by Intel co-founder Gordon E. Moore in a 1965 paper. The trend has continued for more than half a century and is not expected to stop for another decade at least and perhaps much longer."

Further down in the article it says:

" Pixels per dollar. Similarly, Barry Hendy of Kodak Australia has plotted the "pixels per dollar" as a basic measure of value for a digital camera, demonstrating the historical linearity (on a log scale) of this market and the opportunity to predict the future trend of digital camera price and resolution."

here is the full article if you are interested.

http://en.wikipedia.org/wiki/Moore's_law

Well while camera sensors are not computer CPUs and its not about fitting transistors on a piece of silicon it seems that the cost and power of camera sensors works in a similar way, as there seems to be a similar principle involved. Engineers and scientists find ways of building more powerful and bigger sensors and manufacture them with greater reliability. But it takes time to (a) develop the technology and (b) build the necessary production facilities.

If you want to know where its heading, Google Hasselblads digital cameras. They are now making 20 and 30 megapixel (plus) sensors (in fact the biggest is 39 megapixels) for their cameras. I do not know how physically large the sensors are, but the retail price for the biggest one is around $40,000 US.

My feeling is that the market will logically head in the direction of 35mm format sensors as once the pro market goes here the semi pro market will follow. Whether that means the market for P&S cameras will go down this road is a good question.

But the cost of making big sensors will almost certainly fall over time. In any mass manufacturing process, a high proportion of costs are more or less fixed (e.g. research and development, cost of setting up big expensive plants and production lines.) The cost of the actual production of each additional item is not all that much. So over time the price of expensive medium / large format or 35mm format sensors will fall on the market as manufacturers first satsify the market for early adopters who are willing to pay more and then need to look for markets which will only buy the sensors if the price is lower. If they do not do that they need to shut the production lines down and thats in no ones interest. (The economists explanation is that the price will fall to approach the marginal cost of production, if you are interested.)

So I am saying............give it time!
 
Last edited:
My feeling is that the market will logically head in the direction of 35mm format sensors as once the pro market goes here the semi pro market will follow. Whether that means the market for P&S cameras will go down this road is a good question.

I think that statement might have been true six months ago, but now it is in jeopardy. This is due to the fact that the only real advantage the large sensors had over smaller ones (to average consumers) was higher quality images at high ISO. None of them care about DoF tricks, etc, that only a larger sensor can bring.

Now Kodak has announced a 10,000 ISO chip that is much smaller than even the standard digicam - good for 5 mp on a cell phone camera, and without the standard Bayer filter. If true - and first reports indicate it is a winner - then the mass of consumers have no more need of a big sensor at all - ever. It gives them nothing of value.

Larger sensors will continue to be made, yes. In fact, Kodak makes the giant sensors for Hassy that you spoke of. But I don't think they will reach the kind of economies of scale that will enable lower retail prices. Instead, the downmarket needs will be served by last year's technology larger chips in a less-expensive model. So in five years, perhaps you'll see a $5,000 31mp sensor that is typically used in Hassy's best of breed today - since the costs will have amortized and they'll be running end-of-life manufacturing push-out anyway. It won't be demand coupled with higher sales that will push prices down, in other words. The pros who need 31 mp today are comfortable paying the exhorbitant prices those cameras command.

I suspect we WILL finally see some more full 35mm sensor size cameras, now that Sony is set to enter the market with several next year and Pentax is retooling what would have been their 645 to fit that chip in a 35mm sized dSLR (oops, did I say too much)? Yes, Samsung has a full-size CMOS image sensor, and Pentax is getting more in the cross-pollination with Samsung that would have been initially expected.
 
Well, new technology is going to continue to raise the bar of expectation, even among "the mass of consumers." People will see that their showoff neighbour has better pics off his camera, or simply a cooler looking unit, and will want one just like it... And twenty years from now, people will still be wanking on the 'net about their superior specs.
 
keithwms said:
Well, new technology is going to continue to raise the bar of expectation, even among "the mass of consumers." People will see that their showoff neighbour has better pics off his camera, or simply a cooler looking unit, and will want one just like it... And twenty years from now, people will still be wanking on the 'net about their superior specs.

Nope. Your average everybody gather around the Christmas tree and say "cheese" family went from cheap 120/620 cameras to 127 cameras to 126 cameras to 35mm cameras to cheap 110 cameras to cheap disc cameras - consumer market, each one worse in terms of quality that the one before it (except for the 126 to 35mm thing, ok). They only came back to 35mm when disc and then 110 slowly became less popular and harder to find. Now they post their incredibly crappy cell phone photos in Facebook or Myspace and are quite thrilled with it all. They don't care jack squat about image quality, and really haven't since the days of neighborhood slide shows and Paul Simon singing about Nikon cameras and Kodachrome.

