James_Hastings-Trew@tptbbs.UUCP (James Hastings-Trew) (06/09/91)
In a message dated Sat 08 Jun 91 05:11, Mikep@hpmwtd.hp.com (mike Powell) wrote: MP> I still think the idea of supplying DCTV with CDTV is a great MP> one.... the quantity/quality/speed of that type of system MP> for things like 'full motion video' (whatever that is :-) would MP> be a BIG plus... although I have yet to see DCTV anims made MP> from successive digitized video frames.... MP> How well does DCTV do with such a task? I tried it out with a laser disk of computer animation. I used the DCTV software to grab each image as a non-interlaced, 3 bitplane image. In the universe of DCTV, 3 bitplanes means that your colours are limited to 8 million or so instead of 16 million. On full motion image, the reduced colour-set is invisible - it IS visible on stills or slowing moving images with lots of large areas of graduated colour as banding. Using standard Amiga anim tools I was able to produce about 5 seconds of full motion, 30 frame per second video with a 200K size anim file. Mind you, all normal ANIM rules apply - if there is too much difference between frames the delta information can get to be larger than the actual frame information. Still, it was neat to see this work on my A2500. I believe that GVP demonstrated about 3 minutes of full-motion video from Back To The Future III running from a hard-drive using the DCTV unit. DCTV frames present dramatic compression of images, mainly by reducing image information down to what would be visible on a standard NTSC composite display. A single DCTV frame can be 70K compared to the 24 bit original file which can easily be over 1 megabyte of information. I once created a DCTV anim which contained 60 frames rendered in Imagine at 736*480 in 24 bit colour. Each frame was 1.2 megs in size - which reduced down to about 70K. The entire anim ended up at 500K - less than half the size of one of the original frames. There is LOTS of potential here.
bobl@graphics.rent.com (Bob Lindabury - SysAdm) (06/10/91)
James_Hastings-Trew@tptbbs.UUCP (James Hastings-Trew) writes: > I believe that GVP demonstrated about 3 minutes of full-motion video from > Back To The Future III running from a hard-drive using the DCTV unit. DCTV > frames present dramatic compression of images, mainly by reducing image > information down to what would be visible on a standard NTSC composite > display. A single DCTV frame can be 70K compared to the 24 bit original file > which can easily be over 1 megabyte of information. I once created a DCTV > anim which contained 60 frames rendered in Imagine at 736*480 in 24 bit > colour. Each frame was 1.2 megs in size - which reduced down to about 70K. > The entire anim ended up at 500K - less than half the size of one of the > original frames. There is LOTS of potential here. Actually, the demo wasn't GVP. It was done by IVS. It was Back To The Future III running from an IVS Grand Slam SCSI drive controller and a Seagate ST2383N 1.2 Gigabyte drive connected to an A3000 25mhz with 6 meg of ram. The video was played back through the DCTV in 3-bit color resolution in full overscan at 24 frames per second (true full motion video is 30 fps) with a video bandwidth of 12.06 mbits/sec. The audio was full 16-bit per channel digital with 44.1 or 48 Khz samples and a bandwidth of 1.54 Mbits/sec. The sound was 4 channel surround sound and the total runing time of the piece was 4 minutes and 30 seconds. Quite an impressive demo I assure you. You can easily see why it would be benificial to the Amiga community for a crossover DCTV product to be produced. The image while moving was *very* good. It was difficult to tell it was actually playing off the computer and not a laser disk. The image was *much* better than standard VHS tape as the actual resolution surpassed that of conventional VHS and of course, the audio was digital. -- Bob The Graphics BBS 908/469-0049 "It's better than a sharp stick in the eye!" ============================================================================ InterNet: bobl@graphics.rent.com | Raven Enterprises UUCP: ...rutgers!bobsbox!graphics!bobl | 25 Raven Avenue BitNet: bobl%graphics.rent.com@pucc | Piscataway, NJ 08854 Home #: 908/560-7353 | 908/271-8878
mikep@hpmwtd.HP.COM (Mike Powell) (06/11/91)
Damn! I sure wish I had a chance to see that BTTFIII demo! The concept of a DCTV output only module for CDTV would be fantastic! I don't have either of these units, but after talking to the DCTV folks about what their product did and how it accomplished it, I immediately came up with the CDTV/DCTV combo idea... and it still seems like a sure-fire winner to me. I just read in another message about a 'rumor' that says Commodore is looking into a DCTV output module for CDTV... I sure hope so. With the space available on each ROM, it seems to me like there would be enough room for TWO sets of graphics most applications... one being normal CDTV output, and the other in DCTV format... selectable by the user as an option. Or better yet, to detect the module and automaticaly configure for it! But since it isn't hanging off of a Zorro slot, there doesn't seem like it would be possible... CBM... are you listening???? :-) -Mike-
