Re: Open hardware AV to digital conversion

From: Ingo Korb <>
Date: Tue, 13 Jan 2015 19:58:37 +0100
Message-ID: <> writes:

> You are right that S-video will disappear earlier than composite, which
> will follow soon. Upscalers/converters/etc. will still be available for
> some time - then we should have a proper digital output already.

From what I've seen it looks like S-Video is almost gone already -
current AV receivers don't have S-Video inputs anymore, at least my
current TV does not accept S-Video at all (*) and it looks like Lumagen
has dropped S-Video inputs from its current generation of video scalers
too. All of those examples still support composite and (usually)
component video though.

(*) The price comparison web site I usually use doesn't list "has
S-Video input" as filter option (anymore?), it would be a nice data
point otherwise.

[PAL color decoding effects]
> Actually I wouldn't mourn that much over this loss but no - I wanted to

It would be quite bad actually, as far as I know these color decoding
effects have been used intentionally on occasion to create additional

[replaying the signal with an AWG]
> As always ;-) But that's an interesting domain. And idea too. Maybe we
> could get an arbitrary waveform generator that would be able to generate
> one frame repeatedly and then check the influence of various
> modifications.

The AWG would need around 500k points to replay two complete fields if
you use a standard video sampling rate (13.5MHz). For real video you
would probably want 10 bit resolution, but for tests with a C64 8 bit
should likely be sufficient.

On the sampling side, a modern DSO with large memory and the ability to
transfer its memory contents to a PC should be sufficient - I could
probably provide you with some sample signals if necessary.

> Guessing but I'd say because "my" is expected to be different every
> other frame, while yours was surely expected to be the same every

I'm still unsure if an interlaced PAL/NTSC video signal is really
expected to be different in every other frame - there is a half-line at
the beginning or end of the field, but other than that the number of
lines should be the same in each field.

> Interesting - I haven't seen that many of them but I am probably not
> frequenting various web forums enough. Have you seen some good examples?

The initial project probably was the modulator replacement circuit in which had a "stripe compensation" addition while the details
of the circuit and the best way to put it onto a board were still worked
out. Recently I've seen new "stripe compensation" and/or modulator
replacement PCB designs at least on and on
(someone actually built a board this time instead of everything getting
dragged down with discussions) and I seem to remember a third reference
somewhere that I can't find right now.

The modulator replacement is just a small circuit with two transistors
that provides the proper biasing for the VIC-II outputs and buffers the
signals so the onboard modulator can be circumvented. Since that small
metal box is also responsible for a bit of crosstalk between the chroma
and luma signals, the quality is improved. The stripe compensation
(AC-)couples a weak version of the AEC signal into luma to reduce the
visibility of the vertical stripes that are very prominent in the output
of the 85xx VIC-II.

The (reconstructed) original thread for that is at

> And those of us
> who continue to use CRTs for best experience could still benefit from
> better quality picture over S-video/composite.

I actually have a Sony BVM-9045 on my desk right next to a 22" LCD
TV. For some things the BVM is just much more convenient because it's
more tolerant of "weird" signals and switching between inputs with
dedicated mechanical buttons is sooo much easier than fiddling with a
remote or a single "Source" button hidden at the side of the TV. =)

-i 'Wanted: BKM-129X and/or BVM-D9H5E' k

       Message was sent through the cbm-hackers mailing list
Received on 2015-01-13 18:58:37

Archive generated by hypermail 2.2.0.