Only the enthusiasts wank on the net about anything. People like us. And we are not the game. The people on Facebook are the game. They are the market, and that's where the camera companies are aiming the majority of their R&D and marketing dollars.

Enthusiasts always think other people share their values - they don't. They not only don't, but if you try to tell them about it, it annoys them. They don't only not know, they don't want to know.
 
bmattock is correct from what I can see. As a retired guy doing a lot of travelling, what I see is the desire for smaller and lighter cameras in the hands of tourists. Frankly, their 4 x 6 photos are not too bad. Dave
 
Leighgion said:
I have read that the economics of chip fabrication is that the cost of fabrication hasn't fallen, but that each generation of electronics gets cheaper because of shrinking microchips able to do the same job (or better) as the previous generation enable higher yields per wafer.
That semiconductor features get smaller is indeed true, but this doesn't mean the chips get smaller! The functionality that gets crammed into them increases even faster than feature sizes shrink. The bonding pads along the edge of the chip that connect it to the outside world aren't really much smaller today than 20 years ago, and their numbers have increased by a factor ca. 4x over this time. When I was a boy my age, we thought 100.000 logical functions on a die was pretty hot. Now we're just shrugging our shoulders when we stash away a 1.000.000 function module in a corner for some flexibility.

The main reason that cost is reduced in electronic products is that functions that were previously divided over several chips are integrated into one-chip solutions..

Leighgion said:
Digital sensors don't benefit from this because, as we all know, we'd like them to be bigger, not smaller. We can get smaller and smaller sensors, but they aren't delivering the same quality, they're delivering less because in their case size does matter.
Digital sensors are not getting smaller either. The amount of megapixels crammed into them rises. Their size remains the same. There are three factors that make them cheaper:

1. Wafer size increases. 200mm (8") diameter wafers are somewhat long in the tooth, 300mm (12") is mainstream, and 450mm is the next step. 2 x larger wafers means 4x more chips.
2. Once a fab is up and running, the process parameters are constantly tuned to achieve maximum yield (i.e. less drop out due to defects). So over time, the amount of working chips on a given wafer size increases.
3. Once a fab is up and running, the investment is earned back, and the cost of operation decreases. This can be reflected in the price per product.

Leighgion said:
Until I see an ordered and logical rebuttal of this, I tend to believe it. I'm hard-pressed to think of any other application that, like digital imaging sensors, demands greater size to increase performance.
Your logic isn't completely valid. You can increase sensor performance without increasing their size by taking advantage of the possibilities to integrate more electronic functions alongside the photo-sensitive areas.

And yes, you can also increase sensor performance by making larger sensors, but it's no different from increasing computing power by scaling up from a (physically) tiny 8051 to a big Pentium..

Now, how's that for ordered and logical? :D
 
My views aren't any of these. My view is: I use whatever is at hand. I don't care if there's a full frame, half frame, APC sized frame, or whatever in the camera I'm carrying. The only thing that matters is that I'm actually carrying a camera and using it.
 
bmattock said:
Nope. Your average everybody gather around the Christmas tree and say "cheese" family went from cheap 120/620 cameras to 127 cameras to 126 cameras to 35mm cameras to cheap 110 cameras to cheap disc cameras - consumer market, each one worse in terms of quality that the one before it (except for the 126 to 35mm thing, ok).

That's not what I'm getting at. And frankly I think it's a bit too cynical an assessment! People simply want good photos for less money. Is it so wrong? They want to have their cake and eat it too.

10 years go, dSLRs were far out of reach of most casual consumers, they were shooting with coolpixes and such. Now the cost of a "low-end" dSLR is actually lower that the "high end" p&s cameras, so of course people are looking more seriously at the former. A new consumer class is born: the "prosumer" or whatever you want to call it.

OF COURSE less expensive products will always sell in larger numbers, that is basic economics. But that effect is keeping companies alive and fueling new research. With a little bit of perspective, one clearly sees that the price of technology behind more expensive cameras is indeed coming down... slowly. When it reaches the ~$500, that is the the threshold of impulse buying in the U.S.! And people want more camera for their buck, so...

Sure people are using p&s cams in large numbers but... hey, those cameras are in some ways delivering better performance that the first dSLRs did, at the sensor end. It's just that the CoC is too small for good spatial rendering. Most consumers don't know what a CoC is, but guess what, I know a lot of people who don't know a CoC from a CoCK but ask me what do I need to buy to get the blurry backgrounds?!