kdarling@hobbes.catt.ncsu.edu (Kevin Darling) (06/12/91)
> Damn! I sure wish I had a chance to see that BTTFIII demo! > The concept of a DCTV output only module for CDTV would be > fantastic! [...] > I just read in another message about a 'rumor' that says > Commodore is looking into a DCTV output module for CDTV... Whoa :-). First off, a custom demo running on a 25Mhz Amiga 3000 with a 1-2M/sec hard disk might not apply to a 7Mhz CDTV with 150K/sec CDROM. Secondly, the output module is rumored to be for viewing still shots from Photo-CD discs. OTOH, since P-CD isn't due until summer of '92, perhaps they _do_ intend something for the gizmo before then. > With the space available on each ROM, it seems to me like there would > be enough room for TWO sets of graphics most applications... > one being normal CDTV output, and the other in DCTV format... Yes, that certainly would be possible. best - kevin <kdarling@catt.ncsu.edu>
moonhawk@bluemoon.uucp (David Culberson) (06/13/91)
kdarling@hobbes.catt.ncsu.edu (Kevin Darling) writes: > a 1-2M/sec hard disk might not apply to a 7Mhz CDTV with 150K/sec CDROM. This isn't really refuting what you said, it's true, but I read that the CDTV's CD-ROM drive has burst modes of up to FIVE megs a second! Is this true? If so, you could have like a 3 or so second anim for something like a text adventure or something for each place you went, in full 24 bit color, with the DCTV outpput module. It would be AWESOME! If this is not true, don't tell me. Just kidding, tell me no matter what... David Now the world has gone to bed, MoonHawk@Bluemoon.uucp ////|all Darkness won't engulf my head, moonhawk%bluemoon@nstar.rn.com //// |hail I can see by infrared, \\\\///__|the How I hate the night. Yeah, this IS an annoying SIG. \\\\/ |miga
kdarling@hobbes.catt.ncsu.edu (Kevin Darling) (06/13/91)
moonhawk@bluemoon.uucp (David Culberson) writes: > > This isn't really refuting what you said, it's true, but I read >that the CDTV's CD-ROM drive has burst modes of up to FIVE megs a second! >Is this true? If so, you could have like a 3 or so second anim for >something like a text adventure or something for each place you went, in >full 24 bit color, with the DCTV outpput module. It would be AWESOME! If >this is not true, don't tell me. Just kidding, tell me no matter what... I've also seen a CDTV drive spec which has "2 megabytes (burst)" on it. But as cdrom drives always read back at 75 sectors/sec (sectors approx 2K), then the only thing I can think of for such an number to mean, could be how fast two adjacent bytes might be available for a DMA read, at times. In other words, no, I can't believe in a 5MB/sec rate sustained for more than a few microseconds (unless the drive has a cache, and that's likely to be too tiny to be of anim use). best - kev <kdarling@catt.ncsu.edu> Sidenote: I thought the DCTV makers only claim 4 million colors, which is 22 bits. And a competitor's checking seems to show that the color changes are limited per line, and only apply to every other line (altho luminance changes can occur on every line). In any case, it's definitely not true "24-bit output". Probably better to just claim the total color range :-)
James_Hastings-Trew@tptbbs.UUCP (James Hastings-Trew) (06/15/91)
In a message dated Thu 13 Jun 91 09:07, Kdarling@hobbes.catt.ncsu.edu (kevi wrote: K> Sidenote: I thought the DCTV makers only claim 4 million colors, which K> is K> 22 bits. And a competitor's checking seems to show that the color K> changes K> are limited per line, and only apply to every other line (altho K> luminance K> changes can occur on every line). In any case, it's definitely not K> true K> "24-bit output". Probably better to just claim the total color range K> :-) Digital Creations does not like to quote resolutions and colours, as their device outputs a composite video signal, no more, no less. The colours are NOT "pixel" addressable. For a competitor to compare DCTV to a true frame buffer or other "pixel mode" box is unfair - it is comparing apples and oranges. In practice, the luminance changes for each pixel displayed (within the limits of composite video) but the chroma changes much more slowly. At a guess I would say that the chroma is at 1/2 to 1/3 the resolution of the luminance portion of the signal. If you want to see this in action, take a black screen, and draw a RED filled circle on the screen. You should see some MAJOR stairstepping taking place at the edge of the object. Now put a GREY circle on the screen. No stairsteps. The DCTV device can get away with this because this is actually a common video situation - VHS tapes from most consumer VCRs are like this. With a normal "picture" image the effect is invisible. For computer-generated imagery, all it takes is some care in the image creation to avoid extreme situations. Taken for what it is, the DCTV unit is a great box. Don't take it for what it isn't - a 24 bit display.