cmcmanis@sun.uucp (Chuck McManis) (03/06/87)
Video Graphics Basics and the Amiga
In a video monitor, the electron beam scans across the display from
left to right. The frequency of this scan is often referred to as the
horizontal scan rate. The beam also scans from top to bottom, this is
called the frame rate. In any case, the computer display circuitry
and the monitor have these two values designed in as constants. Minor
variations in the output of the computer are accounted for by phase
locked loops in the monitor. Two signals, vertical sync, and
horizontal sync, keep these phase locked loops in step. On the Amiga
these constants are 15,750 Hz, and 60Hz, for the Horizontal and
vertical scan rates respectively.
Ok, so the electron beam is whizzing back and forth at 15750 hz so
there are exactly 63.492 microseconds between, when the beam starts on
a line, and when it is positioned at the start of the next line. It
starts the traversal actually off the left side of the screen, so time
has to be allowed for it to get to the display area, then the beam is
switched on (unblanked) and the computer circuitry begins to feed it
data. The Amiga is running at 7.16 Mhz, and can actually transfer
data on every other clock so the Amiga has time to transfer 227 words
to the display circuitry. Allowing for the beam to get onto the
screen and to retrace back to the next line, the Amiga actually sends
160 words which when broken up by the display system into pixels can
be either 320 5 bit pixels or 640 4 bit pixels per scan line. Because
it doesn't push too hard there are usually borders on the left and
right side of the screen. The morerows program can increase the
number of pixels the Amiga trys to display on the screen. On a good
monitor the screen extends into the borders, on a TV set it extends
off the edges of the screen.
Ok so we can cram probably about 705 pixels on a line, with a good
monitor so what about the frame rate? Well the display circuitry (and
the monitor) is designed to send vertical sync 60 times a second, when
the Vsync pulse is received the beam is sent to the top of the screen
to start a new frame. Well 1/60th of a second is 166.6 milliseconds,
and remember above where we said that the beam is ready to start a new
line every 63.492 microseconds? Dividing those two numbers shows that
the beam will start a new line 262 times before the Vsync pulse is
recieved and it is sent back up to the top of the screen. Now you
also have to allow time for it to go back up to the top and get back
to displayable place on the screen. After making those allowances you
end up with 200 rows visible on the screen. Again with the morerows
program you can alter how many lines the Amiga trys to display, and on
TV's you will end up off the top and bottom if you don't allow enough
time for the beam to position itself.
Then comes interlace, whose idea was originally founded to double the
resolution of the then standard B&W T.V.s. Since the physical screen
had more resolution than the display hardware was producing, and the
beam traverses the screen from top to bottom at a constant rate, how
about starting one frame at the top of the screen and a second frame
half a line down? Well, they did and it is called interlace. What it
buys you is that given the constant horizontal scan rate, and constant
frame rate the same display hardware can 'fool' your eye by putting
the odd lines of a picture on the screen during one frame, and the
even lines (moved down by half a line) on the alternate frame. And
because your eye is so slow you will see both as though they were
there all the time!
That works fine except that your eye isn't that slow. And when there
are 'high frequency' signals on the screen that exceed the sample
frequency then they are on the screen only half the time. A high
frequency object is any object whose feature size is smaller than two
lines. Your eye does pick up this blinking on and off and percieves
it as flicker. Lowering the intensity helps because you have to work
harder to see it, which dialates your pupil which increases the
persistance of your vision.
So how does one increase the actual or perceived resolution of the
Amiga ? There are actually three well known ways and they are
presented in the order of difficulty and expense :
1) Increase the persistance of the monitor's phosphor so that a
frame that is displayed actually continues to glow on the screen
while the next frame is drawing. This requires a simple change
to the chemical composition of the phosphors, however do to the
low demand of this type of monitor not many are made and
manufacturers charge a premium for them, usually double what the
equivalent 'fast' phosphor monitor goes for. The only drawback
to this technique is a slight 'smearing' of the display when it
changes rapidly such as during animation. Generally this is
less objectionable than the flicker.
2) Build a scan doubler. This device is essentially two devices in
one. On one side it is a frame grabber/digitizer, taking the
video red, green, and blue signals and digitizing them into a
twelve bit bitmap in memory. On the other side it is a high
resolution graphics interface that has display circuitry running
with a scan rate of 31.5Khz and a frame rate of 60 Hz. It
simply displays a 640 X 400 X 12 bitmap on a high resolution
screen. The only tricky part of this circuit in the past was
the dual ported memory required, this however has been
alleviated by advent of the video ram produced by Texas
Instruments and others.
3) Increase the speed of the Amiga display circuitry so that it
provides data twice as fast. This is certainly possible, it
requires a 14.32 Mhz 68000 faster versions of the custom chips,
dynamic RAMs with and access time of about 75ns, and an
appropriate monitor. What this change would do is increase the
horizontal scan rate to 31.5 Khz thus doubling the number of
lines one could squeeze into a single 60 Hz frame. The
disadvantages of this technique are that the Amiga could no
longer produce video that was compatible with a video tape deck,
and the cost of the Base unit would go up by a factor of 4.
--
--Chuck McManis
uucp: {anywhere}!sun!cmcmanis BIX: cmcmanis ARPAnet: cmcmanis@sun.com
These opinions are my own and no one elses, but you knew that didn't you.