Some days ago I heard that on new SONY camcorders (TRV60,75,80 with CCD of 2Mpix) smaller pixels are used (to fit them on 1/3,6 CCD)? Do you how does it affects video quality? (I've heard that it badly affects picture quality; actually, I disagree with it, but who knows...). So it would be very nice of you if you could post some info on this article. Thx!

Would the video quality be better than Panasonic MX500's (3CCD,1/6 each) or ~the same quality?

Nah. It will affect low light performance, in theory, but it's not that big of a deal.


"Would the video quality be better than Panasonic MX500's (3CCD,1/6 each) or ~the same quality?"

No. I'd personally rather take that camera (the mx500 is the same as the dv953, correct?) then one of the 1CCD Sony's. You will get better colors with 3CCD's, and I hear that camera isn't so bad in "low light".

The DV format needs about 346K CCD. Once you get to 2 Mbps CCD, the color performance will about equal to a 3 CCD DV system. The larger CCD will have a lot more shallow DOF, to better simulate the film look.

Then perhaps you could explain why even the TRV70's (with 2.1MP CCD) color saturation is so inaccurate.

Joseph. You get better colors when the footage is split into the 3 colors via a prism. The TRV70/80 has only slightly above 1M video effective CCD pixels; with a 3 CCD cam, you take the video effective CCD pixels and multiply that number by 3. So you see, the MX500/PV-DV953 has more video effective CCD pixels than your beloved Sony 1 chipper. However, This Sony 1 chipper is also a good cam.

I don't want to argue with you guys. The one chip is enough to gather all the information; how is it processed further depends on the manufacturer's preferences that have nothing to do with technology or economic limitations.

I also don't want to argue, but the more information a cam can capture, via better lens, more video effective CCD pixels, the more accurate the information will be when processed down, at or below DV's max 540 playback horizontal line limit. A 1 CCD will not capture color as accurately as a 3 CCD camera. I believe several experts here have already explained this to you. Remember?

I'll admit that there's a lot I don't understand when it comes to this topic. Why is it that a good STILL digital cameras only uses 1 CCD for example? The JPEG's that come from my Nikon 950 at 2048x1536 pixels look great to my eyes, way better than frames from a DV camera. I always assumed that the real limitation to DV was the color space afforded by the 4:1:1 compression. Those Nikon frames are about 1.5MB each when I set the quality to 'fine', but a DV frame is only around 100KB and of course 30 of them need to be captured and compressed every second.

I also do 3d modeling and animation. When I compress these with the DV codec for editing in FCP they really suffer. For example, a nice uniform blue sky is broken up with ugly artifacts and a smooth gradient ends up with visible bands. In this case it has nothing to do with CCD's, it's just the toll that DV compression takes on the original clean images.

Digital still cams have a large, state of the art, mega-pixel CD, plus a high quality lens/lenses. They also capture frames instead of 60 fields per second/50 for PAL.

There is another, very important, difference:

Digital stills cameras can take much longer to acquire an image than video cameras. The process can sometimes take more than a second. Video cameras do not have this luxury, since they need to capture 25 (PAL) or 30 (NTSC) frames a second, and move that data out to make room for new data.

This, as far as I know, is the main reason that digital video cameras will never be as good as digital stills cameras with approx. the same resolution (unless a manufacturer would actually built a video camera with separate CCD and circuitry for digital stills, which would probably not be cost-effective).

Ariel

The purpose of having 3 CCDs is to give more resolving ability to the picture as a whole. One 256k CCD is a lot cheaper than one 768k chip (256k x 3). A 768k chip will provide the same color and resolution as the 256k by masking the on board devices for RGB.

The reason digital cameras take longer per frame is they have to transfer data into memory from the CCD. This is not high speed memory like in your computer and CCDs themselves are not fast at transferring data.

Long time ago there were one vacuum tube (Saticon, etc.) sensor cameras and 3-tube cameras. The 3 tubes cameras had great colors. The one tube cameras did not. Then Sony came out with a 1 CCD camera and the colors were on the grade of 3 tube cameras, or better.

Then came 3 CCD cameras at time when CCD's still had low pixel count. So the 3 CCDs improved resolution over one CCD.

Now, with high enough pixel count, one CCD is all you need for the 4:1:1 DV color compression system.

3 CCD cameras have more to do with marketing than with performance.

There is no difference between a still camera and a video camera CCD function, except for the higher speed/frame rate.

Once you get to Hi Def and 4:4:4 like CineAlta SR, 3 CCD's with perfectly aligned pixels are a must. Even on 4:2:2 compression 3 CCD's with aligned pixels will improve color performance significantly. On DV it does not really matter.

