Affordable full frame sensor

[EXA1a]

When you look at the pixel densities (pixel per square millimeter) of all Canon DSLRs you notice a great variety. Calculating a full frame sensor for these densities you would get this:
1D CCD: 7408 pix/mm2, full frame: 6.3 mp
D30 CMOS: 9074 pix/mm2, full frame: 7.7 mp
1Ds CMOS: 12897 pix/mm2, full frame: 11.0 mp (existing)
10D CMOS: 18355 pix/mm2, full frame: 15.6 mp

For my next digital EOS it would be most important to get a full frame sensor so that I could fully use the wide angle end of the spectrum (means: the 94° angle of view of an EF20mm/2.8 would not convert to a 60° like a 32mm lens).

With that in mind I would speculate for a sensor with a D30-like density leading to 7.7 megapixels. I would not need more resolution than this. For the high end even a 15.6 mp sensor seems not to be out of reach.

Just thinking of the next EOS Digital I would want to buy for less than 2000$...

--Jens--

[w10d]

I was reading an article on sensor construction about a month ago (somewhere online), it was quite a technical piece, but a bit disappointing: The conclusion was that construction of full frame sensors was likely to remain costly for the next couple of years at least, with manufacturers concentrating on cheaper wide angle lens design, and cheaper small sensors (by increasing volume sales, etc.).

Wish I could remember where I saw it....

[Belmondo]

Don't count on Canon giving you that all at one time. They'll probably step up to a full-frame sensor a little at a time. That's just good business on their part.

If they do, however, it probably won't be cheap.

(I say all this because I have a well-established track record of being wrong about such things, so predicting it won't happen is almost a guarantee that it will. Let's see if it works.)


---Tom---

[robertwgross]

Let me explain this again for the benefit of those who weren't present earlier.

Building the perfect image sensor is not a "camera thing". It is a "semiconductor thing." Semiconductor chips are fabricated in large factories called "fabs", and they cost something along the lines of US$1B to construct. Lithography is the set of procedures whereby they get from a chip design to something that is optically laid out on the semiconductor die. The current system of lithography that is used now limits the size of the finished sensor chip to about what we have in the Canon D60, 10D, and Digital Rebel. It is very expensive for Canon or anybody else to manufacture sensor chips that are full 36x24mm size. Canon can do those, but they have to do them in a special fab with special lithography tools, and that makes them very expensive.

Now, if the marketplace demands full frame sensors, Canon will figure out how to make them cheaply. It may take a while, but it can be done.

What does it take to make such a demand? Writing a letter does not work. Signing a petition does not work. Continuing to purchase Canon products will work partially, and every time you see the Canon rep, ask him when full-frame sensors are going to be less expensive. Keep beating on the rep that way, and little by little the word gets to the higher powers. Drop little hints to the Canon rep that you would buy this and that... if only there were a full frame sensor.

I haven't seen a Canon rep since about last March, but I dropped such a note into his coat pocket then.

---Bob Gross---

[rdenney]

belmondo wrote:
Don't count on Canon giving you that all at one time. They'll probably step up to a full-frame sensor a little at a time. That's just good business on their part.

If they do, however, it probably won't be cheap.

(I say all this because I have a well-established track record of being wrong about such things, so predicting it won't happen is almost a guarantee that it will. Let's see if it works.)


---Tom---

My prediction is a full-frame body in the $1200 range in three years. But by that time, a DSLR that is $1200 will be considered pricey. The EOS 1 DLSR of that day will be around $2500--a thousand more than a film-based EOS-1. The digital Rebel will be under $500, or $200 more than the film-based version. The Digital Elan will be in the middle. That's why Canon is only building Rebel-quality small-format lenses, when Nikon is rolling out high-end small-format lenses. Canon, on the other hand, is rolling out new, lower-cost L-series lenses that are a stop slower, but still cover the full frame, to entice higher-end DSLR owners who can clearly see the poor quality of their previous lenses for the first time.

Fortunately, I've got nothing riding on my predictions.

Rick "you heard it here first" Denney

[rdenney]

robertwgross wrote:

Now, if the marketplace demands full frame sensors, Canon will figure out how to make them cheaply. It may take a while, but it can be done.

What does it take to make such a demand?
...


The number of folks shelling out $8000 for a 1Ds is what they'll notice. Plus, people buying expensive L-series lenses for their 10D's--lenses that cover the full frame--will also get their attention.

