Pixel Count

[Bill Moody]

Hello everyone! It is a pleasure to read these forums - very helpful.

I need some technical help:

Having reviewed Chris Hurd's article, http://www.dvinfo.net/canonxl2/articles/article06.php, I am wondering some things.

I have a Canon GL2. It works well, but I would like to go to a 1/3 inch, 3-CCD camera because "bigger is better", or so I've heard. The GL2 has a total of 380,000 effective pixels which is more than the XL2 (345,600 in 4:3 mode), as well as the Sony VX2100 (340,000), and is the same as the Panasonic DVX100 (380,000). So, if the GL2 has 1/4" CCDs, why is the effective pixel count greater than the other cameras?

Secondly, I realize the XL2 shoots native 16:9, but why would Canon reduce the active area when shooting 4:3 instead of increasing it to fill the entire 4:3 aspect of the CCD?

Finally, and really the most important question, is it really beneficial to switch to the larger CCDs - are the images truly better? (I borrowed an XL2 and shot some footage, then shot the same footage with my GL2 - I didn't see much difference, albeit it was a limited comparison.)

Thank you in advance for your replys.

[Bennis Hahn]

Originally Posted by Bill Moody

Why is the effective pixel count greater than the other cameras?

You can cram smaller pixels onto a smaller CCD and it will have more pixels then a larger CCD, but b/c they are smaller, latitude and sensitivity will suffer.

Secondly, I realize the XL2 shoots native 16:9, but why would Canon reduce the active area when shooting 4:3 instead of increasing it to fill the entire 4:3 aspect of the CCD?

The XL2 has native 16:9 CCDs, so they are taking 4:3 out of the center of the chip.

Finally, and really the most important question, is it really beneficial to switch to the larger CCDs - are the images truly better?

If you didn't see a difference, then don't switch just because someone says it's better. Personally, I can see it and the form/function factor on 1/3' cameras is better then that of smaller imagers, so that was the deciding factor for me (although it was, way down the list). I have a DVX.

[Bill Moody]

Thank you for the reply, Bennis.

Originally Posted by Bennis Hahn

You can cram smaller pixels onto a smaller CCD and it will have more pixels then a larger CCD, but b/c they are smaller, latitude and sensitivity will suffer..

Are, in fact, the pixels smaller? Do you think the makers of the CCDs are producing chips with smaller pixels on the less expensive, 1/4" CCD chip cameras? It seems to me that the technology for the smaller pixel exceeds the technology for the larger pixel, and therefore should be more expensive. It is like comparing a 2 MP still camera to a 10 MP camera. The CCD/CMOS is the same size, the pixel technology is far superior in the 10MP camera. I don't know for sure, but the <$8,000 video cameras likely have the same CCD technology as far as pixel size is concerned. Is it just a gimmick on the part of Canon et al to tout a 1/3" CCD as a better image camera when there is no apparent additional image quality gained - instead what is gained is just form/function of the camera?

The XL2 has native 16:9 CCDs, so they are taking 4:3 out of the center of the chip.

My point exactly. When the 4:3 is taken out of the center, it produces an area for image capture that is (apparently?) smaller than the area in the GL2 - at least the effective pixel number is fewer. So, when using the XL2 in 4:3 mode you capture less information than when using the GL2, assuming pixel size/technology is the same. This is the heart of my original question - it doesn't make sense to me that first, the effective pixel number would be fewer in the XL2, and secondly, that Canon would use just the center of the chip since the whole chip is in a 4:3 ratio. However, I do understand why they would use existing technology (for the smaller 1/4" chip) in the larger 1/3" chip when gathering image data - it is less expensive to use existing technology.

To me it is not much different than saying I will sell you an 8 cylinder car that happens to use only 6 of those cylinders.

If you didn't see a difference, then don't switch just because someone says it's better. Personally, I can see it and the form/function factor on 1/3' cameras is better then that of smaller imagers, so that was the deciding factor for me (although it was, way down the list). I have a DVX.

My experience with the XL2/GL2 comparison was quite limited. What do you see that is different when comparing the DVX image to the GL2 image (or any other 1/4" CCD, 3-chip camera)? No doubt the form/function of the XL2/DVX are far superior to the GL2, and I guess one could argue those things alone justify the cost difference.

Does anyone know for sure if the pixel size of the GL2 and the XL2 are the same or different, and has anyone posted images comparing the image quality of the two cameras?

Thanks again for the help in understanding these issues.

[Boyd Ostroff]

Welcome to DVinfo Bill!

As part of your research you might spend a little time in our high definition camera forums: http://www.dvinfo.net/conf/forumdisplay.php?f=61

Standard definition cameras have some advantages, light greater sensitivity to light (in some cases), but when you look at a high definition image you will understand the significance of pixel count.

The XL2 is a very nice camera, and the true progressive scan will produce some of the sharpest images possible using the standard definition DV format. But it's a bit of a transitional product, with the XLH1 as the next evolutionary step.

Also see the discussions in the Sony V1 forum: http://www.dvinfo.net/conf/forumdisplay.php?f=140 This new camera seems to redefine what 1/4" chips can do, and it uses CMOS instead of CCD technology. I attended the event where Sony introduced this camera, and the footage they showed on a 36' screen was really impressive.

If you're thinking of upgrading from the GL2 then find a place to see some footage from these new cameras, including Canon's upcoming G1 and A1 models.

But aside from the video format, do you like the size and shape of your GL2? The XL2 is much bigger and heavier. That might be either a good or bad thing, depending on how and where you use the camera.

[Chris Colin Swanson]

Remember that in 4:3 SDTV the resolution is 720x480=345,600. This means that any pixel more than this will not show up when watched on a tv. So to maximize the ccd or to get the largest area per pixel on the ccd the xl2 is designed to only capture the number of pixels that will be used.

[Glenn Chan]

Some cameras have extra pixels since they implement image stabilization... and you need extra picture on the sides to do that.

The DVX100 implements pixel shifting (even though it doesn't necessarily need to, since it has more than 3 X 720 X 480 pixels). I think this is so its autofocus can be more accurate. Pixel shifting sort of increases the luminance resolution.
I believe the downside to the DVX100's pixel shifting is a teensy bit of aliasing.