harald@ccicpg.UUCP ( Harald Milne) (10/16/87)
Scan doublers don't sound cheap, although they sound like the solution
that discerning users like me are looking for. To make RGB using 4-bit ADC's
is ok, but storing that information to the next frame, so it can be dumped
all at once, sounds expensive. Am I missing something here?
--
Work: Computer Consoles Inc. (CCI), Advanced Development Group (ADG)
Irvine, CA (RISCy business! Home of the CCI POWER 6/32)
UUCP: uunet!ccicpg!haraldhedley@cbmvax.UUCP (Hedley Davis) (10/17/87)
Scan doubler:
A device which converts video by repeating each displayed line
twice ( at twice the speed ). This acts to enhance the displayed
image by "filling in the gaps between the scanned lines". Gives
a nice solid appearence, but DOES NOT add resolution. Requires
one display lines worth of memory. Really very cheap ( not
counting the monitor ).
Scan Converter:
A device which converts video from one format to another.
Typically used for deinterlacing picures. Output video rate
is twice that of a normal display. Doubles vertical
resolution. Requires at least one feilds worth of ram.
Not so cheap ( due to memory cost ).
I cannot speak for us ( this IS ironic ), but I know that certain
third parties are pursuing the scan conversion problem for the Amiga
market. Such devices are planned for the A2000, and possibly the
A500/A1000 as well. I have seen prototypes, and they basically
give you a non-interlaced 640x400 display. For the A500/A1000 units,
vendors will probably choose to do a 4 bit/16 color conversion using
the digital RGBI outputs on the video connector. ( High speed ADCs
are expensive, and the analogue stuff would be tricky indeed, as
opposed to a 90% digital solution ). For the B2000 units, the
internal video slot has all twelve bits of digital video avilable,
so the full color solution remains largely digital ( easy ).
The nice thing about scan converters is that they are typically
100% software compatible ( as the software could not tell if
they were there if their lives depended on it ).
Such devices are highly reccomended.
Hedleyminer@dino.cpe.ulowell.edu (Rich Miner) (10/22/87)
In article <2530@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: >Scan Converter: > A device which converts video from one format to another. > Typically used for deinterlacing picures. Output video rate > is twice that of a normal display. Doubles vertical > Resolution. Requires at least one feilds worrh of ram. > Not so cheap ( due to memory cost ). There _will_ be a scan converter/deinterlacing device on the market in the near future (Jan/Feb). One will be on display at Comdex in the dealers show and perhaps in CBM's booth. >third parties are pursuing the scan conversion problem for the Amiga As Hedley points out, it is not a Commodore product, it is being done by a third party in the Boston area. >For The B2000 units, the internal video slot has all twelve bits of >digital video avilable, To be 100% color compatable, the initial version is for the B2000. The card will go in the video slot and require a multi-scan monitor. >Such devices are highly reccomended.I I saw the system many months ago as a 16-color A1000 solution. The results are hard to describe, it produces the cleanest crispest screen you can imaging!! It should transforms the A2000 into a _much_ more professional appearing workstation for CAD and desk-top applications. -- Rich miner@ulowell.edu 617/452-5000x2693 ULowell CPE Imaging Research Lab
spierce@pnet01.cts.com (Stuart Pierce) (10/23/87)
I noticed in InfoWorld that Seiko has announced a new multi-scanning monitor (up to 43 KHz) that supports 1024 x 768. Retail price will be $1095. Stuart Pierce
harald@ccicpg.UUCP ( Harald Milne) (10/23/87)
