Electron S-Video output

discuss both original and modern hardware for the bbc micro/electron
Post Reply
Bitstik
Posts: 117
Joined: Thu Feb 18, 2010 2:59 pm
Contact:

Electron S-Video output

Post by Bitstik »

When researching video output for a project I came across this mod from another forum:

https://www.amibay.com/threads/s-video- ... ron.37853/

My motherboard layout on my issue 6 is slightly different but I soldered the red lead to the right hand pad of Lk4 without connecting the pads - taking the ground and signal from the composite socket.
IMG_4596.jpeg
IMG_4595.jpeg
Tested it via a usb video capture card (reason for wanting S-video output) and it works/ seems very clear. Need to test further.
Hope this is of some use. As the original poster had said- I haven’t seen this anywhere else.
User avatar
egel
Posts: 158
Joined: Mon Nov 25, 2019 1:10 pm
Location: Almere, Flevoland, The Netherlands
Contact:

Re: Electron S-Video output

Post by egel »

I made this mod just last Friday on an electron. But the colours are a lot less saturated than I would like.
Bitstik
Posts: 117
Joined: Thu Feb 18, 2010 2:59 pm
Contact:

Re: Electron S-Video output

Post by Bitstik »

I haven’t soldered the connection to the svideo adaptor yet but this is the mode 2 screen grab with the jumper wires
IMG_4604.jpeg
Barneyntd
Posts: 229
Joined: Fri Feb 28, 2020 9:55 pm
Contact:

Re: Electron S-Video output

Post by Barneyntd »

I've added s-video output to my model B. The circuit I found recommended a 1K resistor on the C+ line, but I've no idea how important that is. I actually removed the UHF modulator (I can't imagine any use for it), and put the s-video socket in its place. The quality through a cheap usb video capture device is not perfect, and the delay is too great for fast games, but it's much better than colour composite.
Completed projects: Multiphrom, My own rom board, Extra VIA, iPhone pauser. Current projects:
vexorg
Posts: 289
Joined: Wed May 24, 2023 5:05 pm
Contact:

Re: Electron S-Video output

Post by vexorg »

In theory, you sholdn't see much difference between composite and s-video. They essentially contain the same information.
David
Barneyntd
Posts: 229
Joined: Fri Feb 28, 2020 9:55 pm
Contact:

Re: Electron S-Video output

Post by Barneyntd »

vexorg wrote: Mon Apr 08, 2024 11:37 am In theory, you sholdn't see much difference between composite and s-video. They essentially contain the same information.
RGB also contains the same information, but the quality is visibly better. The fundamental difference is how many wires you are pushing that information through: RGB has three, S-video has two, composite has one. The device at the other end has to separate the information out into three channels, and this is not a lossless process. S-video is a good compromise because it keeps brightness (Y) separate from colour (C), and our eyes are more sensitive to inaccuracy in brightness than in colour.
Completed projects: Multiphrom, My own rom board, Extra VIA, iPhone pauser. Current projects:
vexorg
Posts: 289
Joined: Wed May 24, 2023 5:05 pm
Contact:

Re: Electron S-Video output

Post by vexorg »

i'd still say there's no real noticable difference between the quality of s-video and composite, regardless of how many wire you have in there.
RGB is more than 3 as you have the sync signals that are also embedded into s-video and composite.
David
User avatar
1024MAK
Posts: 12783
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Electron S-Video output

Post by 1024MAK »

The real difference is in the available bandwidth for the respective signals.

RGB+Sync has four channels, each independent and each with a wide bandwidth without any limiters that compromise the signal by any significant amount.

Monochrome composite video is similar (as long as a colour display disengages its colour decoder).

Composite video is full of compromises.

Mark
User avatar
arg
Posts: 1289
Joined: Tue Feb 16, 2021 2:07 pm
Location: Cambridge
Contact:

Re: Electron S-Video output

Post by arg »

Some of the above analysis may be correct for video in general, but for output from these machines it is simpler.

RGB is 'perfect' - there is no information lost in the output formatting, bandwidth of the signals is greater than the pixel resolution.
S-video loses colour resolution due to the limited bandwidth of the chroma signal, but has perfect luminance.
Composite has the same colour limitations as S-video, but additionally loses accuracy in the luminance - either the receiving device filters the signal to remove the chroma carrier, in which case horizontal resolution is lost as the luma bandwidth is reduced; or else it doesn't and then chroma information spuriously corrupts the luma.

The latter is why Acorn machines don't put colour on their composite outputs by default: if connected to a monochrome monitor (which has no filter as it supports high resolution) and attempting to display a black-and-white picture, there will be a fuzzy overlay from the chroma carrier which the monitor is unaware of and will treat as part of the luma signal.

The chroma carrier is at approx 4.43MHz, compared to the Mode 0 pixel rate of 16MHz (so 8MHz for alternate black/white pixels, and more bandwidth than that if you want them to be flat), so this is quite limiting.

S-video done properly will give you perfect black-and-white pictures, significantly better than composite. However, for colour pictures the advantage will be smaller, and for something like Mode 0 with lots of colour and big pixels the advantage will be minimal.

