a544@mindlink.UUCP (Rick McCormack) (11/25/90)
in an article, Sherman Wilcox writes: >And now, a question for anyone out there who might be reading >this. I keep hearing that we will soon have the technology to >do a project like this on CD-ROM. Does anyone have any idea >approximately how much full-motion video we can expect to >store on a CD-ROM? I realize this depends on many things (as >I understand it, picture size, frames per second, compression >factors, etc). Any ballpark figures? I remeber hearing that we could expect to get less than 20 minutes of realtime video, when they get the new CD-I, CD-X, CD-A discs functional. CD-I is supposedly well into the development stage, but that has been said for 14 months now. CD-X is in use industrially (anyone out there using this beast yet?) I have figures here from a research project this summer - but I can't FIND them yet!!! When I clear my desk, I will post a more definitive answer. _____________________________________________________________ | Rick McCormack | IMAGISTICS BUSINESS THEATRE TECHNOLOGY | | Vancouver, BC | Information transfer - with a purpose. | | CANADA | INTERACTIVE COMPREHENSIVE ENLIGHTENING | |________________|___________________________________________|
pweiss@iwarp.intel.com (Paul Weiss (pweiss@iwarp.intel.com)) (11/29/90)
Hi. "How much {picture,sound} time can ya fit on a CD?" is a
deceptively complicated question. CD-I and DVI use the same
underlying medium (as do CD-DA and CD-ROM, fundamentally) but make
different choices about how to use the bandwidth. The disc can hold
about 640Kb of "stuff", in any case. As far as video goes, DVI uses
fancy (and very proprietary) compression/ decompression to allow
sustained full-screen, full-motion pictures. I think the compression
factor varies in some way with the material, from a low of 10:1 to a
high of about 100:1. (Can anybody do a better job on that? Please
followup!) Although decompression can be done in real time, with
Intel's 2-chip set, compression is a mainframe mega-cycles kind of
proposition, with supercomputer-class machines ideal.
CD-I doesn't specify any video compression, so is unable to support
full-screen, full-motion rates: there just ain't enough raw bits
flowing off the disc to get it. You can have full-motion partial
screen video (to a max of about 1/3 of the screen), or "slow-scan"
styled jerky full-screen video - about 8 or 10 frames/sec, if I
remember correctly. (I'm dipping 'way back on this stuff. I worked
on a CD-I authoring system about 2 1/2 years ago while I was in the
Advanced Development / New Media Workstation group at Sun; the work was
done under contract for Philips, who, along with Sony, is the source
of all the CD-xx standards. The stuff I'm saying may be wrong based
on revs to the Green Book which I don't know about. Of course, it may
also be bogus because - in my mom's words - "I'm up to the 'z' in
Altzheimer's!")
CD-I does a much better job on audio. It supports 4 or 5 different
encodings, ranging from CD-DA tracks on a CD-I disc, to an encoding
which uses half that space and bandwidth, giving twice the time, all
the way to a "voice-grade mono" encoding which would allow 20 hours of
sound. Most listeners can't tell the difference between the CD-DA and
the half-rate encoding. Although the sound quality of the most
efficient encoding is noticibly tinny for music, it's just fine for
voice information. The other encodings I haven't identified lie
between those extremes, and can be used to juggle time-space and audio
quality in situations where the listener's attention is more-or-less
taken up with visual data. CD-I's audio encoding doesn't have nearly
the compute-time requirements as DVI's video encoding, and has been
done successfully on PC-class authoring stations.
For slide-show styled titles, using a combination of still video
images, computer-generated box-and-arrow markup, cartoon graphics
animation, and voice-over, CD-I will be a terrific medium. For things
which want a lot of whizzy full-motion video, DVI will probably do a
better job. Both of them will do a better job than the current
platform-dependent CD-ROMs will. (There is another flavor of
Sony-Philips disc which adds CD-I audio encoding to the CD-ROM
standard, but the thing is still highly platform-dependent.) DVI
shares the same platform-dependence as CD-ROM, at least in theory, but
the new 2-chip set should make the appearence of dedicated DVI players
economically possible. CD-I was designed from the start as
mass-market consumer technology. The question will, I bet, finally
work out to a question of time-to-market and title production. Both
media are usable, neither is perfect. Both benefit from the economies
of scale generated by the success of consumer CD-DA technology, which
has not been available to the "big-disk" videodisk marketplace.
