Re: Adapter to use IIGS's RGB output on modern VGA/SVGA monitors (idea?)

[mjmahon@aol.com (Michael J. Mahon)]

Tim Haynes wrote:

>"Michael J. Mahon" <mjmahon@aol.com> wrote in message
>20030513193743.28006.00000333@mb-m05.aol.com">news:20030513193743.28006.00000333@mb-m05.aol.com...
>> ToasterKing replied:
>>
>> >In article <20030512180919.24954.00000344@mb-m06.aol.com>, Michael J.
>> >Mahon <mjmahon@aol.com> wrote:
>> >
>> >> Some signal processing is required, but perhaps only simple
>> >> analog addition and subtraction.
>> >>
>> >> There are quad video op-amps available that will certainly
>> >> do the job of generating YPbPr from RGB, but you'll have
>> >> to get some power to them, since they drive 75 ohm loads.
>> >>
>> >> The Y signal is a weighted sum of R, G, and B signals,
>> >> where the weights are a standard.  Pb is B minus Y and
>> >> Pr is R minus Y.  All signals have standard polarities and
>> >> peak-to-peak amplitudes.
>> >>
>> >> -michael
>> >>
>> >> Check out amazing quality 8-bit Apple sound on my
>> >> Home page:  http://members.aol.com/MJMahon/
>> >
>> >That sounds like possible a little more complex design than I was
>> >hoping for, especially since my circuit design skills are a little
>> >rusty.  Since I don't know exactly what the specifications of YPbPr
>> >call for, and how to manipulate the signal from RGB, I'm not really
>> >sure where to start here.  If you happen to have any of this knowledge,
>> >could you possibly point me in the right direction?  Otherwise, can you
>> >think of any other easily-achievable design that's a little more
>> >elegant than conversion to first YPbPr, then to VGA?
>>
>> There's a Linear Technology Application Note that has a
>> circuit for converting RGB to YPbPr:
>>
>> http://www.linear-tech.com/pdf/13989f.pdf
>>
>> It requires two dual video op amps and some resistors.
>> It does not show adding composite sync to the Y signal.
>> That will require another op amp between the Y output
>> shown in the diagram and the real Y output, to add (or
>> subtract--I couldn't quickly confirm the IIgs sync polarity)
>> the composite sync to the Y signal in the right proportion
>> (about 1/4 of the Y peak-to-peak amplitude).
>>
>> The circuit assumes +5v and -5v supplies, which is
>> a bit awkward.  The circuit could probably be adapted
>> to work from a single-ended supply, since I believe all
>> the signal components are positive(?), but levels would
>> have to be shifted some to keep the video linear.
>>
>> Assuming that 75 ohm impedances are used throughout,
>> the current requirement from each supply should be less
>> than 100ma, so, though it's not a power hog, it will need
>> access to "real" power.
>>
>> The reason that this "transcoding" is so simple is that
>> we need not do anything to change the timing of the
>> signals, only the mixing of the components.
>
>Michael, I think this group needs to sponsor a couple of months of your time
>so you could finally build us all the IIgs2VGA box/dongle/card/whatever that
>everybody keeps fantasizing about.  :-) :-)

Thanks, Tim...I think.  ;-)

Actually, I don't think I'm the person to do this.  The OP had found
a box to convert component NTSC-rate video to VGA, and so was
looking for a way to convert the IIgs's NTSC-rate RGB to component.

The hard part is the scan conversion from NTSC rates to VGA
rates.  To obtain the resolution and stability that you are
accustomed to with an AppleColor RGB display, very high
bandwidth color signals (10MHz or better) must be sampled,
stored, and read out at VGA rates (while other samples are
being stored).  And the sampling must be either synchronized
with the IIgs pixel clock or very cleanly anti-aliased to obtain
a stable, non-jaggy display.

This is not simple, and it will not be cheap unless it leverages
some high-volume chip to do the job.  I am not convinced that
such a chip exists, with the desired quality level.

I would love to find out that I'm wrong--so I encourage anyone
to experiment with various converters, hoping for success.
I tried a "Video Cheese Box" (Yes, really ;-) and was quite
disappointed with the quality of the conversion of lowly
composite video to VGA.  The composite format, like the
S-Video format, is fundamentally incapable of carrying the
information to make a crisp IIgs display, so our hopes must
be pinned on direct RGB or component video.

As you can see, the market is not making this easy.  ;-)
Part of the problem is that consumer electronics folks feel that
the computer world is a threat to their domain, and so have done
just about anything to avoid providing RGB interfaces, even though
essentially _all_ display devices natively require RGB.

This is creating an opportunity for third parties to create adapters,
of various complexities and qualities, to do transcoding and scan
conversion.

What we can all hope for is that, against the desires of the TV
manufacturers, the standard TV interface becomes scan-rate-
adaptive, RGB--as computer monitors and projectors have become.
Of course, it won't.  The "emerging" standard is some encrypted
digital link (IEEE 1394, or what?), to protect digital content.

Meanwhile, others are trying to get UWB accepted as a wireless
way of communicating among video components.  (And you can
bet that the protocols they choose won't be TCP/IP.  ;-)

It's a mess.  And unless consumers make an ugly noise about
all the complexity and non-uniformity, it will continue to delay
computer-TV convergence for another generation.  ;-(

-michael

Check out amazing quality 8-bit Apple sound on my
Home page:  http://members.aol.com/MJMahon/