dtynan@altos86.Altos.COM (Dermot Tynan) (07/11/90)
Sony has a format for the storage and interchange of motion-picture images (as I understand it). I read a quick blurb on it a while back, and am interested in knowing the format, and if any other organizations support it at this time. The impression I got, was that you took a 35mm film, and converted each frame to this "D2" format, using some very high-resolution digitizer. You then take this whatever, load it onto the graphics workstation of your choice, edit it a will, and finally send the finished (tape??) to Kodak, or wherever, where it will be converted back to 35mm celluloid. Does anyone have a simple breakdown of this format, or know anything about it?? I'm even curious as to what kind of tape they use, or even if it *is* tape. Any help appreciated... - Der -- Dermot Tynan, Altos Computer Systems, San Jose, CA 95134 dtynan@altos86.Altos.COM (408) 432-6200 x4237 "Five to one, baby, one in five. No-one here gets out alive."
buck@drax.gsfc.nasa.gov (Loren (Buck) Buchanan) (07/11/90)
In article <3714@altos86.Altos.COM> dtynan@altos86.Altos.COM (Dermot Tynan) writes: >Does anyone have a simple breakdown of this format, or know anything about >it?? I'm even curious as to what kind of tape they use, or even if it *is* >tape. Any help appreciated... Yes it is tape. Since you posted to this group, I assume what you are looking for is a way to generate animations to digital video. The questions to be asked are: 1. Is anyone marketing a D2 machine with a computer interface for creating animations? 2. What are the specifications for D2? Is it component or composite? How many bits per sample? Etc.? My own interests lie elsewhere, so here are a few more questions: 3. Is anyone marketing a D2 machine with a computer interface which is useable for both video work, and general purpose computer work (backups, data storage, etc.)? 4. How is audio handled? How many samples per second, bits per pixel, is it stereo, etc.? 5. Questions 1-4 with D1 substituted for D2. B Cing U Buck Loren Buchanan | buck@drax.gsfc.nasa.gov | #include <std_disclaimer.h> CSC, 1100 West St. | ...!ames!dftsrv!drax!buck | typedef int by Laurel, MD 20707 | (301) 497-2531 | void where_prohibited(by law){} CD International lists over 40,000 pop music CDs, collect the whole set.
tell@oscar.cs.unc.edu (Stephen Tell) (07/17/90)
In article <2751@dftsrv.gsfc.nasa.gov> buck@drax.UUCP (Loren (Buck) Buchanan) writes: >In article <3714@altos86.Altos.COM> dtynan@altos86.Altos.COM (Dermot Tynan) writes: >>Does anyone have a simple breakdown of this format, or know anything about >>it?? I'm even curious as to what kind of tape they use, or even if it *is* >>tape. Any help appreciated... > >Yes it is tape. I think the tape is physicaly 3/4" wide. The cassetes get rather large for long playing times. >2. What are the specifications for D2? Is it component or composite? How > many bits per sample? Etc.? D2 is digitized NTSC composite; 8 bits sampled at 14.31818. >My own interests lie elsewhere, so here are a few more questions: > >3. Is anyone marketing a D2 machine with a computer interface which is > useable for both video work, and general purpose computer work (backups, > data storage, etc.)? To quote the frequently asked questions posting: "Uhhhhh..." At >$60K for D2 and >$100K for D1, you might be better off with a dedicated tape drive. It would store a lot though, but the computer interface would have to do its own error correction (see below). >4. How is audio handled? How many samples per second, bits per pixel, is > it stereo, etc.? Audio is digital in both formats, I think 4 channels, "CD quality" (14 bits at 44.1Khz). Even though the audio is recorded by the helical scan rotary head, audio and video can be edited separately using a read-erase-write set of heads that fly by in that order. >5. Questions 1-4 with D1 substituted for D2. D1 is digital component; 8 bits for Y, R-Y, and B-Y. Y sampled 13.5Mhz, the others at half that. >Loren Buchanan | buck@drax.gsfc.nasa.gov | #include <std_disclaimer.h> For more info, here's a posting from several months ago from rec.video, in which someone with more knowledge corrected a posting of mine and explained things better than I would have anyway. From thorin!mcnc!rutgers!mit-eddie!snorkelwacker!apple!gem.mps.ohio-state.edu!wuarchive!texbell!texsun!newstop!sun!vector!poynton Sat Nov 18 16:22:20 EST 1989 Article 7012 of rec.video: Path: thorin!mcnc!rutgers!mit-eddie!snorkelwacker!apple!gem.mps.ohio-state.edu!wuarchive!texbell!texsun!newstop!sun!vector!poynton >From: poynton@vector.Sun.COM (Charles A. Poynton) Newsgroups: rec.video Subject: D-1 and D-2 Message-ID: <127857@sun.Eng.Sun.COM> Date: 14 Nov 89 09:22:33 GMT References: <8911071850.AA05885@nsipo.arc.nasa.gov> <10410@thorin.cs.unc.edu> Sender: news@sun.Eng.Sun.COM Reply-To: poynton@sun.com (Charles A. Poynton) Organization: Sun Microsystems, Mountain View Lines: 69 Was: Re: LD "Best Practices" Sumary: D-1 is compoNENT digital 27 MB/s, D-2 is compoSITE digital 14.3 MB/s. Regarding speculation on the details of D-1 and D-2, Stephen Tell did a pretty good job of explaining, here's corrections and some more detail. [pardon the stilted style, this is edited 'vi'] Both D-1 and D-2 are 100% digital. (Well, OK, there's a scratch audio track, but one hopes that it will never be heard outside the post suite.) D-1 came first. It is a COMPONENT digital format, according to CCIR Rec. 601: YUV coding [luminance and two colour differences, Y=0.299R+0.587G+0.114B, U=(.5/.886)(B-Y), V=(.5/.701)(B-Y)], so-called 4:2:2, which means that the U and V components are horizontally subsampled 2-to-1. Luminance is sampled at 13.5 MHz, 720 samples per picture width. Aggregate data rate is roughly 27 MB/s (megabytes per second). The D-1 format was