By the way, I would very much appreciate pointers to companies and
organizations working in computer-based new media in the Portland, OR,
area. I miss working with this stuff, and would like to find out
what's going on around here. Tanx.
Paul
-------------------------------------------------------------------------------
Paul Weiss <- Mangler, systems software, Intel/CMU iWarp project
pweiss@iWarp.intel.com <- I wouldn't even try the "reply" key, were I you.
(503)629-6371 <- ... as in "webby feets"
Standard disclaimer: Yep. <- Strange powers speak through me, but Intel doesn't
-------------------------------------------------------------------------------
--
-------------------------------------------------------------------------------
Paul Weiss <- Mangler, systems software, Intel/CMU iWarp project
pweiss@iWarp.intel.com <- I wouldn't even try the "reply" key, were I you.
(503)629-6371 <- ... as in "webby feets"jimf@idayton.field.intel.com (Jim Fister) (11/29/90)
pweiss@iwarp.intel.com (Paul Weiss (pweiss@iwarp.intel.com)) writes: >Hi. "How much {picture,sound} time can ya fit on a CD?" is a >deceptively complicated question. CD-I and DVI use the same >underlying medium (as do CD-DA and CD-ROM, fundamentally) but make >different choices about how to use the bandwidth. The disc can hold >about 640Kb of "stuff", in any case. As far as video goes, DVI uses >fancy (and very proprietary) compression/ decompression to allow >sustained full-screen, full-motion pictures. I think the compression >factor varies in some way with the material, from a low of 10:1 to a >high of about 100:1. (Can anybody do a better job on that? Please >followup!) Although decompression can be done in real time, with >Intel's 2-chip set, compression is a mainframe mega-cycles kind of >proposition, with supercomputer-class machines ideal. Not to be picky, but it's 640Mb of data on a CD ROM. I'm not to sure about CD-I, but DVI can do about 70 min. of full-screen, full-motion video. Compression on the high end, using the mainframe is about 120:1 now. DVI can also do real-time compression. The capture board on the two-board set does about 80:1 right now, and it'll only get better. It looks to be about standard VCR quality, if you want a reference. >CD-I does a much better job on audio. It supports 4 or 5 different >encodings, ranging from CD-DA tracks on a CD-I disc, to an encoding >which uses half that space and bandwidth, giving twice the time, all [stuff deleted for BW] >taken up with visual data. CD-I's audio encoding doesn't have nearly >the compute-time requirements as DVI's video encoding, and has been >done successfully on PC-class authoring stations. >For slide-show styled titles, using a combination of still video >images, computer-generated box-and-arrow markup, cartoon graphics >animation, and voice-over, CD-I will be a terrific medium. For things >which want a lot of whizzy full-motion video, DVI will probably do a >better job. I'm not so sure, but I don't have the info to argue. >...DVI >shares the same platform-dependence as CD-ROM, at least in theory, but >the new 2-chip set should make the appearence of dedicated DVI players >economically possible. CD-I was designed from the start as >mass-market consumer technology. DVI can handle any transfer media with a BW of 150Mb/s or greater. Our office has successfully transferred RTV over a network, along with stills. Platform dependance is true not, but not necessarily in the future. Just my two cents. I'd say more, but my lack of info on CD-I would only make me seem more of an idiot than I'm normally accused of being. Any CD-I hacks out there? Oh, the fact that the door outside the office says "Intel" doesn't mean that I say what they mean, or that I mean what they say, right? Other standard disclaimers apply. Greetings from the rocking metropolis. JimF
eric@mcrware.UUCP (Eric Miller) (12/08/90)
