We can’t make this assumption quite yet. We can say for sure that your two units have a bit hotter signal on blue channel when in this overdrive error state. But we’d need to take some voltage measurements and recreate the issue across a few units to make a full diagnosis – in particular, I want to know how hot that negative voltage is. Does it go down to potentially -12V? (only an issue with extreme cases) Or do we start seeing some phase invasion happen around -2V (this could be annoying.) But the gist of it is correct: “Signals outside the anticipated input signal range may cause unknown issues.” In the screenshots you shared, it appears that negative voltages below black are what is causing this effect – we just don’t now how far below black they need to be.
We can potentially improve on this in the future by finding a way to incorporate better precision clipping on all inputs in general, but the practical question for right now is whether this is an issue for anyone patching signals within the expected LZX headroom (or really, within an overall +/-2Vpp input scale, which is the headroom we’re aiming for).
2 Likes
I suppose this would be a good question for @syntonie or maybe @pbalj. Having just moved I no longer have a multimeter on hand to test the voltage. Luckily though, we know clipping does resolve the tint, so there is a solution which doesn’t require any alterations to the existing hardware. You’re saying you suspect the tint is from negative voltages, not positive, correct?
1 Like
Right. We’re all good, as far as FKG3’s performance goes. Thanks for taking the time to try the clipping test, that resolved my primary question!
1 Like
Of course @creatorlars! So I have two more questions for now:
- Would you expect SMX3 to have any kind of tint with these weird voltage levels?
- Would ESG3 be able to clip the voltages at the end of the chain?
Not natively, but they might be introduced (as is the case in FKG3) by patching in signals outside the input headroom of the module’s +/-2V input range. We’ll just have to run some tests here in the lab to gather more information in order to characterize exactly what’s going on. It’s odd that the anomaly has a blue tint at all – and that’s what’s making me guess this is tolerance related (a tiny and normal offset from 1% resistor tolerances is being amplified and becoming exaggerated in the error state).
CBV001 is technically a composite video source – with it or any other composite video source, you should be using a video input amp, like the Cadet 3, TBC2, or Visual Cortex, to input your signal. But we’ve found a great hack and workaround on that requirement in this thread with FKG3.
If the question is “which Gen3 module makes the best post-processor / clipping backend for CBV001” I would definitely prefer the FKG3 patch you tested with. You get a lot to work with there: high gain/exposure based luma proc, amplitude/VCA modulation via inverted softness, a DC mixer input via threshold, and you’ve still got the ability to patch different RGB textures (from ramps, shapes, etc) into the background and foreground, etc. Try to patch it that way, and modulate it as a CBV001 post proc, it may unlock more than you initially think.
When “Softness” is high, think of “Threshold = Brightness” and “Softness = Reverse Gain/Contrast”.
3 Likes
In fact with that rig you have in OP you kinda gotta run one of those FKG3 on the CBV001 output. You’ve got a Vidiot style luma proc on the output that way, and it even has a solarizer (the outline/rectify mode.)
3 Likes