When Canon came out with the EOS-D30, I said "Whatever, a digital still camera will never look good because it has a single CCD -- or maybe it will look good when it has 20 million pixels which can be downsampled to look good."

Well, look where we are now with digital still cameras. The EOS 10D, 20D, etc -- look amazing with their single-CMOS. Would we not be happy if it captured motion at 60 frames per second? Would we complain the color is bad, or the noise is too high? No way! But 5 years ago I would not have thought this possible.

I predict we will soon see a single CCD camcorder, perhaps from Canon, which has a single CCD and produces EOS quality images. So if you hear of such an annoucenment and it is from Canon, don't be too quick to snob it like people did to the JVC.

I'm inclined to agree with you, Robert, and within the realm of standard-definition camcorders this is almost a reality already. Any single-chip camcorder equipped with an RGB primary color filter mask will closely approximate the color accuracy of a three-chip camcorder. For example, any of the current Canon Optura series camcorders (which have an RGB filter) will match the look of the older three-chip Canon XL1 and GL1 camcorders under the proper lighting conditions. I think it's just a matter of time before single-CCD performance and color accuracy improves to the point where three CCD's are no longer necessary. The JVC HD10U is a step in that direction already.

Also the manufacturer does not need to worry about CCD alignment on the optical block. I think we will see large single CCDs that actually work well.

I am happy with my OpturaXi (that is, until I started playing the the HDR-FX1).

I'd like to see Foveon CCDs integrated with camcorders, but that may be overkill. ;)

Actually it is underkill.

Canon's cheapest single-CCD digital SLR has better quality.

Foveon's idea is great, but Canon is so advanced they made a single CCD which is better.

Nice to know. I not keeping up-to-date with all advancements in CCD technology, but I remember the hooplas over Foveon, perhaps 2 years ago?

No way to clear the CMOS gates fast enough for 30fps with either Foveon or Canon. If it's CMOS in a video camera, look to Rockwell.

Chris beat me to it. CMOS does not equal CCD. You can bet we'd be seeing much cheaper, higher quality HDV cameras with CMOS chips if they could be easily adapted for video. They can't, at least not right now.

Well Canon makes good CCDs also. They use them in their higher-end EOSs.

Actually no. Canon does not make CCD's at all. Now their high-end digital SLR's use a Canon CMOS chip. However, for every Canon camera or camcorder that incorporates a CCD, it is out-sourced from some other supplier. Canon missed the boat by not getting into the CCD business and it's way too late to get into it now. Canon makes a mask aligner -- the machine that makes a CCD -- but Canon does not make CCD's, go figure.

Let me rephrase that to say that Canon uses CCDs which provide amazing images in spite of just being a single CCD. I think we all agree that if a Canon EOS1v or whatever the good one is was shooting at 60fps, we would consider the output good and would forget about 3-CCD solutions.

<<<-- Originally posted by Robert Silvers : Actually it is underkill.

Canon's cheapest single-CCD digital SLR has better quality.

Foveon's idea is great, but Canon is so advanced they made a single CCD which is better. -->>>

Actually, Foveon chips are as yet unsuitable for HD video cameras.

There are two models of Foveon sensors, the X3 4.5MP (a.k.a. Foveon FO18-50-F19) and the X3 10.2MP (a.k.a. Foveon X3 Pro 10M and Foveon F7X3-C9110). The 4.5MP, has a 4:3 aspect ratio and, you guessed it, about 4.5 million effective pixels (1,420 x 1,060 x 3 layers [R, G, B]) over an effective area of 7.1 mm x 5.3 mm (effective diagonal, 8.8 mm). The chip can yield this resolution, which wouldn't be adequate for 1080i or 1080p (especially the full 1,920 x 1,080), even at 7 fps; at 30 fps, resolution drops to 640 x 480. Also, greater light sensitivity can be achieved, but only by ganging pixels together, effectively reducing resolution at a precipitous rate.

