CNet Tracks the History of the Digital Camera

Abby Donivosif writes "CNet has up an article about the history of the digital camera. It's fascinating to note how far the technology has come in such a short amount of time. 'The camera generally recognized as the first digital still snapper was a prototype developed by Eastman Kodak engineer Steven Sasson in 1975. He cobbled together some Motorola parts with a Kodak movie-camera lens and some newly invented Fairchild CCD electronic sensors. The resulting camera, pictured above on its first trip to Europe recently, was the size of a large toaster and weighed nearly 4kg. Black-and-white images were captured on a digital cassette tape, and viewing them required Sasson and his colleagues to develop a special screen.'"

Comments

Nostalgic?

[Anonymous Coward]

Yes, I recall the days of 320x240 and 640x480. Great times I'm sure.

As a digital photographer, I've come to appreciate the people behind the physical camera. Both technological and artistic.

As for future cameras, I think we'll see initially, 3x sensors allowing for on the fly HDR images. After that we'll go to static video where a framed shot can be spun around to see all the out of frame info.

After that, I suppose we'll get selective depth of field, on the fly image editing, blemish correction and on the fly multi-image splicing allowing for a static family photo to be created via sliced video.

Of course we'll have meta data including temperature, GPS, wind speed, angle, height, surrounding buildings, photographer's personal ID#, satellite upload, etc.

Film will die in the same way that pinhole cameras are dead. Sure, it's around and you can use it but what's the point? The medium isn't the art. It's the person behind the camera.

Re:Nostalgic?

[adolf]

I doubt we'll see 3x sensors for automagic HDR images in anything but very specific (ie, expensive) applications, for a few reasons:

- Cost. The current and obvious trend, at least toward the low end of digital photography, is to reduce sensor size as much as possible in order to manufacture more CCDs with a single wafer. 3 times as many sensors means 3 times as much cost. I suppose one could put all 3 CCDs onto a single die, making the whole package a bit smaller, but I doubt it'd help much with the final cost.

- Rarity. Because most people don't care about HDR photos for their routine shapshots, they technology will have no economy of scale boost to cover the R&D on the manufacturing side for the short runs it would entail.

- Lighting. Split the optical path into 3 equal parts so that you can expose 3 sensors at the same time, and even in a perfect world each of them will get one-third as much light. Without some very dramatic increases in signal-to-noise ratio over what is common now, this isn't likely to be very practical in most uses. (In reality, you'd want them graduated such that one sensor gets overexposed, another gets normally exposed, and the third gets underexposed. But this still means less light available, even for the "normal exposure" sensor.)

One could, I suppose, scan a mirror between the three sensors and cure the light problem by using each one in series. But then, you've got time distortion between exposures, geomatric distortion, focal plane issues, and a bunch of extra jittery mechanical parts (which SLRs have plenty enough of, already) -- solvable problems, sure, but unnecessary ones. And it's not at all clear to me how such a scanning arrangement might ever be better or faster than using a single CCD, and having the camera's software take a quick series of three shots at different exposures.

No, sir. I just don't think it will happen. Better, simpler, cheaper, and far more available results would be produced by improving the dynamic range of conventional single-CCD cameras.

Re:Nostalgic?

[fyngyrz]

Too late. It's already happened. Same location on chip, so the same sensor size, essentially three sensors at three different depths. Sigma SD14, for instance. Price is right in the prosumer zone.

its not really photography

[waterwingz]

Wasn't it the board of directors of Kodak who decide to not go the digital route, summing it up with the statement "If it doesn't contain silver halide, its not really photography" ?

Reverse DLP

[inKubus]

What about the future of the digital camera? The CCD is reaching the end of it's useable life. They are just packing more and more pixels in, when really what you need is more levels of greyscale and a better signal to noise ratio. I'm wondering when they'll get rid of CCD entirely and move to a 4 "pixel" sensor with a DLP chip in between handling the scanning, instead of a bunch of piddly pixels on a 1/3" ccd. The sensors could be larger, with focusing lenses in between. The color isolation would be perfect. Plus you could use variable filtering/exposure PER COLOR based on the ambient light to do true (not digital enhancement after capture) white balancing. There's no reason a DLP couldn't work in reverse, I don't think. Other possibilities include nanotubes "tuned" to certain visible frequencies that cause them to vibrate slightly, etc.

