Date: Wed, 29 Jan 2003 22:35:11 -0000
Reply-To: secretary@dkvj-cons.co.uk
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: DKVJ Cons Co UK <dkv-j_consultancies@TOTALISE.CO.UK>
Subject: Re: Filming HASUG Meetings - Help
In-Reply-To: <20030129060008.536BB15B700@mail10.easyspace.com>
Content-Type: text/plain; charset="iso-8859-1"
Aye, there's the rub.
There are some presentation issues here. When you reduce an image in size,
some process will vote on which pixels to accept, and which to ignore. To
see what I mean, you could draw the capital letter 'E' on a piece of graph
paper. One cell height for each line, two cells height for each line
spacing. You will now have a structure that is seven cells high. Now
reduce that to four cells high. No matter how you calculate it, you either
lose one of the bars, or collapse two of them together. This is an artefact
of digitising in a lossy format. By reducing the size of the image, you
must lose information, and reducing the size of the image will cause
information to be lost.
If all you did was lose bits, that may be acceptable. Unfortunately,
because objects are moving in relation to the camera, the rendering of a
shape will keep changing. You've seen this on television when someone wears
a striped garment in the background, and the garment appears to move and
shift as the camera perspective shifts. It's less obvious in film, but when
film is digitised for television the stripes move between the area where
they are, and then are not captured. So the garment appears to strobe.
When you do this with a presentation, you're likely to lose parts of words
on the display, or make a highly unwatchable picture.
By the way, when I capture my DV material, it is first rendered in full AVI.
In this format, 13 minutes of video takes more than 1GB of disk space. As
it is rendered in MPEG-3 losses are introduced, and a piece of software
makes a decision about a clump of pixels that may render them all in the
same colour. The result is that my pan shot across the lake at last years
SUGI looks fine in native DV and AVI, but gradually becomes more indistinct
as the available space for image information is filled and pieces of my
picture are thrown away. MPEG-1 is substantially more lossy, even though it
did render my pan in a file that easily fit in a small part of a CD!
It's a frustration Charles seems to have experienced as well.
By the way, some of us use another video system called Phase Alternate Line
(PAL). While NTSC is based on 60 fields per second of 525 lines, PAL is
based on 625 line images at 50 fields per second ("more, less often" as my
Californian friend might remark). They also have different colour burst
frequencies, which led to some wag suggesting NTSC is a mnemonic that makes
a derogatory remark about NTSCs colour rendering ability. One must ask
though, how many people are still using 640*480 display sizes? And how many
will be happy watching a picture one quarter of that size? It might work
for bouncing babys, for entirely different reasons, but live video is
another matter. I come back to the question implicit in my first reply;
what are you trying to convey, what are you trying to achieve, and what is
the limitation of the medium?
Kind regards
David
Date: Tue, 28 Jan 2003 19:17:20 -0700
From: Jack Hamilton <JackHamilton@FIRSTHEALTH.COM>
Subject: Re: Filming HASUG Meetings - Help
Did you try AVI or any other formats besides MPEG?
320*240 is small, but you don't have to get much bigger. A standard
NTSC signal is only 525 lines high.
--
JackHamilton@FirstHealth.com
Manager, Technical Development
METRICS Department, First Health
West Sacramento, California USA
| 