The 10.2MP has a 3:2 aspect ratio and about 10.2 million effective pixels (2,268 x 1,512 x 3 layers) over an effective area of 20.7 mm x 13.8 mm (effective diagonal, 25 mm). This resolution is more than adequate for 1080i or 1080p, but it can be sustained at a maximum of only 4.4 fps; at 25 fps, resolution drops to 576 x 384, which, oddly, is a lower frame rate/resolution combination than that of the 4.5MP. In addition, this sensor is probably too large (about 1 inch) even for "conventional" professional HDV camcorders (1/2 to 2/3 inch is about the biggest they'll likely get), although, as it's approximately the same size as the aperature in a 35-mm motion-picture camera, it may be perfect for some of the new "digital cinema" cameras, such as the Kinetta. Unfortunately, as with the 4.5MP, increased light sensitivity is achieved at the cost of resolution.

When the 10.2MP can run at full resolution at 60p (a frame rate said to be supported by the rumored second-generation extension to the HDV spec) and 3 lux or so, then it will be ready for HD/HDV prime time.

Thanks for such an informative post, Lawrence. Nice to see you here, by the way!

If we could just get freakin 24p I'd be happy!

I will take 60p as my first choice for documenting things, and 30p for drama, then 24p last or when it is just needed to go to film (come on, how many people really do that).

Oklahoma was shot in 65mm at 30fps in 1955 and it was considered to be a good thing.

I want better than film quality. People are starting to see this with digital still cameras and won't even WANT to go back to 35mm film. In a few years video people will stop thinking of video as a cheap alternative to film and actually will prefer it.

<<<-- Originally posted by Robert Silvers : In a few years video people will stop thinking of video as a cheap alternative to film and actually will prefer it. -->>>

You know, I'd be happy if more people would simply quit thinking of film as **THE ONLY WAY TO DO IT RIGHT, HARUMPH, HARUMPH** and simply admit that video is a viable option and has it's merits. I think in the past couple of years we've come a long way, but there's still those few out there that tell us videophiles--"You'd all shoot film if you had the choice and/or money...". The only thing that's gonna change those attitudes is new technology I guess. And while HDV isn't the end all, be all--it's a good start in the right direction. I think trying to take something that has been ridiculously expensive and make it affordable to almost everyone will spawn new innovations and hopefully some really amazing breakthroughs.

Sorry, rant done

Yeah hopefully, but film still have leg up on digital video is exposure lattitude. Plus you can shoot IMAX which is like equivalent 8000x6000 pixels in the digital world. Japan already developing a Ultra High Definition video standard that will compete with IMAX, but you wont see for at least another 20 years.

The next major advancement is not going to be in resolution - who cares if you can project 8000 x 6000 pixels? The next advancement will be 3D moving holograms.

My friend invented an overlapping projector system which can seamlessly project more than 8000x6000 pixels. It uses a video camera to adjust itself to reduce distortion and seams.

The next big advancement will be me getting laid! :~)

Film and video are tools, each with their own unique qualities. I am an old school film guy and I still prefer film. If I had my drothers when shooting video I'd shoot HD, or DVCPRO50, or some format that doesn't toss out half of the information, like MiniDV. Yet, I did spend a lot of cash on an XL2 package, to help me get jobs and maybe the other thing too:~).

Arri, Panavision and Dalsa all have working digicams. Arri and Panavision are well positioned to dominate, since they own the lions share of the 35mm camera rental/usage market. I believe Arri already has theirs working on a film or two. Panavision says rental in December. Gonna have to go down there and see...

The Arri D-20, oh yeah, it's a CMOS chip. Super 35 aperture gate- not 2/3", takes any Arri PL mount 35mm cine lens, optical viewfinder, variable eShutter, ramping, speeds from 1 to 60FPS. 150 if you smile purty. It can also output raw Bayer data for "film" mode. 4:2:2, or 4:4:4 output. Definite panty magnet.

Go to: arri.com > Products > New Products.

Panavision's Genesis is a joint venture with Sony and will be manufactured in Woodland Hills, CA. Has many similar features of the D-20 with a 1/3 stop increase in sensitivity- ISO 400, to Arri's 300, and capable of "pushing" to 1600 (heh, isn't that the "gain" switch?). "12.4 Mega pixel, true RGB sensor, not Bayer pattern" . Full compatibility with the Panavision system. Looks like it would be pretty nice on a Steadicam. No tools needed to go from full studio mode to handheld, or Steadicam mode, size and ergonomics designed for it. Takes Primo primes and zooms, she won't do anamorphic, though.

http://www.panavision.com/product_detail.php?maincat=1&cat=36&id=338&node=c0,c202,c203


Pretty nifty, I'm too lazy to go to the Dalsa site, but check it out, their rig is pretty sweet too.