Judging from what I hear, the 1Ds must be outstripping sales targets, even though only a few will spend that sort of lucre. We know the 10D has been a hit.

But there is also a limit to how high pixel density can get before lenses just can't reasonably keep up. The semiconductor industry may dictate chip development as you say, but the optics industry must be under even greater restriction given the maturity of their technology. The 133 pixel/mm 10D sensor is already at the limit of most lenses, even Canon's high-end lenses. The only place for them to go is to bigger chips.

Rick "recognizing that the 1Ds sensor costs $4000--for now" Denney

[w10d]

Personally, I wish they'd make the sensors a little taller before they make them any wider; the 35mm format is not exactly ideal for many things. i don't really see why camera manufacturers are hell bent on retaining those proportions for DSLRs. Seems a shame to always have to throw away a bunch of pixels for final usage - magazines, A4, 10x8, 6x7, all have similar proportions, 35mm doesn't fit any of them.

Just my 2 cents...

[Belmondo]

robertwgross wrote:Now, if the marketplace demands full frame sensors, Canon will figure out how to make them cheaply. It may take a while, but it can be done.

Isn't it ironic that in this age of miniaturization, we actually want something bigger? It's interesting to note that there are more people clamoring for larger sensors than for higher pixel density.

[robertwgross]

I guess we will keep bitching about it until we get the affordable full frame sensor. Then we will shift gears and demand that we get 30 megapixels instead of only 6 or 10 or 20.

---Bob Gross---

[defordphoto]

belmondo wrote:
robertwgross wrote:Now, if the marketplace demands full frame sensors, Canon will figure out how to make them cheaply. It may take a while, but it can be done.

Isn't it ironic that in this age of miniaturization, we actually want something bigger? It's interesting to note that there are more people clamoring for larger sensors than for higher pixel density.


Yes it is. This is where I think many people are wrong in where digital cameras are going. I think we weill see smaller sensors with much higher density. I think we'll end up with miniature SLR's and that the old '35mm' format will go away. I think we'll end up with smaller sensors than what we are using now. Smaller cameras and smaller lenses giving us 2-3-4-5-6 or more times the quality of what we are shooting now.

Of course that's just my opinion. I could be wrong.

[MiG82]

I think you are wrong because I vaguely remember hearing that already 70% of the incident photons are being converted into current. Remember that at the end of the day you are relying on discrete packets of light for your image. You can't keep miniaturising while retaining sensitivity and accuracy.
Of course, this is only my rather incomplete knowledge speaking. I am in no way or form well versed in this topic.

[defordphoto]

It'll be interesting to se where this all goes. Technology is a marvelous thing. About the time we think, "No. They can't do that. That goes against the rules." Then some new technology comes out and blows away everything we thought was right.

[rdenney]

RFMSports wrote:
[Yes it is. This is where I think many people are wrong in where digital cameras are going. I think we weill see smaller sensors with much higher density. I think we'll end up with miniature SLR's and that the old '35mm' format will go away. I think we'll end up with smaller sensors than what we are using now. Smaller cameras and smaller lenses giving us 2-3-4-5-6 or more times the quality of what we are shooting now.



You may be right. But Canon keeps investing in new full-frame lens designs, and that seems like a lot of investment if there is no real long-term commitment to the format. If they have one format for pros and another for consumers, where will be the line that is between them? My prediction is that the 10D's descendents will go with the pro cameras, because the folks buying the 10D's are the ones buying the lenses.

It's likely that Canon doesn't even know for sure how it will turn out. But I haven't seen a Nikon get the sort of visceral reaction that the 1Ds got. I haven't even seen that reaction for the Olympus, despite their advertising blitz.

Rick "seeing the big manufacturers place their vary large bets" Denney

[fredlord]

The new Sony 8 megapixel digicam uses a sensor 8.8 x 6.6mm in size giving a very small 2.7 µm individual pixel sensor size. If that were to be interpolated up to a full-frame sensor you would have a whopping 119 megapixel array for gathering up your creative light. The problem is, as Mr. Gross says, among other problems, the chips can't be easily built to the FF size because all the existing production facilities are geared to a standard size base disc on which is it impractical to build 36 x 24 mm sensors. Until new prodedures and facilities come on line, there will still be no inexpensive FF sensors. It's only my opinion but I think the Kodak FF sensor camera was only a loss leader attempt by Kodak to stay in the game. They obviously cut every corner possible to produce their camera at a lower price than Canon and, sadly, it shows in the final product. Same deal with the Contax FF SLR only they cut no corners, really.

