|
|||||||||
|
Thread Tools | Search this Thread |
July 13th, 2004, 04:02 PM | #1 |
Regular Crew
Join Date: Mar 2002
Location: Kirkland, WA
Posts: 177
|
Pixels question
If a CCD in 4:3 mode were to be exactly 720 x 480 that would work out to 345,600 effective pixels. Given that some CCD's have over 400,000 in their 4:3 dimensions, they must be something other than 720 x 480 which means the image must be re-sampled to get to the native DV resolution of 720 x 480.
What advantage is there to going higher than 720 x 480 if the image will be re-sampled later? Do HD cams use CCD's at a higher resolution than 1920 x whatever and then down sample? This is highly theoretical but I am curious if there is more to gain than marketing hype. Greg |
July 13th, 2004, 04:12 PM | #2 |
Warden
Join Date: Mar 2002
Location: Clearwater, FL
Posts: 8,287
|
Chips are measured in different ways by different manufactures etc. Effective pixels, gross pixels etc. are measuring different things. Some pixels do not fire (not used) and they are used to determine black level or zero voltage. These pictures are not part of the image and are not counted, but are part of the CCD's pixel count.
__________________
Jeff Donald Carpe Diem Search DVinfo.net for quick answers | Where to Buy? From the best in the business: DVinfo.net sponsors |
July 13th, 2004, 11:42 PM | #3 |
Inner Circle
Join Date: Jun 2003
Location: Toronto, Canada
Posts: 4,750
|
3CCD systems: I believe the ideal setup is to have slightly more than the frame dimensions. i.e. for DV, the camera should have slightly more than 345,600 pixels. Some of the extra pixels may be used for image stabilization?
All the best cameras seem to be this way, even the high end Sony DSR series DVCAM cameras (those cost like $20k). 1CCD systems: I'm fairly sure many of these cameras down sample the pixels into 720X480. Many cameras have way more than 345k pixels. Part of this is to boost the megapixels it can shoot for stills (to boost the marketing hype). With cameras like that, a lot of pixels are wasted for shooting stills (hence effective video pixels is a lot lower). Anyways, the practical thing is this: Look at the still grabs from various cameras at the a haru site (japanese site with japanese models). It is a good way to see for yourself the difference in resolution between cameras. http://babelfish.altavista.com/babel....html&lp=ja_en Compare the TRV22 (~380k pixels) versus it's bigger brothers with more pixels (TRV33/38?). You should be able to notice the difference. I can't think of anything better you can do except for: A- Getting user experience of some cameras versus others. The problem with this is that some people over/underemphasize the resolution differences between cameras. B- Getting a hold of the cameras in question and doing a comparison yourself. With frame grabs for various cameras out there on the a haru site and others, I probably wouldn't go to this effort if you are only converned with resolution. However, it is a good idea to play with cameras you want to buy to see if it fits your shooting style and to see the color reproduction on an accurate monitor. |
July 14th, 2004, 03:52 AM | #4 |
RED Code Chef
Join Date: Oct 2001
Location: Holland
Posts: 12,514
|
Keep in mind that the pixels on a CCD block are, I think, SQUARE.
Pixels in DV are NOT. DV NTSC uses a 0.9 pixel aspect ratio. I *assume* they are using more height instead of less width to capture the info. The resolution then would be 720 x 533 for a total 384.000 pixels. Still no 400K, but it is getting close.
__________________
Rob Lohman, visuar@iname.com DV Info Wrangler & RED Code Chef Join the DV Challenge | Lady X Search DVinfo.net for quick answers | Buy from the best: DVinfo.net sponsors |
July 14th, 2004, 08:01 AM | #5 |
Regular Crew
Join Date: Mar 2002
Location: Kirkland, WA
Posts: 177
|
<<<-- Originally posted by Rob Lohman : Keep in mind that the pixels on a CCD block are, I think, SQUARE.
Pixels in DV are NOT. DV NTSC uses a 0.9 pixel aspect ratio. I *assume* they are using more height instead of less width to capture the info. The resolution then would be 720 x 533 for a total 384.000 pixels. Still no 400K, but it is getting close. -->>> Rob, If 640 x 480 takes up the same space with square pixels that 720 x 480 does with rectangular, wouldn't that imply the pixels are using less width but the same height since both are 480 pixels tall? Greg |
July 14th, 2004, 08:06 AM | #6 |
RED Code Chef
Join Date: Oct 2001
Location: Holland
Posts: 12,514
|
Hmmm... my thing would not work indeed since the pixels would
then change their aspect in the vertical direction and not in the horizontal. So the camera must be doing 654x480 in 4:3 and 872x480 in 16:9....
__________________
Rob Lohman, visuar@iname.com DV Info Wrangler & RED Code Chef Join the DV Challenge | Lady X Search DVinfo.net for quick answers | Buy from the best: DVinfo.net sponsors |
July 14th, 2004, 04:21 PM | #7 |
Wrangler
Join Date: Dec 2002
Location: Mays Landing, NJ
Posts: 11,802
|
This topic interests me too, so I recently did some tests to explore it. I shot frames of the EIA1956 res chart with a Nikon 5700 resulting in a 4,613K pixel image from a 2/3" CCD that I downsampled to 720x480 and DV-compressed. I then compared it to frame grabs from a VX-2000 (~340K pixels with 3 1/3" CCD's) and a PDX-10 (~664K pixels with 3 1/4.7" CCD's). You can draw your own conclusions here: http://greenmist.com/dv/res and also view similar tests in an outdoor scene with lots of detail here http://greenmist.com/dv/garden. You might also be interested in how the PDX-10 maps the pixels from its megapixel CCD here: http://greenmist.com/dv/16x9/10.JPG.
Even though the PDX-10 has far more pixels than the VX-2000 the image is somewhat lower resolution. But neither camera approaches the quality of the Nikon. Now my test was largely designed to see how much damage DV compression does to a detailed image. To my surprise the answer was "not much". The real limitation to resolution seems to lie in the CCD's, optics and other camcorder hardware which are designed to record full motion video at an affordable cost. So I'm not sure how much stock you should put in comparing resolution between different camcorders in the same general price class. Pixels are only one part of the story. However I would be very interested to see how images from the new XL-2 compare to my tests. From what I'm reading, it's optimized for 16:9 performance at the expense of 4:3. |
| ||||||
|
|