May the best marketing team win...(let's hope not)

While I agree that we're just about at the point that 3CCD's aren't essential, I think it'll be hard to ween people off of them.

I know that visible results could sell just about any kind of video technology, but do you guys think that users would jump on the single CCD wagon quickly, or would it take a while for everyone to stop equating "3CCD" with "Professional" video anymore?

Actually for professional video 3 CCDs will be better for quite a long time because they have better light sensitivity.

If you build a 1CCD camera and a 3CCD camera using the same CCD chips and put the same lens in front of the cameras, in the case of the 1CCD camera the color filter will block about two third of the light that would go onto the sensors on the CCD. That means that the 3CCD camera will give the same noise and color clarity at about one third the light amount that is required for the 1 CCD camera to give the same noise and color clarity.

This is such a big difference that I don't think it is probable that any professional camera will switch to 1 CCD systems.

Except of course if for example Foveon can decrease the noise from its senser to the same level where CCDs or CMOSes are.

The Arri D-20 is a single CMOS chip and has an equivalent ISO of 300, so there goes your argument.

It is also way bigger than a 2/3 inch chip. It's image area is 24.9mm x 18.15mm. Whereas 2/3 inch chips are 9.4mm x 5.3mm. Pixel pitch for 2/3" is 4.9 x 4.9 microns- 1920 x 1080 pixels. The Arriflex pixel pitch is 8.25 x 8.25 microns- 3018 x 2200 pixels.

The Arri will take any PL mount cine lens. So you can use really sweet glass like the Zeiss Super Speeds, or Cooke S4's.

Get's ya hot don't it?

<<<-- Originally posted by Mark Sasahara : The Arri D-20 is a single CMOS chip and has an equivalent ISO of 300, so there goes your argument.-->>>

My mistake, I meant professional broadcast cameras will not switch. With the exception of the Lockheed-Martin camera all the ultra megapixel movie cameras use 1 CCD. But even the Arri D-20 could be more sensitive using 3 CMOS chips. Unfortunately then it could not use the existing lenses. When digital movie cameras will prevail, then they may go into the pain of making big prism blocks and special 3CCD lenses to increase sensitivity.

I don't know that much about chips, or how they achieve their sensitivity, but why bank on one specific technology? I'm for whatever works, whether it's a single chip, or a chip block, whatever. Things are so up in the air, the technology is changing almost daily.

Besides, I want the option of high sensitivity chips for low light, as well as lower sensitivity for outdoors and high illuminaton situations. With no noise, of course.

It boils down to cost effectiveness. If it is too expensive and complicated to mass produce CMOS chips then, they'll stick with what they've got. Sony is working with Panavision on Panavision's new Genesis camera. Whatever they get from that venture will make it's way into their products.

I don't necessarily want the D-20 to be more sensitive. If I have to shoot outdoors, I don't want to be at T/16 and then have to dump a whole bunch of ND in front of the lens. I want to be at T/2.8, or thereabouts. ISO 300 is fine. If I am shooting Docco and low light situations, yes I would want to have a camera available that can do that and give me lovely pictures.

In a production environment I'll be using lighting, so I don't necessarily need a super sensitive camera.

<<<-- Originally posted by Balazs Rozsa :
. . . This is such a big difference that I don't think it is probable that any professional camera will switch to 1 CCD systems.

Except of course if for example Foveon can decrease the noise from its senser to the same level where CCDs or CMOSes are. -->>>

There are additional reasons that Foveon's chips aren't ready for HD video (see my Dec 15 post). Chief among them are (1) at sufficiently high resolutions, frame rates are too low, and (2) to increase light sensitivity, pixels are ganged together, reducing resolution.

Why are you guys fixated on Foveon? Foveon is for stills only.

Have you been reading what I've written, or followed the links I've provided? The next generation of "Digital Film" cameras all use one chip sensors and have frame rates of:
(in order of top speeds)
Dalsa Origin 36 FPS
Panavision Genesis 50 FPS
Arriflex D-20 60 FPS
And obviously 24, 29.97, 30, etc.