Varicam and CineAlta use 3 CCDs with perfectly aligned pixels.

As far as CCD size is concerned, if you hold 'em out at arm's length, they all appear to be pretty much the same.

<< 3 CCD cameras have more to do with marketing than with performance. >>

Incorrect... nothing could be farther from the truth. A trip to the local retail electronics store will readily reveal that any current three-chip camcorder produces an image clearly superior to any current one-chip camcorder.

It has *nothing* to do with marketing whatsoever.

There goes more care into a 3 CCD camcorder to make the image good. These are more expensive products. The quality goes beyond the CCD count. Yes, 3 CCDs are better. But on a mega resolution chip, you'll get enough info for the DV system to produce excellent picture.

<< But on a mega resolution chip, you'll get enough info for the DV system to produce excellent picture. >>

There's much more to it than that. The type of color filtration mask has more bearing than the pixel count. For example, a sub-megapixel chip with a primary color filter will produce a "better looking" image than a megapixel-plus CCD with a complimentary color filter.

Megapixel-plus chips may produce still image captures with larger print output sizes from a camcorder, but standard definition DV is still 720x480.

Hi,

I think that the post-processing behind a 1-CCD cam is the most important factor that determines the quality. The DSP would need to 'fill' the gap by approximating the other two colors not in the current grid array.

I guess, that explains why 1-CCD cams produces 'color' video noise whereas 3-CCDs models output less prominent monochrome noise.

However, higher pixel count CCD does capture more precise details/colors (better pixel picking) especially when shooting small/complex scene where a specific color falls between 2 CCD pixel.

What exactly is JVC's "high band" CCD? How does it improve the video? Anyone?

Most prob. the 'high-band' processor refers to more advanced DSP processing technology, especially with higher horizontal resolution captures.

Hi,

Just compare from the CCD of DV, there are 3 factors may affect the output quality:
Definitely 3CCD is better than 1CCD
Definitely larger CCD is better i.e. 1/4 > 1/6
More pixel is better although DV output is 720x480 so 2M > 1M > 68K

However the point is which factor is more important? For example Panasonic GS100K is 1/6 3CCD 700K pixel. How to compare it with Sony TRV60/70/80 which is 1/4 1CCD 2M pixel?

Any simple formula can tell us the comparison?
e.g. 3 * 700K * (1/6) vs 1 * 2M * (1/4) ?

Don't go by the total CCD pixels, go by the video effective CCD pixels. After all, it's these pixels that capture the video. Regarding CCD size (and pixel size), the larger these are, the less light they'll need. My old DVL9500s have a 1/3" CCD, with big 380K pixels (360K video effective), so I can shoot in lower light where other cams are blind.

<<<-- Originally posted by Frank Granovski : Don't go by the total CCD pixels, go by the video effective CCD pixels. After all, it's these pixels that capture the video. Regarding CCD size (and pixel size), the larger these are, the less light they'll need. My old DVL9500s have a 1/3" CCD, with big 380K pixels (360K video effective), so I can shoot in lower light where other cams are blind. -->>>
Hi Frank,

Agreed with your point that larger CCD is definitelly better (of course if it is F1.2-F1.4 len is even better). However now compact DV seldom has 1/3" CCD size and 1/4" CCD or 1/6" CCD is common.

It is now perfect if a compact DV can have 3CCD of 1/4" size and with 1M effective pixel at least.

Which compact DV model has such config or it is a dream?

t is now perfect if a compact DV can have 3CCD of 1/4" size and with 1M effective pixel at least. Which compact DV model has such config or it is a dream?

Do you mean 1 CCD cams or 3 CCD cams? With 3 CCD cams it's the video effective CCD pixels X 3. So cams like the PV-DV953/MX5000 and PDX10 are right up there. Over 1000K on a 1 chip? That would be the Sony TRV70/80.

<<<-- Originally posted by Frank Granovski : Do you mean 1 CCD cams or 3 CCD cams? With 3 CCD cams it's the video effective CCD pixels X 3. So cams like the PV-DV953/MX5000 and PDX10 are right up there. Over 1000K on a 1 chip? That would be the Sony TRV70/80. -->>>
Hi, what I mean is a compact DV with 3 CCD and each CCD size should be at least 1/4" and may have 1M pixel for each CCD. It may not be possible now but from the model release speed of 3CCD vendor like Panasonic, 1 or 1.5 year later may have such model.

The PAL MX300 and MX350 are compact with 1/4" CCDs, and the total (X 3) video effective pixels are over 1M.