To sum it up, someone will build a FF sensor for an affordable price eventually. Technology always meets demand so if the demand is for FF, it will happen. The trick is to keep shooting with what you have and move up when you can. In the meanwhile, Sigma has come out with a new wide-angle zoom starting at 12mm on the wide end. I guess we'll have to try and get along with that for a while if we want any semblance of a wide-angle view from our digital SLRs. I do miss my old 17mm FD lens on my A-1s.

[CyberDyneSystems]

I agree that Olympus' direction is the way things will eventually go,. (smaller CCDs with higher density = smaller glass)

But it is too bad that Olympus is screwing it up so bad. The whole reason given for a smaller CCD is to then be able to use more efficient glass,. and yet Olympus is pricing there lenses WAAAAAAAaY out of proportion,. there 300mm lens is priced about the same as Canons 600mm ?????

How does that make any sense?

The only way Olympus would have been able to win anybody over to there system is if it actaully was cheaper,. so far a 10D with "L" glass is a bargain in comparison so the E-1 is pointless.

[Canuck]

Hi!
It is interesting this bit about full frame sensors. I have heard about the 4/3 sensor that Olympus put out and on the top it seems impressive! Look a little further and you have or wonder how it stacks up against the 10D for starters and the D1S on the high end. The other interesting point to make is that (although no expert on the subject, I'd appreciate input) the lenses we use now are made for 35mm cameras and not DSLRs. I recon that was Olympus making that pitch and that was how they created lenses for the digital world and not 35mm. I have to wonder, athough we have seen stunning results from the 10D I wonder if that is kinda hokey? Is there a way to make better lenses for DSLRs? Better still is there a way to up the ante on the sensors themselves? The problem IMHO revolves around one bit...money...call them $$$, call them pounds, whatever. It makes me sad that this kind of crap goes on! I fully understand that they want to make money, but at the risk of alienating the cutomers. I'd like to see Canon and Nikon and the rest go at it, and ultimately reduce prices like mad and make them reveal the technology they are hiding. This is not the ideal world, though. For now I recon we have to live with the 2nd best, what we live in right now.

Cheers from England,
Canuck

[samdring]

Just stirring-up a little but wonder if the market actually wants a full frame sensor!

Any 2 zoom lenses with an overlap create a greater overlap @ 1.6 plus significant addition at the tele end.

17-40 (27-64) plus say 28-85(45-136) give 12mm overlap @ 35mm and 19 with a 10d plus the 85-136.

Do we underrate this flexibility? Am I dismissing wide-angle? Is the digital multiplication really a small part of the digital divide?

[chtgrubbs]

What might really influence Canon is a hint to their Rep that the Kodak 14mp, full-frame camera is looking mighty good to you, even if it does mean dumping your Canon glass and buying Nikon lenses!

[rdenney]

CyberDyneSystems wrote:
I agree that Olympus' direction is the way things will eventually go,. (smaller CCDs with higher density = smaller glass)



What happens when they run into the optical limits of their lenses? Are there better lenses than Canon's L optics on the market? (Answer: not by enough margin to refute my point.) If we can discern the limits of those lenses with the 1Ds, let alone the denser 10D, what will we do with more pixels in a smaller space? We'll still be lens-limited in the sizes of prints we can make.

I don't foresee a return to high-performance primes to get around this problem, either. The zoom lens is here to stay. But improvements in lenses will be incremental.

Thus, pixel density is running into an optical ceiling, and I don't see any way around it.

And discussions over the aspect ratio (3:2 vs. 4:3) are not really germane. It isn't the aspect ratio that is limiting our use of wide-angle. And I sorta like 3:2, and I use it in medium format (6x9) when I'm not using the square format.

And I haven't notice any problem with ligh falloff using a 14mm lens on my 10D.

But what IS nice about the Olympus concept is that if the image processor knows a lot about the lens, it can pull some stunts not possible without that knowledge. For example, perspective correction could be dialed in rather than requiring a shift lens. Or a cheaper lens with barrel distortion (such as a fisheye) could be used to make rectilinear images through in-camera processing.