I have been reading the various discussions of CD capacity lately and I figured that it's about time to add my two cents. Just as Jim Fister is more experienced with DVI than CD-I, my bias is with CD-I, although I have worked on the standardization committees for CD-ROM XA as well as CD-I. This is the general story: CD-ROM/CD-I/DVI all use the same carrier. They all store data within the context of the ISO 9660 standard for storing data on an optical disc. This is the same optical disc that is used to store music (CD-DA) although of course music discs don't have directory structures on them. CD-I and CD-ROM XA (and perhaps DVI) store audio in an ADPCM format which *may be* streamed directly into an audio processor. CD-I allows you to also read that data into memory for buffering and later playback. CD-I allows four levels of audio: CD-DA (16 bit PCM - 44.1 KHz) 72 minutes Hi-Fi (8 bit ADPCM - 37.8 KHz) stereo - 144 minutes mono - 288 minutes Mid-Fi (4 bit ADPCM - 37.8 KHz) stereo - 288 minutes mono - 576 minutes Lo-Fi (4 bit ADPCM - 18.9 KHz) stereo - 576 minutes mono - 1152 minutes At Lo-Fi mono, this gives appx 19 hours of audio. Most music sounds pretty good at MidFi Stereo and most people's stereos cannot reproduce the difference between Hi-Fi and CD-DA. CD-ROM XA has included the Lo-Fi and Mid-Fi modes. I do not know anything about DVI audio. Now for video. A CD, played start to finish, holds 72 minutes of data at 75 sectors per second. Data (text) sectors hold 2048 bytes. Video sectors hold 2324 bytes because they don't contain extra EDC/ECC. Therefore: (72 * 60 * 75 * 2324)/1024 = 735 Meg of video data or (72 * 60 * 75 * 2048)/1024 = 648 Meg of text/program data Full Motion Video on CD-I or DVI uses the full bandwidth of the disc, thus appx 72 minutes of FMV on a CD. CD-I and DV-I systems can use hard disks as well as optical discs for their data storage. In fact, most good hard discs are faster than CDs and there are now some quite affordable HDs that hold more than CDs. CD-I players can read CD-DA, CD-ROM and CD-ROM XA discs. In addition, it takes only a small amount of extra circuitry to handle CD+G or CD+Midi. Several manufacturers are planning these extensions from the beginning. Finally, CD-I is being used today. There are several companies around the world using CD-I for data archival, training, Point-of-Info or Point-of-sales. Although it is *primarily* intended as a consumer technology and it will be hard-pressed to fight the IBM/Mac lock on business systems, it will find some uses in dedicated systems used by businesses. The Wall Street Journal recently carried an article about the Tokyo Electronics Show in Japan where 33 exhibitors were showing CD-I players or software. Over 10 companies are manufacturing players including a hand-held model by Sony that uses a 5" LCD screen or connects to your favorite TV. Personally, I will feel better going out to buy a system that has competition from ten manufacturers than one that is offered only by Intel or Commodore. There will certainly be a much greater force to lower prices and add features. I don't know if I have actually cleared up the discussion at all, but I feel like this was information that needed to be presented. Eric Miller * If this is paradise, I need a lawn mower. Manager, New Media Systems * Microware Systems Corp * - David Byrne
a544@mindlink.UUCP (Rick McCormack) (12/11/90)
in an article, Eric Miller, <4310@mcrware.UUCP> gives some insight into the capacities of several of the CD formats available and soon to be available. I attempted to assemble much of this information from manufacturers and trade magazines, but Mr. Millers summation varies in some significant aspects from my research. I am more than willing to believe his figures than those published in trade articles and magazines - the man makes his living by creating and using these suckers, while reporters make their living by writing about things their editors tell them to. :-) Thanks to Mr. Miller. I have saved his message (I promise not to distribute it in any "for Pay" media) and know of several of my acquaintances who will be glad to see info from the right end of the horse for a change.