[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Information, Compression, and storage.
- To: "telecine internet group" <telecine at alegria.com>
- Subject: Re: Information, Compression, and storage.
- From: "Christopher Bacon" <KA2IQB at worldnet.att.net>
- Date: Tue, 6 Jul 1999 20:01:40 -0400
- Resent-Date: Tue, 6 Jul 1999 19:34:19 -0500
- Resent-From: telecine at alegria.com
- Resent-Message-ID: <"fpyWU.A.9L.ECqg3" at sun>
- Resent-Sender: telecine-request at alegria.com
- Resent-To: multiple recipients of <telecine at alegria.com>
>[...] computer programs there are such things as Lossless compression (e.g.
>PkZip and PkUnZip - although there are others I know). I'm curious about
>compression and how it applies to Transfers and distribution of the
There are a few types of video compression that are theoretically lossless
(meaning that they can be completely reversed). These include DCT, wavelet,
and fractal. It has to be understood that when you zip or unzip files on a
computer, you are dealing with files and it is usually unimportant how long
your computer takes to do the job, as long as it does so with complete
accuracy. Digital video is streaming data, which means that it has to be
handled in real time (at least part of the way), and everybody's system has
to go the same speed. Minor errors are tolerated in consequence. If
processor power was unlimited, these constraints would be irrelevant, but
limitations on the cost of hardware have kept wavelet and fractal
compression out of widespread use and made DCT-based systems less
transparent than they might otherwise be.
Unfortunately, with practical hardware, none of the presently known lossless
compression schemes is good for a ratio much larger than about 10:1 before
artifacts become objectionable. If greater compression is needed,
subjectively better results can be obtained through the use of lossy
techniques. This means that some of the image content is thrown away and
the original image cannot be exactly reproduced in most cases. But in a
well designed system, the lost information would be overlooked by most
people, so this is not objectionable--we hope! The dividing line is
between production and distribution. Low level, "lossless" compressions are
preferred in production and post; high, lossy compressions are used for
final distribution on the theory that the content is going to be viewed, not
further edited or changed.
> Basically I'm curious about, NTSC Broadcast Information rates, compared
>to: Standard NTSC Transfers, 1K transfers, 2K transfers, Data Rate
>Transfers, HD transfers (uncompressed if there is such a thing?).
In analog TV, there is no quantization in the horizontal lines; the lines
themselves represent a quantization of sorts. Hence, there is no way to
directly compare the amount of information carried in digital and analog TV
systems. However, the designers of the composite D-2 format used 143
Mbits/sec (NTSC) or 177 Mbits/sec (PAL), versus 270 Mbits/sec for component
(601) video in either standard, so you can get a comparative idea of the
amount of information these signals contain.
Using the above numbers, a single frame of composite NTSC video contains 477
kilobytes; a single frame of PAL contains 708 KB; and a single 601 frame
contains 900 KB--assuming 10-bit bytes in all cases. The data content of
scanned frames ranges from about 2 MB to more than 70 MB each, depending on
the color bit depth, the design of the scanner, and the size of the frames
being scanned (you'd have to be scanning Vistavision or 70mm to get to the
high end of that range!) HDTV at 1.44 Gbits/sec implies that a single frame
contains 4.8 MB of data.
>I'm curious about how much these figures change once you add in
>compression. I don't understand how much compression 50:1 actually is,
>because It seems that everyone just uses different math to show that
>although they are compressing, they aren't really.
One can get an idea of the data rate following one of the so-called lossless
schemes by dividing the above numbers by the compression ratio, but this is
academic since lossless compression is used mainly in professional VTRs at
present, and outside of tape machine manufacturers, we generally can't get
inside tape formats. Lossy compression, as is used for distribution, is
based on picture content (both detail and motion) and varies from instant to
instant, so there are no hard-and-fixed numbers.
> I'm interested in not only data/transfer rates for the telecine, but if
>anyone also knows about Data/???? rates for broadcast, or other forms of
The ATSC digital TV system calls for a data rate of 19.3 Mbits/sec for
broadcast (38.4 Mbits/sec on cable, so data compression is obviously a big
part of the plan.
>however I am also interested, because I am seeing really bad digitizing
>Artifacts popping up on Television. Most recently watching a program on
>one of the BIG three Networks, and their pre-taped introspective pieces
>all had really bad compression. Jaggy movement, and skin tone gradations
>that made people look like they were part of a topographical map.
Sounds like stuff that was thrown together on low-rent NLE systems. As long
as people are willing to put up with garbage, that's what they are going to
get. The technology is capable of far better, however.
Thanks to Rich Torpey for support in 1999
No advertising/marketing allowed on the main TIG. Contact rob at alegria.com
anonymous messaging now at http://www.alegria.com/HyperNews/get/ubique.html
1047 subscribers in 41 countries on Tue Jul 6 19:34:13 CDT 1999
subscribe/unsubscribe with that Subject: to telecine-request at alegria.com
complete information on the TIG website http://www.alegria.com/tig3/