I'm not sure who the chip manufacturers are, but the cameras work and are getting ready for the marketplace. The Dalsa Origin is also a one chipper and they are the manufacturer of their own chip.
http://www.dalsa.com/dc/dc.asp

I don't think everything is going to be one chip, but I'm no expert either. We'll probably have three chip CCD's for a while, until the technology matures and costs come down.

As the old Vermont saying goes: "Hard tellin', not knowin'."

Sorry, the links are not dynamic.

It strictly depends on what the actual cam can do.

The consumer driven market place has got their fingers in the "3chip" catch phrase. So it is hard now to promote a 1 CCD. JVC brought 1chip back into recognition but market focus has swung back to the 3CCD chant when the FX1 came out.
A major player will have to come out with a $20,000 high end cam with 1CCD to confuse everyone with their $4000 prosumer 3CCD cams so this CCD counting nonsense is forgotten once and for all.

Not easy to do so. Current broadcast lenses are designed for a prism block. You would need to come up with a whole family of new lenses designed for your 1 chipper. Plus your new camera would not be as sensitive to light as its 3 CCD counterpart. Low light sensitivity is one major difference between a 1/4" consumer camera and a 2/3" professional broadcast camera.

<<<-- Originally posted by Mark Sasahara : Why are you guys fixated on Foveon? Foveon is for stills only. -->>>

It's pretty clear from my posts that I'm not fixated on Foveon chips, and I as much said they were unsuitable for HD video.

<<<-- Have you been reading what I've written, or followed the links I've provided? The next generation of "Digital Film" cameras all use one chip sensors . . . -->>>

I have to disagree with you on this one. While it's true that the cameras you mentioned are one-chippers, the notion of "digital film" must apply to all forms of high(er)-definition digitally originated film, not just the high (and extremely expensive) end. Many of the forthcoming HDV cameras, as well as the new Panasonic P2 and DVCPRO-HD models, are sure to be three-chip. Ironically, even though using a single chip is simpler and should be cheaper, chips with sufficient resolution to be used singly seem to be more expensive than comparable three-chip configurations.

<<<-- . . . and have frame rates of:
(in order of top speeds)
Dalsa Origin 36 FPS
Panavision Genesis 50 FPS
Arriflex D-20 60 FPS
And obviously 24, 29.97, 30, etc. -->>>

What, no Kinetta?

I not fixated on Foveon too. I just mentioned it without knowiing all the facts. But this thread still prove to be very useful in understanding greater knowledge of CCD.

<<<-- Originally posted by Balazs Rozsa : Not easy to do so. Current broadcast lenses are designed for a prism block. You would need to come up with a whole family of new lenses designed for your 1 chipper. -->>>

You could always use a lens mount that accepted 35-mm still or 16- or 35-mm motion picture lenses. Or you could adapt the optics used in the 10x and 12x zoom lenses found in some more expensive digicams. Would they be sharp enough for HD video? Probably. One photography publication ran resolution tests on 3.2-, 4-, and 5-megapixel cameras with these zooms; if memory serves, resolution peaked at about 1,500 line pairs, which is pretty close to the theoretical limit of a 4-megapixel single-chip digicam.

> You would need to come up with a whole family of new lenses
> designed for your 1 chipper.

Yes. But that's really a non-issue: especially if they use the 35mm frame size or similar. Canon has no trouble making lenses for their XL-series DV cams. And for a 35mm frame size they already have a wide range of photo lenses that can easlily be adapted. In a worst-case scenario --that is in the case of a new frame size-- it would just create a new market opportunity. It did for Canon.

And, because of the light sensitivity problem that they are already encountering with HDV, they will just have to move consumer and "indie" video to larger sensors at reasonable pricing. This is because higher resolution means lower pixel sizes at a given sensor size, and lower pixel size means less sensitivity.

The Sony FX1/Z1 HDV cameras don't even have full HD resolution yet and they are already much less sensitive to light than the Canon XL1 and the Panasonic DVX100.

My guess is that Sony, being able to do a much better job with a single sensor and MPEG2 than JVC, didn't want to confuse the market just yet. Sony already uses single-sensor RGB filter cameras in their high-end consumer segment with very good results. They just don't want al the Betacam customers to stop paying all the $$$ and buy the beter prosumer stuff!

Ha, saying 'digital film' is like saying 'horseless carriage.' I really can't wait for film to be dead for motion capture like it is for small format still cameras.

>>You would need to come up with a whole family of new lenses designed for your 1 chipper.