In article <1774@dino.cpe.ulowell.edu>, miner@dino.cpe.ulowell.edu (Rich Miner) writes: > In article <2530@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: > >Scan Converter: > > A device which converts video from one format to another. > > Typically used for deinterlacing picures. Output video rate > > is twice that of a normal display. Doubles vertical > > Resolution. Requires at least one feilds worrh of ram. > > Not so cheap ( due to memory cost ). I would dicker with the not cheap comment. Ever heard of interleave? I also made this comment, until I realized that if you delay ram timings through interleaving, You don't need fast rams. > There _will_ be a scan converter/deinterlacing device on the market in the > near future (Jan/Feb). One will be on display at Comdex in the dealers show > and perhaps in CBM's booth. I definately wan't to see this! Guess I'm off to Comdex! > >third parties are pursuing the scan conversion problem for the Amiga > As Hedley points out, it is not a Commodore product, it is being done by a > third party in the Boston area. I looked at my obsolete A2000 schematics, and looked at the update leaflet on it. The video connector doesn't bother me, but the Extended video connector does. Raw audio? Is it the rummored removal of the audio cutoff filter? If so, will the form factor of this add on not interfere with the extended connector so some other card can exist in it's slot? Or will it pass/propogate these signals if it does? > >For The B2000 units, the internal video slot has all twelve bits of > >digital video avilable, > To be 100% color compatable, the initial version is for the B2000. The > card will go in the video slot and require a multi-scan monitor. So I can solve my no A2000 monitor problem by buying one now? If I have to, I'll get a SONY since SONY knows how to do overscan better than ANY company I know. You know, FULL picture to the square corners. > I saw the system many months ago as a 16-color A1000 solution. The results > are hard to describe, it produces the cleanest crispest screen you can > imaging!! It should transforms the A2000 into a _much_ more professional > appearing workstation for CAD and desk-top applications. > Rich miner@ulowell.edu 617/452-5000x2693 ULowell CPE Imaging Research Lab -- Work: Computer Consoles Inc. (CCI), Advanced Development Group (ADG) Irvine, CA (RISCy business! Home of the CCI POWER 6/32) UUCP: uunet!ccicpg!harald
hedley@cbmvax.UUCP (Hedley Davis) (10/24/87)
In article <3248@ccicpg.UUCP> harald@ccicpg.UUCP ( Harald Milne) writes: >In article <1774@dino.cpe.ulowell.edu>, miner@dino.cpe.ulowell.edu (Rich Miner) writes: >> In article <2530@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: >> >Scan Converter: >> > Not so cheap ( due to memory cost ). > > I would dicker with the not cheap comment. Ever heard of interleave? >I also made this comment, until I realized that if you delay ram timings >through interleaving, You don't need fast rams. > I meant that it was not cheap relative to a scan doubler. Yes I have heard of interleave. I think you need fast rams. If you can do it using low bandwidth rams ( IE normal DRAMS ( not video shift register rams )), for less cost than the neccesary number of video shift register rams, I am sure many people here as well as all the other computer companies would be very interested. Hedley PS: I'd be interested anyway. Maybe you'll show me a trick or tood.eER
brianop@well.UUCP (Brian McBee) (10/24/87)
You will probably see scan doublers from several parties in the near future- they are supposed to do wonders for your TV picture also. -- ------------------------------------------------------------------- What the eye beholds CI$ MCI: 72406.1363 And the heart covets, BIX PLINK: Brianop Let the hand boldly seize!. USENET: well!brianop
grr@cbmvax.UUCP (George Robbins) (10/25/87)
In article <3248@ccicpg.UUCP> harald@ccicpg.UUCP ( Harald Milne) writes: > > I looked at my obsolete A2000 schematics, and looked at the update > leaflet on it. The video connector doesn't bother me, but the Extended > video connector does. Raw audio? Is it the rummored removal of the audio > cutoff filter? If so, will the form factor of this add on not interfere > with the extended connector so some other card can exist in it's slot? > Or will it pass/propogate these signals if it does? The capability to disable the audio filter is built-in on the current A500 and US A2000 boards. The extra video connector is just another connector in line with the orignal video connnector that allows you to pick off some more signals. The video connector is a separate slot that doesn't connect to the regular expansion bus. A card in the video slot can pick up one or both connectors. -- George Robbins - now working for, uucp: {ihnp4|rutgers|allegra}!cbmvax!grr but no way officially representing arpa: out to lunch... Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
