Can you help?

indiandy · #1 (**permalink**) 10-27-2006, 05:51 AM

HI
Can anyone help with the following question

Digital video is a fixed format: 4:3 PAL miniDV video is 720 x 576 = 414,720 pixels.

- Does a camcorder really need ten times as many pixels?
- Does the digital image processing algorithm really use information from all these pixels or just from a tiny sub-set?

Thanks

Mark W · #2 (**permalink**) 10-27-2006, 11:06 AM

Not sure what you mean there?

A dv camera does 'use' all those pixela but the codec compresses the video so that the finished picture isnt anywhere near as good as the resoloution figure.

For example - digi beta SD has the same picture quality but looks much more detailed becuse it isnt compressed anywhere near as much. I have seen this 'low res' (ie not hdv) format projected and it looks amazing.

The colour res in dv is very poor, anyone who has done keying on dv footage or played with curves / sat in their editor will have noticed this, colour bleeds at edges and gets all blocky because of the very poor colour resoloution.

TBH it isnt the absoloute res that makes me want hdv but the improved colour res really.

indiandy · #3 (**permalink**) 10-27-2006, 11:29 AM

1/ Is all 4:3 PAL Mini DV video 720x576 pixels

2/ Why do you need more than the 414,720 pixels in total, if you only film for video.

Sorry I did not make myself clear, hope this helps

Mark W · #4 (**permalink**) 10-27-2006, 11:50 AM

Ah, I see, you mean why do some cams say they have a higher pixel count yet output at dv res.

Simple answer I think MARKETING . In the still camera field pixel count does ammter, although not as much a lenses do, but people have come to expect that more pixels means better.

One of the best dv cams around, sony dsr300 pro camera, about 8000 quids worth only has CCDs with 410,000 pixels, nuff said.

edit

Having said that if you want to shoot 4:3 and 16:9 then a ccd with more pixels stops quality loss in w screen.

Evereddie · #5 (**permalink**) 10-27-2006, 01:25 PM

Pixel count DOES matter to a point. It is better to have the electronics do the averaging using smart algorithms, than the pick up pixels. For instance, if fine detail (a small dot lets say) happens to fall right between 2 pixels, it may be missed or it may 'light up' both pixels. Not a preferable situation since it will not be represented in the proper way. If there are more pixels to fall on then it will be 'caught' in the optical process of the pickup and the electronics can use smarter algorithims to work it out.

Also as said above DV isn't the best and if you really want pro footage that a high end programmer would want, then it has to be something that records component. Beta or the like.

Mark W · #6 (**permalink**) 10-27-2006, 02:08 PM

These is a lot too this, as a rule larger count ccds are also tend to be more noisey, and lees sensitive to light.

The quality of lenses, and the processing chips is far more important than pixel count. I wouldnt advise anyone to take any notice of pixel count when buying a camera, it is the quality of the ouput that matters, not how it got there, if you see what i mean...

indiandy · #7 (**permalink**) 10-27-2006, 08:26 PM

Thank you both for your help, you have really explained that in a way even I can understand.

Andy