Re: Apple IIGS info and stuff

[dempson@actrix.gen.nz (David Empson)]

Supertimer <supertimer@aol.com> wrote:

> spec@vax2.concordia.ca (Mitchell Spector) wrote:
> >In article supertimer@aol.com (Supertimer) writes...
> >>10. Display:
> >
> >>        --640x200 with 4 pure colors: this mode is bland
> >>          and is not often used.
> >
> >    Actually it is the same mode as below, only the same four
> >colors are copied and repeated in each of the 4 mini-palettes
> >so you end up with four pure colors.
> 
> No it is not.  You do not dither with this mode.  Read on.

Yes it is.  The video hardware is in exactly the same mode.  The only
difference is how many different colours are present in the
mini-palettes.  I wouldn't call this a different mode, just a different
way of using the same mode.  It all boils down to your definition of
"mode".

> >>        --640x200 with 16 dithered colors: in this mode, the
> >>          pixels in the graphic screen are grouped into
> >>          even and odd columns.  The even columns can have a
> >>          palette of 4 pure colors out of a of 4096 possible.
> >>          The odd columns can have a second palette of 4
> >>          pure colors.  The GS dithers the adjacent colors
> >>          for 4x4=16 dithered colors.  This mode is widely
> >>          used in productivity programs and also in Apple's
> >>          Finder for the GS.
> >
> >    White and black get repeated in each of the 4 mini-palettes,
> >so you have them in both the odd and even pixel columns. That is
> >why you have readable text in the 640 mode (it'd be difficult to
> >read if text were dithered). So there are 6 pure colors usually
> >present in the 640 mode, of which only 2 remain pure (black and
> >white)--the other 4 get mixed to make the 16 dithered colors.
> 
> Yes, colors are repeated in the mini-palettes, but this is
> still the MODE used.  If 4 colors were used over the compete
> screen as one palette, then you would have the other MODE.

The computer is not doing anything special to implement dithering,
though the software is setting up the right colour patterns to cause
this effect.  It is a characteristic of the monitor's limited horizontal
resolution and the behaviour of the human eye.

If you managed to get a higher resolution monitor connected, the
individual coloured pixels would be more visible, but the eye would
still tend to average the colours.

> As a comparison, you CAN'T get CGA mode on a PC to
> replicate the IIGS 640 mode with 16 dithered colors.

Well, no, because CGA can't do more than two colours in 640 mode, for
starters (but it could dither those colours).  You could probably do it
in EGA or VGA, but there wouldn't be much point as you have enough pure
colours available.

> > [Text mode]
> >
> >    Using scanline interrupts, you can do neat tricks like having
> >the foreground, background and border color changed on a line by
> >line basis (much like the 256 color mode, only you have multi-
> >colored ASCII text in this case).
> 
> Now this is something I didn't know.

Neither did I, and I suspect Mitch is wrong.  To my knowledge, scanline
interrupts only work in Super Hires mode, since they depend on the scan
line control bytes which are part of the SHR screen memory image.

It _is_ possible to change the text colours on the fly with the text
screen, but as far as I know it is done by monitoring the horizontal and
vertical scan location registers.

> Do you know of any programs that use this multi-colored ASCII?

At least one of the FTA or similar demos, but I forget which one.

> >>     The difference between a UniDisk and a IIGS 3.5" drive
> >>     is that the IIGS drive is controlled directly by the
> >>     computer while the UniDisk has a separate processor.
> >>     The UniDisk is thus much slower (up to 4x slower) than
> >>     a IIGS 3.5" drive.
> >
> >    The UniDisk 3.5 is actually closer to half the speed of
> >of an Apple 3.5 drive.
> 
> It is pretty bad when you factor interleave and if the disks
> are HFS, access time is closer to 4x slower on the UniDisk.

It isn't quite as simple as that.

The ideal interleave factor for an Apple 3.5 Drive on a fast enough
machine is 2:1, while the UniDisk's ideal interleave is 4:1 on any
machine.

If you are accessing a 2:1 interleave disk, the UniDisk will be
significantly slower than an Apple 3.5, since it will require an entire
revolution to read adjacent blocks.  This will give a speed penalty of
anything from 4 to 6 times.

If you are accessing a 4:1 interleave disk, both drives will be about
the same speed, except in cases where the driver for the Apple 3.5 Drive
is able to read the entire track in one pass: ROM 3 firmware, GS/OS,
SuperDrive card; the Apple 3.5 Drive is about four times faster in these
cases.  The UniDisk will take twice as long as the Apple 3.5 drive to
transfer the data for each sector, so there is a slight delay before
data can be processed.

If you access the appropriate interleave factor disk in each drive, then
the Apple 3.5 Drive is at least twice as fast as the UniDisk 3.5
(depending on track caching issues).

Note that 4:1 is the recommended interleave for an Apple 3.5 Drive on a
IIgs under ProDOS, since there is more software overhead in reading
adjacent blocks (another driver call is required).  A 2:1 interleave
disk might actually be slower in this case.

Note also that when writing, the track caching of Apple 3.5 Drivers
cannot be used, so it drops back to interleave comparisons.

An interesting aside: writing to 1:1 interleave high density disks with
a SuperDrive controller under ProDOS-8 with a PC Transporter is VERY
slow, as it isn't able to get the data ready quickly enough.  I always
reformat my PC disks at 2:1 interleave if I'm using them to exchange
data between a real PC and the PC Transporter.  (Reading is no problem,
as the SuperDrive card caches the track.)

-- 
David Empson
dempson@actrix.gen.nz
Snail mail: P.O. Box 27-103, Wellington, New Zealand