>You could always use a lens mount that accepted 35-mm still or 16- or 35-mm motion picture lenses.

I was talking about broadcast cameras because the camera in question was a $20000 camera.

If you are talking about consumer cameras I agree one chippers have a bright future. With only 2 megapixels 1 chip cameras could produce very nice 1 megapixel downsampled images. If you used a 1.8" hard drive to record 100Mb/s data from the camera at maximum 60 progressive frames a second, you would have an inexpensive Varicam.

The only problem would be you would need a computer to shrink the amount of data to a smaller size before arhiving your footage to DVDs. For example in the form of WMV9 files. But technically speaking such a camera would be quite inexpensive to produce (if high enough numbers could be sold) and still would give you quality images.

Kodak or Olympus could do it because they have nothing to loose in the high end video market

> The only problem would be you would need a
> computer to shrink the amount of data to a
> smaller size before arhiving your footage

Embedded soon enough. Who would have thought 3 years ago that Sony could pull off real-time high-res MPEG2 compression in something you can hold in your hand and pay with three month's worth of salary. So the computer is just a temporary nuisance.

> Kodak or Olympus could do it because they have
> nothing to loose in the high end video market

I agree and I would add Nikon to that list. Furthermore, I think JVC and Canon also have little to loose. And Sony and Panasonic will pretty soon have everything to loose if they don't switch their semi-pro and consumer line to HD and large sensors. It's better for them to self-cannibalize but at least keep the customers at less money then to let those other companies take away all their prosumer clients AND their pro clients.

It's like when the internet came along. The small companies that were able to adapt fast are the ones that were able to ride the wave and survive. Big companies like IBM and even not-so-big like Microsoft almost missed it. IBM actually lost it's practical monopoly on computing and had to become a niche player.

I dont think, 3ccd systems will ever go away entirely

beacuse the way i see it is unless you find an alternative way to sample full colour information (1:1 pixel ratio) then you have to stick to 3ccd.

yes those full 35mm cmos/ccd images are excellent BUT who's saying that those images would not be even bettter if their was 3X 35mm full cmos/ccd..

problem is data..(or too much, to be precise) its hard enough capturing data from ONE 35mm sensor (connected to hard drive the size of a fridge ;)) things will get better in the future..

I forgot about Kinetta, Don't know a lot about it. Not sure how close Jeff is to releasing a production model. I hope it's soon.

My reference to "digital film"(it's in quotes for a reason, punkin') is that Arriflex and Panavision's cameras are" film style" digital cameras. They are set up like and look like film cameras, they have optical viewing systems and use many of their respective film system parts and accessories, plus those of third party mfrs.

This makes the transition for film crews going from film to digital easier. There is a lot of money riding on these folks to get it right the first time, or else the producer is unhappy because the crew is "wasting money". There is a lot more going on on a film set than just shooting a film. So, user interface and user friendliness are paramount.

These people have a lot invested in Arri and Panavision film systems, both in knowledge and equipment that interfaces with these two systems-lenses, follow focus, remote focus, matte boxes, etc. It just makes sense to make them as similar as possible, that way the transition is smoother.

My main interest in a digial camera is as a professional DP. These are tools that I use on a daily basis for my work, that are (hopefully) viable technological steps forward. There are a lot of good things out there, but I take the wait and see approach. I look at what is going on and filter out a lot of the BS and see what will fit my needs and keep track of which way the film and video industries are going. unfortunately the best and the brightest are not always recognized or utilized; or you get the VHS vs Beta argument and whoever has the better marketing department wins.

For many jobs I am renting a camera because neither I nor the production can afford to own the higher end gear. For me as a DP, owning a film camera is one thing, the technology won't really change that much, it'll still take film and expose it twenty years later. Owning an HD camera seems risky, since the technology seems to change all the time. Plus I'm not that rich, yet ;~)

My answer to the original poster's statement of: "I don't believe 3 CCD's are needed anymore" is agreement, apparently not.

<<<-- Originally posted by Mark Sasahara : For me as a DP, owning a film camera is one thing, the technology won't really change that much, it'll still take film and expose it twenty years later. Owning an HD camera seems risky, since the technology seems to change all the time. -->>>

True enough about the HD cameras. As far as the film camera, well, I own a super-modified Arri 2c and I can't rid of it fast enough! (it's been in the shop for a year having a new video door built for it, which is taking forever). I know a lot of cameramen who have dumped their film cameras in the last couple of years; do you know Rob Draper, for instance?