Of course, there's nothing preventing Canon from doing this, either. Canon EF lenses identify themselves to the camera already, though the camera doesn't do much with the information.

I can foresee cameras with lens profiles that are entered as data, and lenses coming with a disk so their data can be added to the camera's database.

But pretty soon sensors will have to get bigger to have more useful pixels, because the image will be limited by lens quality. The very best lenses resolve perhaps 100 line pairs/mm. No more than 200 pixels/mm are required to record that quality, and the 10D is already at 133. And for most of the image made with a 200 pixel/mm sensor, viewing it at full pixel resolution would be exceedingly unsatisfying.

Rick "who needs more area, not more density" Denney

[CyberDyneSystems]

The aspect ratio should just be square to take full advantage of the lens dimensions,. imagine never having to turn your camera for a portrait... then we can just crop in software.

As far as optical limits,. I realize they exist,. I have read discusion both on CCD/CMOS pixel density limits and of lens limits,.
.. but how close are we to those limits?

Isn't this just another hurtle that will be overcome?

Or are we that close to the physical limits with a 35mm lens,. are they really as small as a lens can be and still look good?

[MiG82]

Why are we talking about rectangles at all? :)
Rectangles make best use of film, but that's not a concern with a single sensor behind the lens.
We could have a circular sensor which makes use of all the light captured by the lens. This would give you the most freedom for cropping.

The only thing I wouldn't like about it is that I can't be bothered cropping every single image.

Is there an electronics issue with reading a funny shaped array (i.e. one that isn't well described by so many rows and columns)

[w10d]

A circular sensor would include all of the edge of the lens - not an area that is usually of the highest quality, so such a camera would need the highest quality lenses, and then you'd be cropping most of the time anyway.

A square sensor makes sense, though square formats have never been popular with the majority of photographers, and that could put manufacturers off.

Perhaps a 'rotating' sensor, like the RZ back? With such a design you could have in-viewfinder crops (like the RZ), with custom functions allowing you to select square (full sensor) 5x4, HDTV, Panoramic, or whatever. With big enough storage you could even have the option to save uncropped RAW files with a crop 'as shot', like WB settings....

[MiG82]

But the corners of a rectangle go into the lower quality region anyway.

The variable aspect ratio could be cool, but I can see a lot of people being unhappy with "wasting" expensive pixels when anything except for 1:1 is chosen.

[Jesper]

RFMSports wrote:
Yes it is. This is where I think many people are wrong in where digital cameras are going. I think we weill see smaller sensors with much higher density. I think we'll end up with miniature SLR's and that the old '35mm' format will go away. I think we'll end up with smaller sensors than what we are using now. Smaller cameras and smaller lenses giving us 2-3-4-5-6 or more times the quality of what we are shooting now.

Of course that's just my opinion. I could be wrong.

At this moment there are a number of disadvantages of small sensors with high pixel densities.

Because the pixels are smaller they have much more noise than large sensors - that's why most digicams with small sensors don't go higher than ISO 400 or 800 and cameras like the EOS 10D are usable even at ISO 1600. Something has to be invented to dramatically improve the S/N ratio of small, high-density sensors.

Second, if you want the same depth-of-field as 35mm, you just need a frame the size of a 35mm frame. Small sensors have a larger DOF at the same f-stop, and that's not what you want if you are for example making portraits (although the extra DOF is nice for macro shots).

[w10d]

mig82 wrote:
But the corners of a rectangle go into the lower quality region anyway.

The variable aspect ratio could be cool, but I can see a lot of people being unhappy with "wasting" expensive pixels when anything except for 1:1 is chosen.

Which is why lens tests always show centre and edge tests, with corner being the weakest spot, especially on cheaper lenses.

With a sensor capturing the centre square you don't have to throw anything away - but the history of film formats shows that many people like to crop in camera. Hence 120 film being so popular in 645 and 67.

(wouldn't a circular sensor be a 'waste' of expensive pixels?)

[Erin]

Just remember that we are talking toys here, 35mm equivalent portable cameras.

If you want nice big file sharp files, just buy a 16 meg Kodak DCS Pro back to fit your Contax, Hassy, or any other modern pro 6x6. Its only 16 grand US now!

96 meg 16 bit files that open up in photoshop to 17'x17' at 240 dpi straight from the camera. Then ress up a little to your hearts content!