As for why the modified electron gives poor colour saturation, it will be because the output impedances don't match. The composite (monochrome, S-video 'Y' component) is designed to drive the standard 1V into 75R, with the amplifier around Q8 giving a suitably low-impedance output. The chroma on LK4 is designed to feed into the _input_ of that amplifier, so is at a rather higher impedance. Since this is also a 'hack' facility. Providing another buffer amplifier, or tweaking the component values around Q7 should allow a more correct chroma signal.
vexorg
Posts: 289
Joined: Wed May 24, 2023 5:05 pm
Contact:

Re: Electron S-Video output

Post by vexorg »

The other side of if it is most CRTs will be old and tired, which is why I doubt you'd see much difference.
I'd think you'd need something professional to see the difference, the old style TVs were not that high quality inside.
And anything modern with an LCD / flat display will have it's own issue converting from analog to digital.

Even RGB on a proper monitor back in the day struggled with mode 0.
David
Barneyntd
Posts: 229
Joined: Fri Feb 28, 2020 9:55 pm
Contact:

Re: Electron S-Video output

Post by Barneyntd »

vexorg wrote: Tue Apr 09, 2024 9:20 am Even RGB on a proper monitor back in the day struggled with mode 0.
Mode 0 (or 3) is kind of the point. Colour composite mucks up the signal enough that I don't think any display can produce readable output: the resolution simply isn't there. With s-video these modes are not crisp, but they are perfectly useable (at least in black&white), and 80 columns is a genuine advantage when editing. I have my keyboard switches set to boot up in mode 3, because it's what I use most, but I don't think I could with composite.

You describe mode 0 back in the day as a struggle, but I remember when I got my cub the enormous difference it made being able to edit with 80 columns, and I really hated when memory limits forced me back to mode 7. The point of s-video is that, although the colour is only as good as colour composite, the black&white is as good as RGB, since sync signals only happen off screen.
Completed projects: Multiphrom, My own rom board, Extra VIA, iPhone pauser. Current projects:
vexorg
Posts: 289
Joined: Wed May 24, 2023 5:05 pm
Contact:

Re: Electron S-Video output

Post by vexorg »

On the monitior I had, not cub, the characters were readable, but you had a bit of divergence at the edges of the screen that made the pixels shimmer a bit.
David
atsampson
Posts: 67
Joined: Fri Dec 16, 2022 4:54 pm
Location: Dundee, Scotland
Contact:

Re: Electron S-Video output

Post by atsampson »

Barneyntd wrote: Tue Apr 09, 2024 12:45 pmThe point of s-video is that, although the colour is only as good as colour composite, the black&white is as good as RGB, since sync signals only happen off screen.
Sync signals only happen off-screen regardless - there's a short horizontal sync pulse at the start of each line, and a longer vertical sync pulse at the start of each field. Both of these will be off-screen unless you've turned the width and height right down.

The reason you can lose luma (black-and-white) detail with a composite signal is that the luma and chroma (colour) signals overlap in terms of frequency. Chroma is quadrature-encoded as variations in the amplitude of two fixed-frequency sine waves 90 degrees out of phase for the two colour components, so fine detail in luma can produce signals at the same frequencies used to encode chroma. The very simplest way to separate these at the receiving end is to filter them in terms of frequency - e.g. say that everything from 4.0-4.9 MHz is chroma, and the rest is luma - which looks OK for typical TV pictures, but means that sharp edges in the luma, like you'd find in high-resolution text, are instead interpreted as chroma, producing spurious colours.

Better-quality monitors and capture devices use more elaborate methods of separating luma and chroma, taking advantage of repeated information between lines and frames, and a modern adaptive decoder can achieve essentially perfect separation of luma and chroma on still images, so you'll see less difference between composite and S-Video on a modern display - at least when there's not much motion. (And sometimes you'll get very odd effects when the adaptive decoder guesses incorrectly about what the signal's supposed to be...)
4-3is4-me
Posts: 4
Joined: Tue Feb 20, 2024 7:28 am
Contact:

Re: Electron S-Video output

Post by 4-3is4-me »

Bitstik wrote: Mon Apr 08, 2024 9:55 am I haven’t soldered the connection to the svideo adaptor yet but this is the mode 2 screen grab with the jumper wires IMG_4604.jpeg
I have been thinking of experimenting with the same mod, I'm also using a capture card. Something which I also considered while researching was the wikipedia entry for s-video has a composite to s-video conversion circuit which places a 470pf cap on the chroma signal between the two. I found further reference to this on an old post in a german forum, which may have been the source.
mfaxford
Posts: 11
Joined: Tue Dec 12, 2023 12:55 am
Location: Near Bristol
Contact:

Re: Electron S-Video output

Post by mfaxford »

Barneyntd wrote: Mon Apr 08, 2024 11:12 am The quality through a cheap usb video capture device is not perfect, and the delay is too great for fast games, but it's much better than colour composite.
I had tried looking at a similar mod on my BBC Master. I then found that the cheap usb capture device that advertised being composite and s-video didn't really do s-video. Internally it had the Luma and Chroma pins connected together (so was really just composite with a mini-din plug).
BBC Master: Now with PiTube and Pi1MHz. Disc Drives and Cub Monitor needing some checking and attention
Risc PC: needing some work to cleanup
User avatar
1024MAK
Posts: 12783
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Electron S-Video output

Post by 1024MAK »

It is possible to build a small 'buffer' (to provide an improved signal level and low impedance output).

Mark
Post Reply

Return to “8-bit acorn hardware”