#Hi res monitors and dp
33 messages in this thread
I noticed in Grigley's column or somewhere in InfoWorld (just waiting for the
sub to exoire, gang) that J. Tramiel has some Japanese manufacurer churning out
VERY high res monitors at very low cost. And I vaguely remember someone like
Dave Haynie making reference to what could happen when high res monitors etc.
are more available. My question is, given an inexpensive meg bye meg pixel
monitor, what hardware change would be needed by an A-500 and an A-2000 to let
"serious" desktop publishing be done witht these machinses?
..Jim
Jim,
Humm…1 Meg x 1 Meg resolution would be 128Kbytes per raster line and
with million raster lines, that would be 128 Gigabytes! per bit plane…I think
you meant 1024 x 1024 resolution (1K x 1K) which is 128 bytes per raster line
or 128KBytes per bit plane.
To add that capability with the A2000/500 should only be a chip change
if and when the chip becomes available and presuming it will fit into the
current socket. A much different video output (sync rates, etc) would also be
needed….on the A2000 that shouldn't be a problem with the video slot, I'm not
sure about the A500 on the video output whether the current circuitry could or
couln't be modified.
Don
Don't think you'll see 1K * 1K with the current chipset – there is
not enough memory bandwidth to use the same display generator
with much more than 700 wide displays. To go to 1K you need
a separate display card.
How big a deal is changing the chipset? I think the monitor was actually
something like 1024 x 960 (sorry about the 1 meg by 1 meg howler.)
If such a monitor were cheaply available, how much more would be involved (in
money, labor, hassle, etc.) to upgrade the 500, the 1000, and the 2000 to being
able to use it? I guess Tramiel is thinking of CAD, but being able to see and
work on a whole page for Desktop Publishing would be very nice indeed.
// Jim
Changing the chipset can't do much without changing the basic
architecture of the Amiga, either increasing the bus width or
changing the way memory is accessed. Now the hi-res modes pretty
well max out the memory bandwidth, and increasing the number of
memory accesses can't be done without changing something major.
About the only options are going to less colors (it'd have to
be four colors to get "de-interlaced" hi-res), or to slow down
the display (which is what the new monitor does). My guess is
Commodore isn't willing to invest in a real long-range fix, but
will just go for a short term fix. Argh.
.
Regarding Tramiel's stuff, he is famous for showing things at
COMDEX that are not likely to ever ship, or if they do ship it
takes a long long time. Gotta commend Jack for at least taking
a long-term view of things, which Commodore hasn't done for the
past 18 months or so, but I wouldn't plunk down bucks for a
Tramiel machine until it's actually available in stores.
Tell you what — I _know_ my users could swallow a four color screen if it was
1024 by whatever. Now — are y'all ready for a NASTY hardware hack idea? Hmmm?
This has been done already here, to another machine, but it's all the same.
Here we go. You set up the display (as far as the AMiga knows) for 640x 400y
4bp. Now, this results in said video being applied to the color circuitry four
bits at a time. At this point, you steal it, jam it into a 2×2 shift register,
lock down the now-unused most or least significant two bits of the color
palette, and hum those four bits out as two pairs of two bits. Such a device
needs access to a dot clock… of some sort. It also needs access to a bit
somewhere that tells it to run or not run. Which _may_ be available in the
machine, or you could go positivly _nuts_ and make this thingee an auto
configure device, with appropriate register and all — put it some where the
COPPER could smack it in the teeth… and _really_ integrate it into the Amiga
properly. Should cost about 25-50 bucks to do right. This would give you 1280
by 400… nothing to sneer at, believe me. SoftCircuits did _exactly_ this kind
of thing for the radio shark color computer — we stole the incoming data to
the 6847 VDG and completely reformatted it for 128 by 256 16 32 grey level
video. If anyone is curious how, the schematic is called "CoCoGrey" and is in
the Scheme demo posted here. How about that? I would think that an end user
cost of <$100 should be attainable — and that it would then be totally
transparent to _any_ software that didn't specifically want to use it. If you
_DID_ want to use it, you could have 1280 by 400 or 1280 by 200 (talk about
mismatched x/y ratio!) without the proposed mod knowing or caring. You could
also implement extensions to this idea — like a local palette to control the
other two bits that get locked when the 4->2 bit conversion is made. Any
hackers out there wanna take a swing at this? SoftCircuits is not set up for
hardware manufacturing — but _if_ no one goes for this, mebbe we will anyway.