Come on guys, wev'e just about reached the practical limit for our dinky toys at 6mp!

[photography By Evangelos]

I am not sure if it will take 3 Years for a full frame chip to hit the market. But I am sure canon will be the first to bring it to market and make it affordable. With the Kodak Pro 14 N's low full fame price I think we will see a canon model in the $3,500 range very very soon or with in the next 12 to 18 months or sooner.

[EXA1a]

Erin wrote:
Just remember that we are talking toys here, 35mm equivalent portable cameras.

If you want nice big file sharp files, just buy a 16 meg Kodak DCS Pro back to fit your Contax, Hassy, or any other modern pro 6x6. Its only 16 grand US now!

96 meg 16 bit files that open up in photoshop to 17'x17' at 240 dpi straight from the camera. Then ress up a little to your hearts content!

Come on guys, wev'e just about reached the practical limit for our dinky toys at 6mp!

Good point!
But... most of us ain't no professionals and don't have a Hassy and don't want to spend a fortune on pro equipment. Just want to get the most out of reasonably priced portable 35mm stuff. I like the discussion of this thread and I've learned a lot.

The square thing really bugs me and I spent a few thoughts on it:

It's common knowledge that the typical 2x3 format of a 35mm camera is a bad compromize anyway... too narrow for landscapes and too wide for anything else.

The film 35mm SLR bodies are constructed in a way that the mirror inbetween the lens and the shutter can swing up without touching the rear lens element and can produce a full frame image on the focussing screen (= in the viewfinder).

A sqare sized sensor of 22.7mm x 22.7mm would easily fit into a full size 35mm body (not into the D30/D60/10D/300D bodies, mirrors, shutter, etc. are too small!).

Taking the current 10D chip as a basis, a square 22.7x22.7 size sensor chip would deliver these pixel numers:

3072x3072 9.4mp at a 1x1 ratio
2633x3072 8.1mp at a 6x7 ratio
2304x3072 7.1mp at a 3x4 ratio
2048x3072 6.3mp at a 2x3 ratio (current 10D/300D)

Such a camera could have a cropping function but most of us would prefer to do the cropping on the computer.
Looks good to me, and the 22.7x22.7 sensor probably could be produced in the current cheapy-chip factories, right?

--Jens--
(10D is still tempting for me D30 owner)

[Belmondo]

It is my belief that the 35mm format (i.e. sensor size) will remain the holy grail of digital photography for many years. Indulge me with an analogy I make using American railroads.

In the United States, and many other parts of the world, the standard gauge for railroad tracks is 4ft. 8 1/2inches. Why? Because someone arbitrarily decided almost two centuries ago that rails should be the same distance apart as the ruts in old Roman roadways. Or at least, so the story goes.

Is that the optimum? Is that what they'd use now if there were no existing railroads and they were starting from scratch? Who knows? No matter, the American railroads aren't even thinking about changing. With hundreds of thousands of cars and locomotives, and tens of thousands of miles of tracks to replace, they will continue to improve and adapt their technology within the constraints of the existing standard, whether optimal or not.

And so it will be with digital SLR's, I expect --- at least into the foreseeable future. Is the 35 X 24mm frame size ideal for digital photography? I suspect not, but it’s here now, well entrenched, and unlikely to change anytime soon. 35mm has been the de facto standard for film photography since World War II, and the SLR has evolved into the preferred vehicle for its utilization. Autofocus SLR systems are relatively new in the grand scheme of things, but they’ve been around long enough for a large number of people, especially professionals, to be heavily invested in them. This includes film bodies and more importantly, lenses.

Although digital photography is a new technology, the digital SLR’s function as currently utilized is nothing more than a replacement for a 35mm film body in an SLR system. Accordingly, People like Canon and Nikon are not going to leave a large, well-established user base in the lurch by abandoning the 35mm format as the standard for their film-based and digital SLR systems. A company like Olympus can possibly afford to do it because, frankly, they have little to lose at this point based on their current share of the professional and high-end consumer markets.

I think we can expect to see sensors in popular-priced SLR’s gradually grow in physical size to equal that of a standard 35mm frame. After that, or perhaps concurrently with that, look for pixel density to continue to increase. I’m sure there’s a theoretical limit to how many pixels they can squeeze onto a sensor of given size, and I’m sure they’ll figure out a way to exceed it. (---just as they’ve done with so many other things that ‘couldn’t be done.’)

