home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!sdd.hp.com!hplabs!felix!fritz!ung
- From: ung@felix.filenet.com (Bill Ung)
- Newsgroups: rec.video
- Subject: Re: CD-I to compete with LD?
- Keywords: CD-I, LD
- Message-ID: <20935@fritz.filenet.com>
- Date: 18 Nov 92 02:22:21 GMT
- References: <1992Nov9.140656.8026@cs.wisc.edu> <27972@castle.ed.ac.uk> <1992Nov11.161446.2677@ifi.unizh.ch> <4013@eastman.UUCP>
- Reply-To: ung@fritz.filenet.com (Bill Ung)
- Organization: FileNet Corp., Costa Mesa, CA
- Lines: 54
-
- In article <4013@eastman.UUCP> rao@kodak.com (Arun Rao) writes:
- |In article <1992Nov11.161446.2677@ifi.unizh.ch>, hafner@ifi.unizh.ch (Michel Hafner) writes:
- ||| In article <27972@castle.ed.ac.uk> sss@castle.ed.ac.uk (S S Sturrock) writes:
- ||| |Rumour alert.
- ||| |A friend of mine met a chap who is experimenting with fractal compression
- ||| |who showed him a 486 PC laptop with the whole of Casablanca compressed onto
- ||| |a 1.44 meg disc. The image was small but only because the PC didn't have
- ||| |the performance to uncompress the image to a large display, the fellow
- ||| |claims that every detail is there and only needs a more powerful computer
- ||| |to show it.
- ||| |If this is true, and I have no reason to disbelieve it, and this can be got
- ||| |to market CD-I will be dead along with all the other compression systems
- ||| |used today for video.
- |||
- ||| I do have reason to disbelieve it. Let's do some calculations:
- ||| Let's say "Casablanca" is about 90 minutes long, then there are
- ||| 90 * 60 * 24 ~ 130000 frames. Every frame has, let's say, 300 * 450 Pixels
- ||| ( mediocre spatial resolution ) a 2 Bytes ( D1 standard ) ~ 270000 Bytes,
- ||| gives a total of ~ 35 Gigabyte or a compression factor of over 24000!!!
- ||| This is not possible! I have heard of factors around 100 for single images,
- ||| but not many thousands for sequences. "Casablanca" compressed to 1.44 Mega-
- ||| bytes has very small single frames with reasonable quality. Further blow up
- ||| will add detail, but not detail identical to the detail of the original
- ||| frames. At best it looks strange, at worst it gets unwatchable.
- ||| M.H.
- |
- |
- | ... but fractal compression is another story altogether, if we're
- | to believe researchers. Compression ratios of 10,000:1 are routinely
- | claimed for fractal techniques, but I have not heard of anyone doing
- | anything quite so, well, *practical* with them.
-
- Also, I'm not sure if the compressed movie is just a bunch of frames.
- There are packages that allow you to store numerous versions of a file.
- The first time it stores the file's image, the next time, it merely
- stores the DIFFERENCES between file1 and file2. In the movie business,
- you're going to get a LOT of frames that are VERY similar, surely it
- would be much more efficient to store just the differences between
- frame1 and frame2, than to store the whole frame (assuming that the
- hardware was fast enough to extract the data). If this technique is
- being used, along with the amazing fractal compression techniques, it
- may very well be quite possible.
-
- | -Arun
- |
- |
- |Arun Rao, PhD
- |Senior Research Scientist
- |Recognition & Information Processing Group
- |Eastman Kodak Company
- |Rochester, NY 14650-1816
-
- Bill Ung
- ung@filenet.com
-