As for the "low end" dSLRS... well I think some people need to just shut up and take photos. I have a d40x, it was way less expensive than several p&s cameras, and it's a great piece. I personally know of a good half dozen people who sought my advice on whether to buy it after seeing that I was pleased (or dare I say, astonished?) with it. It is selling like hotcakes. Would this have happened if it had been introduced 5 years ago at a $1000 price point? Of course not.

I leave you with one final observation from the lab where I make my living dreaming up new solid state devices. The current sensor and LCD technology on the shelf is totally rudimentary: there will be many more revolutions in digital imaging before we're done. And each revolution will be eye candy to the consumer and make them want to ditch their older camera. What won't change much is the glass, but the sensors and LCDs will go through many more revolutions and the technology will inevitably filter dwn to "the masses."
 
keithwms said:
That's not what I'm getting at. And frankly I think it's a bit too cynical an assessment! People simply want good photos for less money. Is it so wrong? They want to have their cake and eat it too.

Well, cynical or not, it's the truth. People do NOT want good photos. People want photos. They will settle for TERRIBLE photos and love them. I am shown and emailed photos my friends and relations have taken with cell phones - hideous! They love them. Call me cynical, that's fine. But people like crap. Pure crap. They have ZERO interest in 'good photos'. They just want photos.

10 years go, dSLRs were far out of reach of most casual consumers, they were shooting with coolpixes and such. Now the cost of a "low-end" dSLR is actually lower that the "high end" p&s cameras, so of course people are looking more seriously at the former. A new consumer class is born: the "prosumer" or whatever you want to call it.

This happened before. From the mid 1950's to the late 1960's, the middle and upper middle classes discovered slide projectors and SLRs (and also cassette recordings and Super 8 movies) and they went nuts. This was the age of Kodachrome and Nikons. Look at the camera magazines from that era; they didn't cater as much to the clued-in 'enthusiast', they catered to the newbie.

So we see some of the more well-heeled buyers getting into dSLR territory again. But that's still not your average consumer, any more than it was in the Kodachrome years. Most people back then had a Keystone 126 cartridge camera and were happy with it - and most people today are thrilled with their cell phone camera or happy snap low-end Canon digicam.

OF COURSE less expensive products will always sell in larger numbers, that is basic economics. But that effect is keeping companies alive and fueling new research. With a little bit of perspective, one clearly sees that the price of technology behind more expensive cameras is indeed coming down... slowly. When it reaches the ~$500, that is the the threshold of impulse buying in the U.S.! And people want more camera for their buck, so...

No, again - people (and by this I mean Joe Sixpack) do not want more bang for their buck. They want a photo. They want it cheap. AND they want it small. Put it in a pocket or purse, or better yet, make it part of the cell phone they carry around everywhere anyway. Lower-priced digital SLRs are attracting more people into trying them that are disposed to learning more about photography anyway, and that's a good thing, but they hold no interest to the average Joe.

Sure people are using p&s cams in large numbers but... hey, those cameras are in some ways delivering better performance that the first dSLRs did, at the sensor end. It's just that the CoC is too small for good spatial rendering. Most consumers don't know what a CoC is, but guess what, I know a lot of people who don't know a CoC from a CoCK but ask me what do I need to buy to get the blurry backgrounds?!

No, only the people who have more than a vague interest in photography ask things like that. Most don't even notice the blurry backgrounds or the deep DoF that most digicams render.

Your friends may well ask you such things, but Joe Sixpack burps in your general direction and heads for the fridge to get another beer.

As for the "low end" dSLRS... well I think some people need to just shut up and take photos. I have a d40x, it was way less expensive than several p&s cameras, and it's a great piece. I personally know of a good half dozen people who sought my advice on whether to buy it after seeing that I was pleased (or dare I say, astonished?) with it. It is selling like hotcakes. Would this have happened if it had been introduced 5 years ago at a $1000 price point? Of course not.

My parents would not have bought it. They loved their Kodak Disc camera. They thought my photos looked great, but they had no interest in things like focusing, setting exposure, deciding what f-stop to use, etc. They just wanted to take a picture. And that is what 99.99% of all people want. A) to take a picture and B) to be able to keep the camera in their pocket.

I leave you with one final observation from the lab where I make my living dreaming up new solid state devices. The current sensor and LCD technology on the shelf is totally rudimentary: there will be many more revolutions in digital imaging before we're done. And each revolution will be eye candy to the consumer and make them want to ditch their older camera. What won't change much is the glass, but the sensors and LCDs will go through many more revolutions and the technology will inevitably filter dwn to "the masses."