Thoughts from an insomniac at 4:00am.

[MiG82]

w10d wrote:
(wouldn't a circular sensor be a 'waste' of expensive pixels?)

Obviously. I left that unsaid because the circle wastes even more

[Longwatcher]

The camera I want is as follows

I want interchangeable lenses - or a morphing lens (way out in the future tech range) - I get this with my 10D.

I want a camera that I can use a fish eye lens on with as well as at least a 1200mm zoom. I get the telephoto end with the 10D, but not the fish eye end.

I want as much resolution in my photographs/images as technologically possible. I reconize this is balanced by cost and file size. The 10D meets my current minimum requirements to be able to produce a 16x24" print without noticing pixels. Currently, in a I won the lottery world, I would like a sensor giving me a 40x60" image at 300DPI native or bigger.

I would like the LCD display (and I would like it to fold out to different angles for my convenience). My 10D being an SLR does not do this.

Currently the 10D fills some, but not all of my desires. Even the 1Ds does not meet all of my desires in a digital camera, but at least it would fill the FF issue so I could get a full range of effects.

No camera meets all of my requirements.

As to sensors. I have been told by image scientists, who should know what they are talking about that you get the the overall best SNR of visible light at about 8um size of the detector (pixel). Thus you will note the high end SLRs have around that size for their detectors.

I am of the opinion that around 25MP you will hit the size limit for a FF, SLR in a 35mm body.

Also remember that camera body size is important.

There seem to be four basic camera body sizes in the digital realm.

Pen Cameras - or how small can we get the sensor and still see a decent shot. (A gadget camera)

P&S Cameras - keep it convienent so you can carry it everywhere, but otherwise get a decent picture out of it in snap shot mode. (A consumer camera)

35mm size - get a highest quality possible, while still making it portable. (note I put the Sony f717/828 in this category), despite pixel size) (A pro/semi-pro/adanced amatuer camera)

Studio size - just plain highest quality possible. (A studio/art camera)

Just my opinion and rambling.

[w10d]

belmondo wrote:
35mm has been the de facto standard for film photography since World War II, and the SLR has evolved into the preferred vehicle for its utilization. Autofocus SLR systems are relatively new in the grand scheme of things, but they’ve been around long enough for a large number of people, especially professionals, to be heavily invested in them. This includes film bodies and more importantly, lenses.


35mm has been the standard for SOME types of photography, mainly where cost and portability are paramount. Landscape/Still life is still often taken on 10x8 and 5x4, fashion/beauty with 120, etc., etc.

35mm format doesn't fit many photo paper sizes (which I think fit your railway analogy better - they really haven't changed since the dawn of photography, and have their roots in art history), or printer paper sizes. There is no technical or aesthetic reason for camera manufacturers must sick to 35mm proportions, but obvious reasons to stick with 35mm equipment.

Will they take the chance to give us sensors that are more likely to fit the image's final useage? Who knows, but they didn't have a problem developing the APS format when they wanted to sell more compact cameras, so I guess it's possible...

[Belmondo]

w10d wrote:

Will they take the chance to give us sensors that are more likely to fit the image's final useage? Who knows, but they didn't have a problem developing the APS format when they wanted to sell more compact cameras, so I guess it's possible...

--- and we know how successful the APS has been.

You raise an interesting point about end use size. I can see some advantage to a sensor with a height to width ration of 4 X 5 as long as it uses the maximum potential area of the standard 35mm lenses. Using my middle school algebra, I come up with a frame size of 33.1 mm X 26.5 mm --- 2.5 mm taller, and just under 2 mm narrower. This would give an aspect ration of 4 to 5 and would use all the lens.

Also, think about this: if the camera had a 33 X 33 sensor, you could switch the camera from landscape to portrait mode just by pushing a button; you wouldn't have to rotate the camera 90 degrees. You could also 'mask' the image in the viewfinder to properly reflect the orientation of the image. Just think about it: no more sticking your elbow up in the air to take a picture in portrait mode. And as EXA1a points out, you could take a square picture and do the cropping yourself. If the camera would have to be a little taller to acommodate the taller sensor, so be it. The overall package would still be much smaller than a 10D with the the Big Ed battery grip on it, even with a second battery designed into it.