There's also the liquid lenses such as Varioptic, which are going to change what we know about photography. Coupled with GIS/GPS I think we're in for a great next century.

Re:Reverse DLP

[wickerprints]

The problem with using a "reverse DLP" mechanism for light capture is that it's just not possible to scan several millions of pixels at higher shutter speeds. The method neither scales well with resolution nor time, unlike existing CMOS/CCD technologies which does scale very well with time, and reasonably well with resolution. Even worse, there is a more fundamental problem--the incoming photons through the aperture enter at different angles and energies. How do you properly distinguish them with a sufficiently precise moving object, when the camera is handheld? Eventually, one reaches a limit where the ability to trigger a sensing element falls below the noise threshold. Simply increasing the sensor size does not completely resolve this issue, although it does mitigate the need for extremely sensitive pixels. I forsee the short-term future of digital light capture to go in the direction of improvements in sensor design and optics, perhaps even implementing layered sensors (higher frequency photons penetrate more layers, thereby resulting in better frequency capture), rather than the commonly-used Bayer filters that invariably result in information loss. Long-term, I see things like capturing phase information, leading to natural light, full-color holography, thereby rendering photography obsolete.

Re:Reverse DLP

[femtobyte]

DLP works because you can fool a human eye into seeing a quickly-scanned point of light as a continuous image. Freeze a point in time, and the DLP is projecting a single bright point of light focused in one place. It's not practical to reverse this for taking photos, since the light coming from real-life objects that you want to photograph is usually coming from all points on the object at once. While you are scanning around the object looking at one small point at a time, you are wasting all the photons coming from the rest of the object. To most efficiently use the light reflecting from a photographic subject, you want a sensor that can simultaneously record the light coming from all points on the subject at once --- e.g. a CCD sensor at the focal plane of the camera, just like we have now.

The idea of a scanning, DLP-like camera isn't without merit. One could imagine a futuristic replacement for flash photography where a set of colored lasers are scanned across the subject at high speed, and micromirrors guide the reflected spot of light onto a single-pixel sensor. This way you can record both color reflectance information and 3D topology using a much more simple and compact sensor system than current CCD cameras. But in order for this to work, you have to be providing the illuminating light yourself, making sure that it is focused on the one small spot that you are currently scanning. Such a camera would be horribly inefficient for photographing any "naturally" lit scene where the illumination is spread over the entire subject (e.g. a sunset or a diffusely lit portrait).

first picture?

[jmcnaught]

It's too bad they didn't include the first digital picture, that would have been neat to see. I couldn't find it on google, but I didn't really spend that long looking.
Hopefully they still have it kicking around somewhere. The comments in the CNET article suggest they know what the picture was of but I guess they couldn't find it either.

Re:first picture?

[matfud]

One of the first (and still the most commonly used image in image processing) is actaully porn

http://en.wikipedia.org/wiki/Lenna

matfud

Wait! Kodak?

[blind biker]

(by the way, excellent article and photos, really enjoyed it!)

So, you're saying Kodak had the first digital camera in their house (and later, they produced Apple's digital cameras - read the article, you'll see..), and Kodak is today in commercial difficulties because their film business is failing - because of digital cameras' success?

While I have the greatest admiration for Kkodak's engineers and workers, to Kodak as a company I have to say: WHAT WERE YOU THINKING???

History of the camera in life

[Eccles]

I put together a "slide" show recently of my son's life for a family event. We had few, grainy pictures from his younger days, and lots of high quality pics from more recent times. (Didn't have enough time to scan film photos.) It's like the Calvin and Hobbes where Calvin's dad claimed the world was black and white when he was younger, then got grainy color and then finally high quality color around the time Calvin was born.

Re:Micron DRAM chip

[fyngyrz]

I still have a couple SWTPC KC tape controllers, they both still work - I had kind of a old computer fest here a few years back, wrote 6809, z80 and 6800 emulations, gathered up all my old software and so forth. Was interesting. I was able to recover every tape I'd made; I thought the oxide would fall off, but no, they played back fine. I even read back a paper tape of BASIC; now that was a bit of a flashback. I keep the paper tape in a sealed can. It's some kind of oiled paper, holding up very well indeed.