standardized back in the days that the industry thought it would make the transition from composite-analog to component-digital in one swell foop. No such luck. Ampex identified an opportunity to make a quick buck, whoops, do the industry a service by providing digital capability at a lower price, and worked up the D-2 format. D-2 uses exactly the same physical casette s D-1, but the recording format is entirely different. D-2 is COMPOSITE NTSC digital: digitized NTSC. The composite signal is sampled at four-times-colour-subcarrier, i.e. about 14.318 MHz. Aggregate data rate is, of course, 14.318 MB/s. All the impairments of NTSC, but with the reliability and generation performance of digital. Great for "on-air" use where the user is stuck with NTSC whether he likes it or not; not-so-great for post production applications where the user really wants separate components for production flexibility. Not only is the NTSC footprint permanently imprinted on a D-2 recording, but the luminance range from black to white encompasses only about seven bits [the remaining codes being required for sync and chroma excursions], as opposed to very nearly eight bits for D-1 (well, OK, 235 codes of 256). Oh yeah, almost forgot, 757 samples per picture width, 768 "active" but a few are blanking. There are three sizes of D-type cassettes, called S, M, and L (use your imagination to figure THAT out!). D-1 gets [I think] 22, 40, and 90 minutes, respectively. D-2 has half the data rate, half the tape speed, therefore twice the capacity on a tape; a maximum of about three hours or about 150 GB (gigabytes) per tape. Neither D-1 nor D-2 is totally error-free like a computer tape, however. Error correction techniques are used to record redundant information to enable perfect reconstruction of small numbers of error bits; large error bursts can be detected but not corrected. So, the samples written on tape are dispersed spatially ("shuffled") so that a single error burst affects isolated single samples dispersed spatially throughout a frame. The resulting error samples are concealed, that is, replaced by information interpolated from their neighbours. It is this error concealment which ultimately limits the generation performance of the tape. Pardon me for rambling, this is probably much more than you needed to know. On the other hand, this technology will be arriving on your doorstep in the form of a consumer digital VCR quite soon, and so you rec.video enthusiasts might as well learn about it now. Maybe since the wall is now being torn down, we can turn some of our defense budget toward television system R&D in America! Personal views only, of course, C.
bobe@videovax.tv.tek.com (Bob Elkind) (07/17/90)
The previous articles explaining D1 and D2 formats were pretty good, this article contains some corrections and amplifications: D1 and D2 are both 10-bit video standards. They had originally been 8-bit, but the Society of Motion Picture and Television Engineers (aka SMPTE) has fixed that. Much of the professional equipment out there supporting D1 and D2 are 10-bit boxes (frame synchronizers, signal generators, switchers, etc.). Tape recorders are 8-bit devices (I believe) in general. 8-bit video is represented in 10-bit D1 and D2 formats by setting the two LSBs to zero. D1 comes in two flavours: NTSC and PAL. Both flavours are indeed component video sampled at 13.5 MHz for Luminance (the information to which human eyes are most sensitive) and 6.75 MHz for the two Color-difference channels. NTSC flavour of D1 is still 262.5 lines per video frame, two frames per field, 59.94 Hz video frame rate. PAL flavour D1 is 312.5 lines per frame, two frames per field, 50 Hz video frame rate. So you see, D1 is still quite different between the NTSC and PAL standards. D2 is also in NTSC and PAL flavours. D2 is basically digitized baseband (composite) video sampled at 4 times the colour sub-carrier frequency (appr 3.58 and 4.43 Mhz for NTSC and PAL, respectively), or 14.3818 and 17.8 Mhz respectively. Baseband composite video is pretty much what comes out of the RCA video connector on the rear panel of your VCR. Both D2 bandwidths are much lower than the 27Mhz D1 data bandwidth (13.5 MHz + 6.75 Mhz + 6.75 Mhz), and therefore the tape storage costs are lower for the composite standards. This is a significant operating cost factor! By the way, D1 format is also called 4:2:2 by many manufacturers, particularly Sony. Serial transmission formats for D1 and D2 are being formalized at this time by a SMPTE committee. The data format is basically the same as the basic non-serial D1 and D2 formats, but some extra information is inserted at every horizontal line, provisions for conveying audio information in the same signal have been made, and provisions for miscellaneous (auxiliary) data have also made. Serial video is still 10-bit data,and since it is bit-serial the transmission bandwidths are 10x that of non-serial (or parallel) D1 and D2. This is not you basic RS-232 serial interface. I hope this exposition has been somewhat useful. There's a lot of interesting folklore, politics, and humor in the developmental history of video standards; and there is much more of the same going on right now (anyone tracking HDTV developments out there knows what I mean).
dmc@videovax.tv.tek.com (Donald M. Craig) (07/17/90)
For more information about D1/D2 formats, and many other digital television topics besides, I recommend: The Art of Digital Video John Watkinson Focal Press, 1990 ISBN 0-240-51287-1 The companion volume is also excellent: The Art of Digital Audio John Watkinson Focal Press, 1988/1989 ISBN 0-240-51270-7 Focal Press books can be ordered from: Butterworths 80 Montvale Ave. Stoneham, Ma. 02180-2422 (617) 438-8464 FAX: (617) 438-1479 cheers, Don Craig Tektronix Television Division