[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Apple video bandwidth question

MdntTrain wrote:

On Aug 28, 4:55 pm, Alex Freed <alex_n...@mirrow.com> wrote:

MdntTrain wrote:

For Apple II Hires 280x192 monochrome video, isn't the monitor
bandwidth needed 3.59mhz with a dot clock of ~7mhz?
I thought we'd posted here some time ago that dot clock was twice the
bandwidth requirement, but I was reading in a Sencore computer monitor
analyzer manual today that to calculate a monitor's bandwith, you do
the following:
Example for Apple:
horiz scan = 15.7khz
total scan time = 1/15.7khz = 63.69us
active video time = 40/65 * 63.69us = 39.19us
display time per pixel = 39.19us / 280 pixels = 0.14us
bandwidth required for single pixel = 1/0.14us = 7.144 MHZ.
Yes, the dots would surely need to be clocked out at 7.144mhz, but
something seems wrong with their method as it doesn't to take into
account half cycles.
Or does dot clock equal bandwidth?  I just forgot! lol.
 jS

The NTSC TV had the luminance bandwidth of less than 3.57 MHz because
the color subcarrier was at 3.57. No comb filters at the time either.
Better monitors could turn the filter off if no color was detected and
extended the bandwidth to say 5-6 MHz.

So in practical terms 3.57 is enough to see every pixel in 280x192 mode,
but in theory the spectrum of a square wave is infinite. So you need
infinite bandwidth to pass it with a 100% accuracy. Fortunately a lot
less will work just fine.

BTW the 80 col mode uses 14 MHz pixel clock and so it doesn't look good
on a color TV but looks OK on a monochrome monitor with a bandwidth of
about 6 MHz or so.

-Alex.- Hide quoted text -

- Show quoted text -


So to restate, needed bandwidth is 1/2 pixel clock rate, as highest
bandwidth requirements would be alternating black/white lines, and
your first half cycle would perform the white line, while your second
half cycle, the black line... am I understanding correctly?


That is correct.

However, the rise time that the monitor supports--or the "squareness" of
the edges, if you prefer--depends inversely on actual bandwidth (in this
case, we will assume 3dB bandwidth, with a moderate rolloff slope, say
6dB/octave).

A conventional formula is: risetime = 0.35/bandwidth.
> http://www.tek.com/Measurement/App_Notes/Technical_Briefs/bw_rt/55_18024_0.pdf 

The reason that risetime is important for Apple monitors is that the
appearance of a pattern with single-line horizontals and single-pixel
verticals will have noticeably different brightness of the horizontal and
vertical parts if the risetime of the monitor limits the amount of time
that a pixel is at or near 100% response.

This effect is most evident on the Apple II with 80-column text, where
a lower bandwidth monitor will show bright horizontal font lines and
relatively dim single-pixel-wide verticals or diagonals.

A good NTSC-standard monochrome monitor for the Apple II will have at
least 12MHz bandwidth for crisp 80-column text with equal brightness
horizontals and verticals.

This allows a single-pixel about 30ns of rise time, followed by about
110ns near full amplitude, which is more than 3/4 of the pixel width.
A lower bandwidth monitor will cause this pixel to spend proportionately
less time near full amplitude, resulting in a blurrier and less-bright
pixel.

-michael

NadaNet 3.0 for Apple II parallel computing!
Home page:  http://home.comcast.net/~mjmahon/

"The wastebasket is our most important design
tool--and it's seriously underused."