Split TV signal to make a video wall

Question

I'm working on this project for University. We are generating visuals by sending audio signals from the computer to an old tv. Left channel to TV's Vertical, right channel to TV's horizontal. We send simple sound waves (generated in PureData), example, 400hz to the left channel and 200hz to the right channel. And also some FMs.

We are getting this kind of visuals:

http://youtu.be/gc67zXuBw9A

http://youtu.be/7Uqq8SJhm9o?t=1m29s

The thing is that we want to have this in a larger scale, and I thought of making a video wall, the same signal splitted into 9 (or 8 or whatever) TVs.

I have no idea on how to make this, any ideas??

Thanks!

Answer 1

Since you're using fairly low frequencies, this will be straightforward but tedious. What you'll need to do is to divide each channel into 3 segments - left,horizontal center and right, and bottom, vertical center and top. Then feed the 9 combinations (permutations, actually) to your 9 monitors.

You make the 6 segments using op amps. Let's say, just for reference, that your horizontal and vertical waveforms as now used have a range of +/- 1 volt. Then what you do is provide 3 offsets - -2/3, 0 and 2/3 volts. Then put each offset signal through a gain of 3. You can combine these functions, and do the whole thing in a single quad op amp, 1 op amp each for vertical and horizontal.

^{simulate this circuit – Schematic created using CircuitLab}

And, of course, I messed up the schematic. The op amp I'd intended is a TL084 - the TL081 is a single. Note that for loads of 10 kohm the TL084 will work out to about 100 kHz, which should be more than adequate for your needs, and you may even need to slow them down a bit by putting a small capacitor across the feedback resistors to suppress high-frequency pickup.

ETA - And another boo-boo on my part - R10 should also be 30k, rather than 10k. Damned dyslexic fingers!

For the real thing, you'll need the gains and offsets slightly bigger than shown in order to accommodate the dead zones between the TVs. Assuming 25% dead space, the offsets should be +/- 0.278, and R2 should be 12.5k.

You'll note that there is none of the logic / screen selection which Phil Frost mentioned. The reason is simple - for this you don't need it. Signals which are out of bounds will simply be off the screen.

Answer 2

Sure you could do it, however it won't have the same elegance as what you have now. It looks like you are directly steering the electron beam, so you have a horizontal control (X) and a vertical one (Y), and the beam is more or less linearly positioned according to those inputs. So if you say,

$$ X = \cos(t)\\ Y = \sin(t) $$

You get a circle. Neat, elegant.

If you want to split that up into multiple screens, it's not linear anymore. Say your signal generator outputs a value between 0 and 1. In your example of nine TVs, there are three in each axis. So we must divide this range of 0 to 1 into three components:

• \$0 \le x < 1/3\$: left column active
• \$1/3 \le x < 2/3\$: middle column active
• \$2/3 \le x < 1\$: right column active

Then after having decided which TV is active, this number needs to be scaled back to the range of 0 to 1 for the individual TV that is active. You could do this by multiplying by 3 and keeping just the fractional part.

In pseudocode:

function split(x):
  if x < 1/3:
    column = "left"
  elif x < 2/3:
    column = "middle"
  else:
    column = "right"
  return column, (x * 3) % 1

This returns which column of TVs you need to activate, and then gives you the number between 0 and 1 to feed to that TV. Then, you need to do the same for the vertical axis. Knowing which column and which row of TVs, you activate that one and deactivate all the others, and feed it the transformed value.

Since only one TV is on at a time, you might be able to feed the same X Y signals to all the TVs, and only the one that is enabled will show it. This way you don't need nine audio cards for your computer.

You could implement this with analog electronics with some comparators to determine which row (or column) you are in, some logic gates to combine the row and column into an enable signal for the individual TV, and some op-amps to scale the signal after dividing it. Or, you could do the processing in PD.