1

I have an old 16:9 CRT TV that has a zoom function. If the input signal is 4:3 letterboxed, in zoom mode it will only show the middle part stretched to the full screen leaving the black bars above and below the visible part of the screen. It's also possible to move the picture up and down in small steps to find the exact center.

How exactly does it do this? Is it possible to preprocess my video signal with a homebrew circuit to achieve the same result (the source is 4:3 letterboxed, the TV is set into true 16:9 mode, and the video manipulation circuit goes between the TV and the signal source)?

tpimh
  • 486
  • 4
  • 15
  • It should be easy for an analog TV to manipulate the signal like this. You may recall that a CRT TV continuously generates X and Y "coordinate" signals that steer the electron beam to each "pixel". Naturally there is the possibility to offset or amplify the X and Y signals. It is not like an LCD, where you need some kind of digital processor to make sure the right data flows to the right pixels. If they simply amplify the Y coordinate then... the Y coordinate is amplified, and there can be a knob to select how much it is amplified. – user253751 Jul 24 '23 at 09:48
  • @user253751 in light of the question's request for a circuit that goes "between the TV and the signal source" you've got it pretty much backwards. You can't "offset or amplify the X and Y signals" because they only exist inside the TV; they're generated from the timing of the analog video signal. This makes scaling very difficult in the analog domain, and far easier to do digitally. – hobbs Jul 24 '23 at 09:59
  • @hobbs the first part was "How exactly does it do this?" – user253751 Jul 24 '23 at 10:01
  • I don't think that's that simple. There are 525 (486 visible) or 625 (576 visible) scan lines in the source signal for NTSC and PAL respectively. In letterboxed video 61 or 72 lines respectively on both top and bottom would be completely black. That leaves only 364 or 432 lines respectively to stretch to the entire screen. My math may be a bit off, so correct me if I'm wrong. – tpimh Jul 24 '23 at 10:09

1 Answers1

4

How exactly does it do this?

It should be easy for an analog TV to manipulate the signal like this. You may recall that a CRT TV continuously steers an electron beam to hit each "pixel". There are actually no pixels on the screen, but the circuitry generates them by pointing the electron beam at the right place.

So if you want the electron beam area to be twice as tall, or a centimeter further up, that is quite easy to do by amplifying the vertical steering signal. It is not like an LCD, where you need some kind of digital processor to intercept the data for one pixel and send it to a different pixel. A CRT just puts the pixels in different places to begin with.

Is it possible to preprocess my video signal with a homebrew circuit to achieve the same result (the source is 4:3 letterboxed, the TV is set into true 16:9 mode, and the video manipulation circuit goes between the TV and the signal source)?

Not as easily. Although the CRT's circuitry generates the beam steering signals and makes it widescreen by adjusting where the pixels go, there's no way for the video signal to tell it that. You can't make a video signal that says please shift all the pixels up 5 centimeters; you have to put the picture in different pixels just like an LCD does.

Well actually, there is one. If your TV understands this code, it can tell the TV this a letterboxed widescreen picture, so please zoom automatically.

Do some films automatically display in widescreen instead of letterboxed? Then your TV probably understands this code.

The Wikipedia page shown before has many references describing the code. Your video signal processor would need to count the lines in the video signal, then count a specific position within the line, and then inject some binary pulses. It should be doable with almost any microcontroller, and a little extra circuitry (to make sure the uC can detect the timing pulses at the right voltage and inject the bits at the right voltage). Actually doing it will be a project outside the scope of this answer.

user253751
  • 12,469
  • 2
  • 22
  • 37
  • Thanks for a comprehensive answer! Although WSS would work with my particular setup, I was thinking of a completely standalone device, something that would skip the top black part, then let through 3 scanlines, then draw a scanline twice, repeat until the bottom black part starts again. I guess this is just too complicated to manipulate analog signal like this, so nobody really implemented it as a standalone device. – tpimh Jul 25 '23 at 07:18
  • @tpimh You could easily do something like that with a non-simple digital device that would read in the entire frame, change it and output an entire frame, adding at least one whole frame of delay. If the signal timing is not affected by the device, then it has to send the first active scanline to the TV, before it receives that scanline from the video source, which is time travel. If the signal is delayed by the device to avoid time travel, then you have a much more complicated device with a bunch of memory and processing. – user253751 Jul 25 '23 at 07:52
  • maybe it would be possible to make this device in analog using a multi-tapped delay line. Each time you wanted to duplicate a scanline you'd switch to the next delay, one scanline longer, and at the start of the frame, reset the delay back to zero. Maybe you would have a mechanical rotary switch spinning around 60 times per second to select the right delay at the right time :) Still not simple. – user253751 Jul 25 '23 at 07:57
  • You may wish to upvote the question. Mine is the only upvote and Q&A pairs like this are excellent knowledge sources. – Russell McMahon Aug 08 '23 at 23:47