–Ben–
Am I hearing an Apple ][ type hack here, or am I misreading what you're
describing?
"Well, lessee, to draw a line you turn on these bits, then you have to go down
there and turn on THOSE bits, you see, 'cause it's not really an orthogonal
raster, and it's shared between two bit planes, but you also gotta flip that
control bit over there…"
'Course with the ][ it was also "… and you can't draw THIS color in EVEN
columns or THAT color in ODD columns, and oh yes, you can't draw THESE colors
next to THOSE colors…"
8)
Rick
No, in the Amiga you could use any color next to any other color, simply given
that you were limited to four colors total. As for the orthoganality of the
resulting pixel/bit orientation, you would need to do some custom rendering
routines… in this machine, with the blitters barrel shift capability, things
could work quite nicely. You could even use the current routines to manipulate
your window borders and so on as long as you were willing to use two colors…
otherwise, _everything_ would need a rewrite. This kind of thing would, in my
mind, occasion a rewrite of the text render functions and the line/circle
render functions… with those done, you could use the "new" mode and get some
very high definition use out of it. And for not much effort… I would think
that the use of such a display could be made transparent to the software trying
to use it via the Amiga's replaceable vectors for exec/kernal functions…
OpenLibrary(FunkyMode); OpenScreen(FunkyNewScreen); life as usual, except
x=0-1279, y=0-399, bp=0-1 for windows and 0-3 for graphics primitives. Of
course, if you want to render right into the bitplanes, you could, but then you
_do_ have to pay attention to the bits/bitplane organization. In any case, if
we do it, or if someone else does and provides us with one, we'll do the
primitives. –Ben–
You'd still need to change software to use the color mapping things,
which would be non-trivial. The nice thing about the de-interlacer
is that it works with existing software with no changes.
But it doesn't change your resolution — it just replaces a slow phosphor
monitor — arguably a different appraoch completely. What I am proposing would
allow (for instance) _much_ better looking text… something that wouldn't hurt
to look at. 640 is minimal for quality text display… look closely at a laser
typesetter's output. It's _much_ higher that equivalent to 640… –Ben–
CBM was showing a 1280 x 800 mono board at Comdex. No info on pricing and
availability.
Sounds like the CBM mono board is actually a custom monitor
with a converter board built in. I haven't seen the board, but
from what I heard it sounds like a cluge – you wind up using
the Amiga's memory bandwidth to get the full screen display, IE
system performance is degraded same as it would be in a HI-RES
16-color display. Also, screen update is at 15 HZ's (it takes
4 frames to update the display). Guess I'll refrain from
further speculation 'till I actually see it C-:
I didn't see it either — was just told about it (since I didn't go to Comdex).
On Monday a friend is going to bring the CBM press kit he got at the show over
for me to see. Hopefully it's talked about there. If not, I guess it's a fair
question to ask Hazy about next Wed. nite.
A difference should be pointed out about the Mono?? Monitor that CBM showed
at Comdex, and other hi-res mono monitors for the MAC.
Not only the image was very sharp BUT it also shows 4 shades of gray which is
a GREAT difference from those large Mac monitors (which are only 2 color).
Also that the monitor works with both the 500 and 2000 with the new software.
From what I was told all the electronics are IN the monitor. The amiga puts
out a 640x400x4 image converted to a 1000x800x2.
If only Dave of C-A was around to explain it better.
In either case (I WANT ONE!!!!).
I haven't heard from Dave lately… hope he's okay!… but he's supposed to be
here in formal conference tomorrow night, and I'm sure he'll talk about that…
if someone asks him… <hint hint>
Rick
From what I heard, the C= monitor and software uses four "tiled"
screens, which are demuxed and stored internal to the monitor.