You may know tech, but you don't know people. People are not interested in learning technology - they just want to use it - and only when it fits their lifestyle, which is increasingly dumbed-down for them. They want big TV sets, they don't want big cameras. For years, people stopped taking photos altogether, compared to the days of Kodachrome. That's because 110 was on the wane, APS was dying, and 35mm cameras, even PnS single-use cameras, were more than they wanted to carry around. Now they are taking photos again, and with new websites like Facebook and Flickr and MySpace, everybody suddenly wants photos of the places they do and the people and things they see. That's cool. But if it does not fit in pocket, it isn't going with them. These people are not going to be seduced by dSLR cameras at incredibly low prices - THEY ARE TOO BIG.

You and I are enthusiasts. Our friends tend to be technology buffs and intelligent enough to grasp the benefits of a big sensor and a dSLR. We get it. We're not the market. The market is a big gigantic idiot with barely enough sense to pour piss out of a boot with the instructions on the bottom. Manufacturers serve the market - they have to. That's not cynical - that's being realistic.
 
I just won't waste my time.

I teach device physics, and I teach traditional b&w photography... at the university level... in the same semester. I am not a walmart clerk, and neither am I some white-coated wonk who never gets out of the lab.

Look at the OP's question, comment on it, and let others comment on it without getting all lathered up. Share your perspective but respect others'.

<end of silly thread for me>
 
keithwms said:
I just won't waste my time.

You wasted it telling me you're not going to waste it? No, you want to say that you have no argument anymore, and it makes you mad.

I teach device physics, and I teach traditional b&w photography... at the university level... in the same semester. I am not a walmart clerk, and neither am I some white-coated wonk who never gets out of the lab.

If you teach device physics, then you're a brainiac, and hence, not in touch with people who watch "American Idol," which is most of the booger-eatin' morons out there. You have no idea what they want.

Look at the OP's question, comment on it, and let others comment on it without getting all lathered up. Share your perspective but respect others'.

I did. I also responded to YOUR comments to ME. I respect your perspective to the extent that I responded to it and addressed your individual statements with counter-arguments of my own. What you're actually saying is "agree with me, or I'll take my toys and go home. And note that I am actually a rocket scientist, please." Yes, yes, I respect your intelligence. I disagree with your conclusions. I take it you're not all that used to having your logic shredded in front of your eyes. Sorry.

<end of silly thread for me>

Don't go away mad...
 
ernesto said:
From the camera makers side: Technical and Economical reasons.
"The technology is not ready" "It would be too expensive"

In silicon chip manufacturing, larger chips are far more expensive than smaller chips. Specifically, if you double the chip area, the cost in increases by a factor of about 16! (I.e., cost is proportional to the fourth power of area.)

Here's an online app that shows how many chips can be produced from a silicon wafer:
http://tams-www.informatik.uni-hamburg.de/applets/yield/index.html

If you increase the chip size, you'll see that the yield drops dramatically. Also, here are some lecture notes explaining chip yield vs size vs cost:

http://bnrg.eecs.berkeley.edu/~randy/Courses/CS252.S96/Lecture05.pdf
 
Regardless of sensor size, FWIW, one must not overlook the oft overlooked warning sticker on the bottom of most digitals, which reads:

"Do not smash camera on the person next to you's head. You will damage the camera and void the warranty".

Be mindful of this if your camera is still under warranty, I can tell your from experience. If the warranty of your camera has expired the sticker is "moot" and you may bash away.
 
Last edited:
NickTrop said:
Regardless of sensor size, FWIW, one must not overlook the oft overlooked warning sticker on the bottom of most digitals, which reads:

"Do not smash camera on the person next to you's head. You will damage the camera and void the warranty".

Be mindful of this if your camera is still under warranty, I can tell your from experience. If the warranty of your camera has expired the sticker is "moot" and you may bash away.

It was funny the first time you posted it. Kind of. Now it's not.
 
sonofdanang said:
The interesting thing about polemic is that it gives life to generalization. Joe Sixpack? - I like the characterization. But it's one end of a spectrum. Hideous photos? Perhaps. Depends on the yardstick. It's not that most people 'don't want good image quality' it's that they are interested in content -- traditional measurements of IQ simply don't matter once things are above a threshold. You will find as many thresholds as users. I don't do facebook. I'm a long-time usenet-er. I shoot almost entirely digital but have a IIIc with a 50/3.5. My wife does Facebook, shoots with a Leica (M3, and R4), and a little $100 Canon P&S. Guess what's most often in the purse? And why?