Finally, I'm not dismissing other formats as standards within their 'niche.' It's just that the vast majority of cameras produced, sold, and still in play within the last 60 years have been 35mm. Let's not quibble over technicalities.

Tom

[w10d]

belmondo wrote:
you could switch the camera from landscape to portrait mode just by pushing a button; you wouldn't have to rotate the camera 90 degrees. You could also 'mask' the image in the viewfinder to properly reflect the orientation of the image.

Finally, I'm not dismissing other formats as standards within their 'niche.' It's just that the vast majority of cameras produced, sold, and still in play within the last 60 years have been 35mm. Let's not quibble over technicalities.


Yes, the 'rotating without rotating' idea was what I was suggesting earlier (like the RZ backs) - I guess this thread has got a bit too long :)

OK, I won't quibble, I think we're sort of agreeing anyway!

[Belmondo]

w10d wrote:Yes, the 'rotating without rotating' idea was what I was suggesting earlier (like the RZ backs) - I guess this thread has got a bit too long :)

OK, I won't quibble, I think we're sort of agreeing anyway!

Amen. (BTW the thread isn't too long -- my attention span is too short.)

Tom

[rdenney]

CyberDyneSystems wrote:
The aspect ratio should just be square to take full advantage of the lens dimensions,. imagine never having to turn your camera for a portrait... then we can just crop in software.

As far as optical limits,. I realize they exist,. I have read discusion both on CCD/CMOS pixel density limits and of lens limits,.
.. but how close are we to those limits?

Isn't this just another hurtle that will be overcome?

Or are we that close to the physical limits with a 35mm lens,. are they really as small as a lens can be and still look good?

Yeah, square format--that's the ticket! (I do lots and lots of serious work on 6x6 cameras, but they seem to also be a diminishing breed.)

Pro-quality lenses typically have useful resolution in the 60-100 lines/mm range. My high-quality medium-format lenses are in the 40-60 range.

The 10D sensor can delineate resolution in the 50-70 range, depending on orientation, at least theoretically.

But read all those lens reviews on Luminous Landscape. That fellow can tell us that the 16-35 L zoom is a little better than the 17-35 L zoom based on test shots on a 1Ds, which has lower pixel density than the 10D.

Thus, I think more pixel density will just allow us to see our lens faults more clearly. I think we are pretty close to where more pixel density will not mean more useful detail in our photographs. We would do nearly as well (or perhaps even just as well) by resampling the image to a higher resolution. And it may be that software that simulates edges like Genuine Fractals might actually work better than the same pixel density on the sensor.

The best 35mm lenses are a little better than the best medium-format lenses because the MF lenses would be prohibitively expensive if made to the same standard. Even the high-end Zeiss glass in Photodo has ratings in the high 3's, rather in the middle 4's like the best fixed primes for 35mm cameras. Considering their cost, I suspect that we will be seeing only marginal improvements in the future, though we will see faster lenses and wider zoom ranges that rise to that level of image quality.

Rick "who thinks that the best current sensors are already leaving us limited by our lenses" Denney

[agit-prop]

RFMSports wrote:
belmondo wrote:
robertwgross wrote:Now, if the marketplace demands full frame sensors, Canon will figure out how to make them cheaply. It may take a while, but it can be done.

Isn't it ironic that in this age of miniaturization, we actually want something bigger? It's interesting to note that there are more people clamoring for larger sensors than for higher pixel density.


Yes it is. This is where I think many people are wrong in where digital cameras are going. I think we weill see smaller sensors with much higher density. I think we'll end up with miniature SLR's and that the old '35mm' format will go away. I think we'll end up with smaller sensors than what we are using now. Smaller cameras and smaller lenses giving us 2-3-4-5-6 or more times the quality of what we are shooting now.

Of course that's just my opinion. I could be wrong.

Anyone remember the old Pentax Auto 110 SLR system?

http://www.submin.com/110/collection/pentax110/cameras/cameras/auto110super_2575351.jpg

taken from http://www.submin.com/110/collection/pentax110/cameras/super.htm

A 110 negative is close to the size of most CCD sensors. Perhaps Pentax was 25 years ahead of thier time with that chassis and mount system...

Hmmm. I *do* have an old auto 110 kicking around here. Maybe I should look for a doner digicam and build a digital back.