farren@gethen.UUCP (Michael J. Farren) (10/25/87)
In article <2595@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: >I think you need fast rams. If you can do it using low bandwidth rams >( IE normal DRAMS ( not video shift register rams )), for less cost >than the neccesary number of video shift register rams, I am sure >many people here as well as all the other computer companies would be >very interested. It's definitely possible, and, in fact, is easy. 120ns? Hell, you could probably use 250ns RAM with ease. Remember, you are not con- strained by requiring the CPU to be able to access the RAMs, so you could design them as a 6-bit deep array, and load/read all six bit planes at once. This would reduce your bandwidth by six times; the scan conversion would use twice the bandwidth; net result, three times less bandwidth required than for "normal" video. The difficulty with this scheme is that you have to use a lot more RAM chips, but you get paid back by the fact that you could use smaller RAMs, which are con- siderably more than four times cheaper than either video RAMs or 60ns dynamics. You could even use CMOS RAM and save a lot of power and heat. Don't know what the tradeoffs would be, exactly, but I'm willing to bet that you could design a winning card like this. -- ---------------- Michael J. Farren "... if the church put in half the time on covetousness unisoft!gethen!farren that it does on lust, this would be a better world ..." gethen!farren@lll-winken.arpa Garrison Keillor, "Lake Wobegon Days"
harald@ccicpg.UUCP ( Harald Milne) (10/26/87)
In article <2595@cbmvax.UUCP>, hedley@cbmvax.UUCP (Hedley Davis) writes: > In article <3248@ccicpg.UUCP> harald@ccicpg.UUCP ( Harald Milne) writes: > >In article <1774@dino.cpe.ulowell.edu>, miner@dino.cpe.ulowell.edu (Rich Miner) writes: > >> In article <2530@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: > >> >Scan Converter: > >> > Not so cheap ( due to memory cost ). > > > > I would dicker with the not cheap comment. Ever heard of interleave? > >I also made this comment, until I realized that if you delay ram timings > >through interleaving, You don't need fast rams. > > > I meant that it was not cheap relative to a scan doubler. > Yes I have heard of interleave. Sorry, I didn't know I was talking to a CBM engineer. This discussion got quite threaded. Sorry Mr Davis. > I think you need fast rams. If you can do it using low bandwidth rams > ( IE normal DRAMS ( not video shift register rams )), for less cost > than the neccesary number of video shift register rams, I am sure > many people here as well as all the other computer companies would be > very interested. I believe I have that solution using slow 256k drams. I don't care about other companies, just the Amiga. The solution I have in mind only pertains to the A2000. > > Hedley > > PS: I'd be interested anyway. Maybe you'll show me a trick or > too.... According to my calculations, the pixel rate (the time that digital RGB information has to arrive from ram) is not a factor. If it is, I have got that covered anyway, and very simply, using 2 way interleave. The problem as I understand it, is the need for dual port video rams. When displaying this uninterlaced information, it is necessary to simulatuneously display from the very same ram that is being written to. Thus the need for dual port rams. By using additional buffering in extra cheap ram, this problem is solved along with the need to use dual port drams. Thus a lower cost. This is of course at the expense of PC real estate, to have the extra cheap chips on a PC card. I think that this solution also solves another problem, the timing of this mess, address generation, etc. Also by my calcutations, the problem splits very neatly. I takes less than 512k bytes of ram to store an entire Hi-res digital RGB encoded picture, including max overscan. The cost for cheap 256k drams for 512k of memory is less than $100.00. That's $200 for 2 stored deinterlaced frames of RGB. See what I am getting at? -- Work: Computer Consoles Inc. (CCI), Advanced Development Group (ADG) Irvine, CA (RISCy business! Home of the CCI POWER 6/32) UUCP: uunet!ccicpg!harald
daveh@cbmvax.UUCP (Dave Haynie) (10/27/87)