Hi Charles! I like your work.

I know the name, but I don't think I've met Rob.

I think there will eventually be a technological stasis, my main point being that unless you have the bucks to keep re-investing in new video gear every few years, then, you're better off renting. I'm not getting that kind of work yet, but I'm definitely working on it, though.

<<<-- Originally posted by Balazs Rozsa : Actually for professional video 3 CCDs will be better for quite a long time because they have better light sensitivity.

If you build a 1CCD camera and a 3CCD camera using the same CCD chips and put the same lens in front of the cameras, in the case of the 1CCD camera the color filter will block about two third of the light that would go onto the sensors on the CCD. That means that the 3CCD camera will give the same noise and color clarity at about one third the light amount that is required for the 1 CCD camera to give the same noise and color clarity.
>>>

One factor people seem to be forgetting is the accuracy of determining the color of the object. Each resolved pixel of the final image will be defined as components of Red Green Blue light. How to determine how much of each RGB?

3-CCD - If you have a prism you can accurately split the light into beams of each color and put a CCD to capture each beam. Remembering high school physics class the prism does an excellent job of this. Light is split exactly by its frequency/wavelength. The CCD that captures Green for instance will not require the same dense filter to remove the other colors since the prism has already split the light beam and redirected the other colors to other CCDs.

1-CCD - If you have only one CCD you must paint a tiny color filter in front of each pixel to perform the filtering. So the cluster of RGB pixels are all together, little blobs of color allowing the 3 sensors to be sensitive to the different colors. but you are depending on a less accurate method to split the light into RGB. Also you have reduced the sensitivity to low light situations since you have essentially thrown away 2/3 of the light by absorbing it into the painted filters instead of spliting it up and redirecting it to the 3 CCDs of the other solution.

Just my 2-cents...

Paul Shard
Producer - Distant Shores

<<<-- Originally posted by Paul Shard:
3-CCD - If you have a prism you can accurately split the light into beams of each color and put a CCD to capture each beam. Remembering high school physics class the prism does an excellent job of this. Light is split exactly by its frequency/wavelength. The CCD that captures Green for instance will not require the same dense filter to remove the other colors since the prism has already split the light beam and redirected the other colors to other CCDs.

1-CCD - If you have only one CCD you must paint a tiny color filter in front of each pixel to perform the filtering. So the cluster of RGB pixels are all together, little blobs of color allowing the 3 sensors to be sensitive to the different colors. but you are depending on a less accurate method to split the light into RGB. Also you have reduced the sensitivity to low light situations since you have essentially thrown away 2/3 of the light by absorbing it into the painted filters instead of spliting it up and redirecting it to the 3 CCDs of the other solution. -->>>

Example: white light coming in.
3ccd: 1/3 of the light is green and that gets into green ccd.
1ccd: 2/3 of the light is filtered from green pixel, so 1/3 of the light gets into green pixel.
Where's the difference just considering sensivity?

<<<-- Originally posted by Toke Lahti :
Example: white light coming in.
3ccd: 1/3 of the light is green and that gets into green ccd.
1ccd: 2/3 of the light is filtered from green pixel, so 1/3 of the light gets into green pixel.
Where's the difference just considering sensivity? -->>>

1ccd: 2/3 of the light is filtered from green pixel, so 1/3 of the light gets into green pixel. Plus for every other pixel which is not a green pixel the green light is not getting to the pixel at all. Overall 1/6 of white light.

For red and blue only every fourth pixel is red or blue so the ratio is even bigger.

<<<-- Originally posted by Toke Lahti :

Example: white light coming in.
3ccd: 1/3 of the light is green and that gets into green ccd.
1ccd: 2/3 of the light is filtered from green pixel, so 1/3 of the light gets into green pixel.
Where's the difference just considering sensivity? -->>>

I think it is as follows... the 3CCD is noticeably more sensitive - somewhere in the neighbourhood of 3 times because the sensors are much smaller in the 1CCD camera - less light gets to them (a bit better than 1 f-stop sensitivity improvement)

3CCD - each CCD in this system is say 1/3 inch. Imagine it divided as a grid 720 by 480 pixels (for normal DV camera). Each pixel represents one pixel of the image in the one color (say green). There are 345,600 pixels in each CCD so each sensor receives 1/345600th of the total green light. But the prism redirected 2 thirds of the light to start with so we could say 1 millionth of the total light will hit the green pixel in question.