If you were to watch the display on a normal Amiga monitor, you
would see it cycling through four different screens at a 60 HZ
display rate, which are combined to give one display that is
updated at 15 hz's in four quadrants.
Yes, this is true. The new hi-res monitor has a 15 Hz refresh rate. They say
the monitor will be about $500, and plugs into the standard RGB port on back,
and requires only new software to get the higher resolutions. Try that on your
Mac, Jack. 🙂
Don–
Must've been a Freudian slip. I checked and it was more like 1024 x 960.
Still, it would be nice to have the sort of full page wysiwyg monitor that the
MAC can have for professional DP work.
// Jim
"1 Meg x 1Meg resolution would be… 128 gigabytes!"
Not to mention one other fact: assuming we're only talking about decimal 1 Meg
and not binary 1 Meg, and assuming that you are willing to drop your vertical
refersh rate to 50 Hz to ease circuitry requirements while still avoiding
flicker, you would require a horizontal sweep rate of *60 MHz,* and a video
bandwidth of around… oops, my calculator don't go that high. 8)
Rick
Rick,
Forget the video scan rate…care to try to figure out how long it
would take for 4 Turbo Amigas each filling 1/4 of the screen to fill out an
entire screen? My rough calculation states about 1.5 hours. And that presumes
no addressing limits on the Turbo by using paged RAM.
Don
Hehe. I slipped on my statement. I used 60Hz instead of 50Hz. The actual
horizontal scan rate "only" needs be 50 MHz, not 60. Wow, that makes it
do-able! <Silly grin>
Rick
Rick:
I have not picked on you for at least a couple of weeks now but '50
MEGAhertz horizontal scan rate?' Of course that would reduce the EMI problem a
little (except your video might bother the X-band hams <grin>).
73, bill
And you are still not picking on me. YES. 50 MEGAhertz horizontal scan rate.
Read the thread, Bill… read the thread. 8)
Rick
Rick:
Don't have the time right now but would be curious what the video rate
for that system is running (seems like it would have to be 12 Gig just to
maintain the current NTSC resolution). Of course at that rate a lower video
bandwidth would still allow rather high resolution using …say 10 frames or
more. 73, bill
Hehe… DON'T bother to do any figuring until you HAVE time to read the thread.
As a famous cartoon star once said, "That's a joke, son!"
Rick
Rick:
OK I'll go back to sleep <grin>. 73, bill
A de-interlacer can do the trick, if 700*450 – ish screen resoloution is
sufficient. I've been beta-testing the version that will be marketed by
Microway, and love it. This allows you to use a MultiSync style monitor with
"hi-res" interlace mode, but no flicker.
Can you explain what a de-interlacer is? Also, when this product due
that you're beta-ing?
David Masterson
The de-interlacer is a box or card that takes the Amiga's interlaced NTSC
video, and doubles the scan rate so it can be used with a Multisync monitor
without having the flicker that you will get using normal monitors. This
allows you to get near-workstation quality video out of the Amiga without
requireing any software changes, and using standard readily available
Multiscanning monitors such as you might use with an IBM EGA/VGA card.
Wait a sec', something doesn't sound quite right here (I'm not
technically inclined, though). If the Amiga is outputting its information for
a screen at one scan rate and something grabs this information and outputs it
in 1/2 the time, sounds like somethings going to spend half its time waiting
around and this might affect things like highly active graphics. Not a good
explanation of what I'm trying to say I know, but I hope you get the idea of
what I'm trying to get across. Care to elaborate?
David Masterson
There would be no such problem. The Amiga is spending X time sending _4_ bits
of info out. We take the four bits, send 2 of them during 0 -> .5X, and then
the next two during .5 -> .999X, which neatly fits jsut as you'd like it to.
The principal has been tested and works just fine… it's simply a hardware job
at this point. –Ben–
The de-interlacer buffers the display in it's own memory so it can supply both
video fields in the same time as a single field is normally displayed.