in article <3248@ccicpg.UUCP>, harald@ccicpg.UUCP ( Harald Milne) says: > Keywords: Interlace, monitors, scan doublers > > In article <1774@dino.cpe.ulowell.edu>, miner@dino.cpe.ulowell.edu (Rich Miner) writes: >> In article <2530@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: >> >Scan Converter: >> > [...] Not so cheap ( due to memory cost ). > > I would dicker with the not cheap comment. Ever heard of interleave? > I also made this comment, until I realized that if you delay ram timings > through interleaving, You don't need fast rams. I think maybe you do. Consider for a moment the Amiga. At 640 across by 4 deep, the Agnus chip is fetching data for Denise as fast as a 3.58MHz DRAM can supply it, or real close. If you want to double the scan rate (why you're building this thing in the first place), you're now fetching video information at 7.16MHz. Or, put another way, one fetch every 158ns. And that's not leaving any time to put the data in there, via interleaving. Now you're down to about 79ns. I think the video DRAM will be cheaper than the high speed static RAM you'd need for this otherwise. >> >third parties are pursuing the scan conversion problem for the Amiga >> As Hedley points out, it is not a Commodore product, it is being done by a >> third party in the Boston area. > > I looked at my obsolete A2000 schematics, and looked at the update > leaflet on it. The video connector doesn't bother me, but the Extended > video connector does. Raw audio? Is it the rummored removal of the audio > cutoff filter? Yes and no. The /LED line is used on the A2000 and A500 to supply audio without the filtering under program control, for those applications that know they're using the higher sampling rates and possibly external filtering. The A2000's RAW AUDIO outputs on the video connector are in fact the unfiltered signals. These are always unfiltered, regardless of the /LED setting. The first connector contains the lines are are filtered or unfiltered based on the /LED line. > If so, will the form factor of this add on not interfere > with the extended connector so some other card can exist in it's slot? > Or will it pass/propogate these signals if it does? Huh? There's ONE video slot, with two connectors in line with each other. Only one video card goes in any machine, and a video card will not conflict with any other expansion device; it's on the other side of the machine. > Work: Computer Consoles Inc. (CCI), Advanced Development Group (ADG) > Irvine, CA (RISCy business! Home of the CCI POWER 6/32) > UUCP: uunet!ccicpg!harald -- Dave Haynie Commodore-Amiga Usenet: {ihnp4|caip|rutgers}!cbmvax!daveh "The B2000 Guy" PLINK : D-DAVE H BIX : hazy "Computers are what happen when you give up sleeping" - Iggy the Cat
hedley@cbmvax.UUCP (Hedley Davis) (10/27/87)
In article <3290@ccicpg.UUCP> harald@ccicpg.UUCP ( Harald Milne) writes: > > Sorry, I didn't know I was talking to a CBM engineer. This discussion >got quite threaded. Sorry Mr Davis. > You can talk to us anyway you please, but this mister stuff has gotta go. > [ lots of technical stuff suggesting technique for scan conversion > using inexpensive ram buffers in some innovative fashion ] > > See what I'm getting at ? > ( Well, not really, but I'll send you email so you can explain this to me privately so that I won't lose face on the net. ) :-) Seriously, discussion moved to email to reduce net.bandwidth. Interested parties can contact me for copies of correspondence. Hedley
harald@ccicpg.UUCP ( Harald Milne) (10/28/87)
In article <2616@cbmvax.UUCP>, daveh@cbmvax.UUCP (Dave Haynie) writes: > in article <3248@ccicpg.UUCP>, harald@ccicpg.UUCP ( Harald Milne) says: > > If so, will the form factor of this add on not interfere > > with the extended connector so some other card can exist in it's slot? > > Or will it pass/propogate these signals if it does? > > Huh? There's ONE video slot, with two connectors in line with each other. > Only one video card goes in any machine, and a video card will not conflict > with any other expansion device; it's on the other side of the machine. This almost answers my question. So in your mind there is a video card that uses both slots? The video connector slot has enough signals to make a composite video signal, correct? I thought just to do this would take up only the video connector. This is what I meant by form factor. You could have 2 cards in these slots, except the extended video connector has no power. This means you have to be careful of what you put in these slots. And this is my concern. It also appears that there is a byte wide bus available. Are A2000 genlocks also using these connectors? -- Work: Computer Consoles Inc. (CCI), Advanced Development Group (ADG) Irvine, CA (RISCy business! Home of the CCI POWER 6/32) UUCP: uunet!ccicpg!harald