1CCD - for our 1/3inch camera there have to be three sensors in each of the locations to make up the image. Each with a filter. So total sensors on the chip are 720*480*3=1036800 sensors. Therefore 1 millionth of the total light hits our same green pixel. BUT the filter now removes the red and blue light leaving 1/3 of a millionth of the total light.

As I reread this it seems needlessly complicated and potentially confusing. anyone feel free to chime in with a simpler explanation :-)

Maybe another way to see it is that all the light entering a 3CCD camera eventually hits a sensor but 2/3 of the light entering a 1CCD camera is absorbed away by filters and never strikes the sensors.

Paul

To make it a little more clear (I think)...

1CCD - the color filter throws away 2/3 of incoming light before hitting pixel

3CCD - the prism deflects 2/3 of light away from each CCD & towards the others. Than all three CCD's charges are processed (combined) into one signal.

Is that right?

> I think it is as follows... the 3CCD is noticeably more sensitive -
> somewhere in the neighbourhood of 3 times because the
> sensors are much smaller in the 1CCD camera - less light gets
> to them (a bit better than 1 f-stop sensitivity improvement)

Correct. Perhaps somewhat less than three times, because some light is lost to the color dividing optics, but three sensor cameras are usually more sensitive than same-size 1 sensor cams.

Actually I think this is why Sony had to make one of it's new palmcorders 3-CMOS. One high-res CMOS at such a small size probably doesn't have enough sensitivity for your average birthday party.

For years we have underestimated the progress of inovation. When I first saw the JVC HD-10, I was impressed with the balance of priorities. The single CCD has a separate white and green pixel areas. Since green is more than half luminance info, this already gives a pretty high percentage of the CCD area to luminance. The yellow and cyan (or is it magenta?) elements also contribute a high percentage of the luma (especially the yellow). Such a combination allows a better luma S/N at the expense of the chroma. This is exactly what happened in the HD-10, as the chroma S/N suffers more.

I thought this was an excellent compromise, as there is seldom much chroma present in very low light. And there is more to come! It's risky to make predictions that a particular technology will not be superceded. I specifically remember that I thought that CCDs would never replace Plumbicons.

<<<-- Originally posted by Toke Lahti: Example: white light coming in.
3ccd: 1/3 of the light is green and that gets into green ccd.
1ccd: 2/3 of the light is filtered from green pixel, so 1/3 of the light gets into green pixel.
Where's the difference just considering sensivity? -->>>

So I need to continue a little bit:
Let's assume that 1ccd and 3ccd has same size pixels.
Then they will get same exposure.
If you want same resolution from both then 1ccd should of course have more pixels than one chip from 3ccd. And that means bigger chip for 1ccd.

If I had a choise I would take a camera with 2/3" 1ccd than 1/3" 3ccd.

> I would take a camera with 2/3" 1ccd than 1/3" 3ccd.

But you would get a better image with a camera with three 1/2" CCDs, and that would probably be less expensive than a single 2/3" CCD. Sensor size is critical to the cost of a camera system because it directly affects the size of the lenses and because larger sensors are much more expensive to make because yields are much lower, this means more CCDs have to be made to get a good one, thus more money is spent. This is how things have been for the SD world for some time now with CCD technology.

With higher resolutions, aligning three sensors accurately is probably more difficult and costs more money, hence the single sensor concept makes more sense than in the SD world, but large CCDs are expensive. However, large CMOS sensors are common now in digicams. And CMOS technology has gone a long way in the last few years. I think we will end up using giant (i.e. 35mm) single CMOS sensors for HD video, but not just yet. It will take some time, those thing are still expensive at large sizes, even if not as expensive as CCDs.

In fact the sensor itself isn't the biggest cost of a camera these days.
Think that a normal sensor costs around 200 dollars for a 1920x1080 one.
The cost itself goes on housing mechanics, IC and the like.
The sensors used on a consumer camera indeed cost just tens of dollars....
ICs cost more than them.Not to mention all the motors and mechanicals involved in tape mechanisms.