You're using anecdotal evidence. I use trends that are clearly provable. What sells? Camera phones and cheap digicams. What have camera execs been fretting about and film lovers using as yet another reason to hate digital? Not many people print their photos. What are the most popular shows on television? Brain-dead 'reality' shows like "Dancing with the Stars" and "American Icon" and, God help me, "Desperate Housewives." You know smart people. So do I. But the majority of people are booger-eatin' morons of a rather low order. The proof is in the Leica-versus-cheap Digicam sales figures.

Alex Majoli shoots with anything but, lately with point and shoots. Technique makes the difference.

You miss my point, I think. Perhaps I didn't explain it well. When I say the poopy-drawer snap-shot artists make 'terrible' photos, I was not referring to photographic talent. I was referring to the technical quality of the image itself, whether it is of odd twins in some suburban backyard holding hands, or the burial at sea of the latest goldfish victim of young Bobby over here. The image quality of a cell phone camera is rather suckish - regardless of the abilities of the photographer or the compelling drama of the subject, it makes a Holga look like an M7.

I'm not saying most droop-drawers suck as photographers (although I have no doubt they do), I am saying their tool of choice is terrible in terms of quality, and they don't care at all.

With reference to the raging, you're both right. And wrong. The world is a huge greyscale of abilities and levels of awareness. For a dopple-metaphor, just look at the fw corrections for fall-off in the M8. The same thing applies to the piss-boot operators. Give Joe Sixpack an M8 or a MF machine all tethered with a 40 MP sensor and he will still produce dreck. Maybe. It's a sense of design and the ability to bring technique to bear that differentiates Leonardo Da Vinci from that idiot down the street building a fire pit out of essentially the same stone. But it's still just stone. Image quality is nothing without content. One the other hand, I wonder why disk film and most of the other ultra-small film formats perished?.... Hmmm.

You make a good point with your query as to why the smaller film formats vanished, but I would tend (without proof, I admit) to believe that this was simply due to the inability to achieve market penetration to the extent that they were ubiquitous. As each began to flag in sales, instead of propping them up with marketing, the camera manufacturers and Kodak went to the next format and more-or-less abandoned the old.

I also accept that everything is on a nice bell-shaped curve, including the imbecility of most human beings. But the majority are within a few brain cells of each other, and slightly above the average housecat.

To the original post:

There is no one small-set of reasons. It's a large set. The market is part of it. But also is the quirkyness of the human need to explore options - from the techlab to the street. The threshold is being crossed rapidly, again and again. Part of the problem is that we are thinking of digicam sensors applied to a 35mm film model. One of the things that differentiated Leica from other 35mm offerings (apart from Leitz optics), is the slightly different geometery in the camera and lens working distances. The whole 35mm geometery changes DoF issues. Use the tool that is appropriate for your needs. Nobody serious about landscape photography uses 35mm formats......except those who are applying a sense of design to the short comings of any particular piece of technology. And that happens in any format, in any technology, in any endeavor.

All true, and all important - to us. Not to Joe Sixpack. I must insist; he neither knows nor cares about such things. The proof is all around us. We who are intelligent and who are involved in photography tend to make assumptions that others feel as we do, that they understand our issues, and that we represent the market that the major camera makers are trying to connect to. We're not.

I think that we will see all kinds of interesting things in the near future. Perhaps non-planar sensors that compensate with focal lengths to achieve better edge performance in very small form factors. I'll pick it up, try it, and if I can get something out of it, I'll consider it a candidate for my tool kit.

I agree. I am looking forward to innovations too. My point, contrary to the O/P, was that large-format sensors are not going to become the default standard for digital cameras, nor do 'people' want the advantages that a larger sensor can bring. They might have - if a larger sensor had remained the sole method of obtaining higher ISO / lower noise; but Kodak shot that in the head with their newer, and even smaller, sensors announced recently. Now there won't be a good reason not to keep making sensors EVEN SMALLER for the masses. So those sensors will drop in price - not the ones we want, other than through the usual efficiencies achieved in manufacturing that tends to drive tech prices down.

Who knows? Who cares?

Go out and take some pictures. It's more fun to rage about content!!!

Everyone thinks I am mad when I post. I'm not 'raging' about anything. I speak emphatically.

And I shot a few hundred images today. I do that nearly every weekend. For all my blah-blah-blah, I am constantly practicing the art of photography in an attempt to improve my abilities and to enjoy my hobby.
 
Back
Top Bottom