grr@cbmvax.UUCP (10/29/87)
In article <3373@ccicpg.UUCP> harald@ccicpg.UUCP ( Harald Milne) writes: > In article <2616@cbmvax.UUCP>, daveh@cbmvax.UUCP (Dave Haynie) writes: > > in article <3248@ccicpg.UUCP>, harald@ccicpg.UUCP ( Harald Milne) says: > > > If so, will the form factor of this add on not interfere > > > with the extended connector so some other card can exist in it's slot? > > > Or will it pass/propogate these signals if it does? > > > > Huh? There's ONE video slot, with two connectors in line with each other. > > Only one video card goes in any machine, and a video card will not conflict > > with any other expansion device; it's on the other side of the machine. > > This almost answers my question. So in your mind there is a video card > that uses both slots? The video connector slot has enough signals to make a > composite video signal, correct? I thought just to do this would take up only > the video connector. This is what I meant by form factor. You could have > 2 cards in these slots, except the extended video connector has no power. > This means you have to be careful of what you put in these slots. And > this is my concern. It also appears that there is a byte wide bus available. > Are A2000 genlocks also using these connectors? No, the idea is that there is one slot, which in the original A2000 design had one edge card connector that basically picked up the same signals as were present on the external 23-pin video connector. When we revised this design to incorporate the newer custom chips developed for the A500, we realized that we hadn't been very imaginative, and that there were many interesting things that could be done if you had access to all the digital RGB bits, some clock and control signals. We then added a second connector so that a single video card could pick up either the basic or the basic and extended video signals. There was no intention that the slot should be occupied by two "half-cards", it not clear what utility there would be with such a setup. Normally, this slot would be empty or occupied by a composite color encoder card. If you put some interesting new device in the slot, you would have to remove the color card, but it the card is a genlock or something similar, it may provide its own composite output. -- George Robbins - now working for, uucp: {ihnp4|rutgers|allegra}!cbmvax!grr but no way officially representing arpa: out to lunch... Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
mph@rover.UUCP (Mark Huth) (10/29/87)
In article <2616@cbmvax.UUCP- daveh@cbmvax.UUCP (Dave Haynie) writes:
-in article <3248@ccicpg.UUCP>, harald@ccicpg.UUCP ( Harald Milne) says:
-> I would dicker with the not cheap comment. Ever heard of interleave?
-> I also made this comment, until I realized that if you delay ram timings
-> through interleaving, You don't need fast rams.
-
-I think maybe you do. Consider for a moment the Amiga. At 640 across by
-4 deep, the Agnus chip is fetching data for Denise as fast as a 3.58MHz
-DRAM can supply it, or real close. If you want to double the scan rate
-(why you're building this thing in the first place), you're now fetching
-video information at 7.16MHz. Or, put another way, one fetch every 158ns.
-And that's not leaving any time to put the data in there, via interleaving.
-Now you're down to about 79ns. I think the video DRAM will be cheaper than
-the high speed static RAM you'd need for this otherwise.
I think that's where the interleave comes in. Instead of fetching
sequential samples from the same ram chip, you fetch seqential samples
from different banks of ram. Maybe I have four ram banks, so each ram
is used every 316ns. With sequential addressing like you have in
video display applications the interleave technique works real well.
Just use some of the low order address bits to generate the chip
select signals.
