[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Resurrecting the Apple I
- Subject: Re: Resurrecting the Apple I
- From: mjmahon@aol.com (Michael J. Mahon)
- Date: 08 Dec 2002 09:00:14 GMT
- Newsgroups: comp.sys.apple2
- Organization: AOL http://www.aol.com
- References: <asooq3$3fd8$1@node21.cwnet.roc.gblx.net>
- Xref: archiver1.google.com comp.sys.apple2:25090
Dan Foster asked:
>In article <qhbs402fvj.fsf@ruckus.brouhaha.com>,
>Eric Smith <eric-no-spam-for-me@brouhaha.com> wrote:
>>
>>The Apple ][ design, on the other hand, shows an absolutely remarkable
>>amount of cleverness. It took me a *long* time to fully understand the
>>design of the system timing generation (the lower left corner of the
>>large schematic) and why some things were done in a particular way. For
>>instance, why every 65th clock cycle is slightly longer than the others.
>>But once I finally understood, I was in awe of Woz's design skill.
>
>I've seen Woz's ingenuity in the other parts of the design so I'm not
>surprised. But since you've gotten the better of my curiosity... ;)
>
>So, why is the 65th clock cycle slightly longer than the others?
To keep the hi-res pixels in phase with the color subcarrier, which
in the NTSC system is one-half cycle out of phase with the previous
line. ;-)
The reason for this NTSC system choice is that it puts the energy
of the color subcarrier exactly in between the signal frequencies of
the luminance signal. The 3.58... MHz signal was chosen to be an
odd multiple of one-half the line (horizontal) scan frequency--thus the
one-half cycle phase shift on each subsequent line.
This relationship also permits a comb filter to precisely separate
the chrominance signal from the luminance signal. It also makes
areas of solid color appear to be "crosshatched" instead of vertically
striped on a black & white set that doesn't filter the chrominance
signal out (as was true for all B&W sets made prior to the NTSC
"compatible color" standard).
In order to maintain a stable (non-complementary) color mapping
for even and odd hi-res pixels, Woz had to delay the start of each
horizontal line generation by on-half cycle of the subcarrier frequency,
or 140ns. Each line is 65 cycles long, so every 65th cycle is "stretched"
by 140ns.
-michael
Check out amazing quality 8-bit Apple sound on my
Home page: http://members.aol.com/MJMahon/