On Fri, 25 Apr 2003 23:41:25 GMT, "David McCall" <davidmccall@attbi.com> wrote:

>"Neuman - Ruether" <d_ruether@hotmail.com> wrote in message

news:3eab3167.24299426@newsstand.cit.cornell.edu...

>> On Thu, 24 Apr 2003 15:54:18 +0100, "Robin Davies-Rollinson"

 

>> If so, then it also does not crop the image, unlike

>> what the original poster said... I don't know, since

>> 16:9 is unimportant to me when using a 4:3-proportioned

>> sensor and display...

 

>I think "cropping the image" is a bit of a misnomer.

>

>Many cameras have CCDs with what appears to be a relatively

>arbitrary pixel count, but I think the pixels may be square, as

>opposed to rectangular. They use a 4 x 3 section of the chips

>for normal video, and a 16 x 9 array for the widescreen image.

>Early cameras may have had something closer to a 1 to 1

>relationship between the vertical resolution of the output and

>the actual cell count on the chips. And they used all or most

>of the available horizontal cells to create 3 x 4 or 16 x 9. To

>achieve the wide screen "effect" they would not use some of

>the available rows at the top and bottom, and then stretch what

>was left to fill out the full resolution. The result being that there

>was a loss of vertical resolution. In effect, 16 x 9 could not be

>thought of a "wider than 4 x 3, but rather, it was shorter, but with

>the same width. To be considered truly "16 x 9" you would expect

>that your field of view would get wider when shooting 16 x 9, while

>the height stayed constant. Some cameras (TRV-950 or PDX10

>for example) use megapixel chips that have more than enough

>on chip resolution to create a 16 X 9 image without the vertical

>resolution being compromised below the 1 to 1 relationship of

>CCD cells to Output pixels. The PDX10 actually takes a wider shot

>in the 16 X 9 mode than it does in 3 X 4 mode. The TRV-950 stays

>about the same width for both, but is still not stretching at all. There

>are plenty of cells on the CCD to stay above the 1 to 1 relationship.

>I don't know anything about the CCD cell count on the VX2000, but

>I know you don't get a wider view without an anamorphic lens.

>Just shorter.

>David

 

(Much of the following is a restating of what

you said, but possibly from a different

"perspective"...;-)

OK, I was curious enough to check the VX2000 just

now, and (at least in terms of what is seen in

the VF, which *could* be different on tape, with

"correction" seen in the finder - but I'm not

interested enough in this right now to check...;-),

the VX2000 appears to just crop the 4:3 image

to 16:9, throwing away part of the CCD info.

As for how others do it, it can be done with a

CCD with sufficient pixels for either mode for

full resolution (almost, in the PDX10, but not

quite...), with a CCD designed for 16:9 image

proportion (JVC DV GY700, et. al.), by an

anamorphic squeezing of the image on a 4:3

CCD (using rectangular pixels [stretched

horizontally] in the end result), or by an

anamorphic squeezing of the image before it

enters the camera's lens... The first is the

best solution (though higher than normal pixel

counts can have down-sides...), giving the best

flexibility while maintaining versatility; the

second is the best solution for 16:9-only

shooting; the last two are similar in resolution

effect (and both require reproportioning when

editing), but have practical differences:

the on-CCD anamorphic application permits

various lens accessories to be added easily

(WA, fisheye, macro-achromats, etc.), but

the basic picture is not as wide; the on-lens

anamorphic application makes using additional

lens accessories much more difficult - and

results will be more dependent on the quality

of the added anamorphic attachment, and on how

well the attachment and camera lens match and

perform together. All of these are likely

better than just cropping the image, though,

in terms of greatest total final resolution...

BTW, using an anamorphic solution does not

necessarily imply to use of a front lens

adapter, or a wider view - and the anamorphic

solution can result in odd visual effects,

as with the shape of out of focus highlights

being not symmetrical in all directions...

But, I still prefer 4:3 except for very large

displays viewed close...;-)