Mark Huthcharles@hpcvcd.UUCP (10/30/87)
> > I would dicker with the not cheap comment. Ever heard of interleave? > > I also made this comment, until I realized that if you delay ram timings > > through interleaving, You don't need fast rams. > > UUCP: uunet!ccicpg!harald > > I think maybe you do. Consider for a moment the Amiga. At 640 across by > 4 deep, the Agnus chip is fetching data for Denise as fast as a 3.58MHz > DRAM can supply it, or real close. If you want to double the scan rate > (why you're building this thing in the first place), you're now fetching > video information at 7.16MHz. Or, put another way, one fetch every 158ns. > And that's not leaving any time to put the data in there, via interleaving. > Now you're down to about 79ns. I think the video DRAM will be cheaper than > the high speed static RAM you'd need for this otherwise. > > Dave Haynie Commodore-Amiga Usenet: {ihnp4|caip|rutgers}!cbmvax!daveh If you fetch 32bits at a time and deliver 4 bits per pixel, you only have to read your RAM at 895KHz, even after doubling the scan rate. With half of the time for CPU access, this leaves approx 500ns per access. Can you say cheap RAM? Charles Brown hplabs!hp-pcd!charles
hedley@cbmvax.UUCP (10/31/87)
In article <4410014@hpcvcd.HP> charles@hpcvcd.HP (Charles Brown) writes: >> > I would dicker with the not cheap comment. Ever heard of interleave? >> > I also made this comment, until I realized that if you delay ram timings >> > through interleaving, You don't need fast rams. >> > UUCP: uunet!ccicpg!harald > >If you fetch 32bits at a time and deliver 4 bits per pixel, you only >have to read your RAM at 895KHz, even after doubling the scan rate. >With half of the time for CPU access, this leaves approx 500ns per >access. Can you say cheap RAM? > Charles Brown hplabs!hp-pcd!charles Video comes out of the amiga in bursts during a line time, and goes away for retrace. I hope you weren't planning to save the data somewhere to run the rams full tilt during retrace also. The bandwidth during the video time is 14.32 Mbits per bit per pixel. ( four pixels per 3.58 mhz cycle in highres mode ) For 4 bits this is 57.28 Mbits/sec. Video to the monitor is twice this, or 114.56 Mbits/sec. Total bandwidth needed is 57.28+114.56 = 172.84 Mbits per sec. Assuming 32 bit wide memory, the memory needs to run at 172/32 = 5.375 Mwords/sec => 186 ns cycle time. How'd you get this 500ns number ? Also, what kind of drams you going to use ? Look at the memory layout. Consider further the amount of shifting you need. ( 32 bits wide x 4 bits in + 32 bits wide x 4 bits out => 256 bits of shift register storage ). Also, you need to time align all this stuff. You're gonna need even more bandwidth to achieve that. How many chips did you say this would take ? The goal here is inexpensive overall. Not inexpensive in RAM only. Hedley
farren@gethen.UUCP (10/31/87)
In article <2671@cbmvax.UUCP> hedley@cbmvax.UUCP (Hedley Davis) writes: >In article <4410014@hpcvcd.HP> charles@hpcvcd.HP (Charles Brown) writes: > >The bandwidth during the video time is 14.32 Mbits per bit per pixel. > ( four pixels per 3.58 mhz cycle in highres mode ) I assume you mean 14.32 Mbits/sec, instead of Mbits/bit. Other than that, you are mostly correct. Those four bits, though, are actually four bytes, read serially every 1.12 microseconds (four color clocks), and "written" to the screen, via the lookup tables and A/Ds, as eight pixels, each with twelve bits of color information. The net result is eight pixels written every 1.12 microseconds, or approximately 7.15 Megapixels/second (twice the color clock, or 140 ns/pixel). If you organize your scan converter ram as twelve parallel channels, one per bit of color information, then you only have to read one pixel every 140 ns. Use four-bit-wide RAM, and you only actually access the RAM every four pixels, therefore have 560 ns access time. You have to write them out at twice the rate, of course, so you actually have three pixels to deal with every 140 ns, which works out to a 186 ns access time, which is tight, but not unattainable with a little clever design. Use two four-bit-wide chips in parallel, and you lower your access time to 372 ns, and you can get away with 24 RAMs, each 64K X 4, at a total cost for RAMs of less than $75 when purchased in single quantities, and probably less than $50 in production quantity. This calculation done using 4464 chips, with pricing from Microprocessors Unlimited, not the cheapest supplier around. Other considerations, such as the cost of the shift registers, timing circuitry, A/Ds, and such will drive up the cost. This thing isn't going to be cheap, no matter how you do it. I figure a minimum manufacturing cost of $100 (probably optimistic), which translates to something around $400 retail cost. It'd make a good hacker's project, though - does C/A plan on selling prototyping boards for the video slot? -- ---------------- Michael J. Farren "... if the church put in half the time on covetousness unisoft!gethen!farren that it does on lust, this would be a better world ..." gethen!farren@lll-winken.arpa Garrison Keillor, "Lake Wobegon Days"