What's "different" about bias tremolo?

[Mark Hammer]

I'm familiar with the electronic aspects of how bias trem is achieved, compared to other forms of tremolo.  But what is different about it, sonically, that motivates people to try and achieve it?  I'm not dissing it.  Rather, I'd like to have some sense of the parameters, and both quantitative and qualitative dimensions of it.  Given that classic bias trem involves the output stage of an A/B tube amp, is it the behaviour of the tubes that impart the quality in some manner, and not simply variations in the amplitude of the signal being sent to them?  If one was trying to mimic it in the analog solid-state domain, what would you have to do to the signal to make it "sound" like bias tremolo?

[PRR]

Not just louder/softer. Different distortion flavor.

You never(?) see it on organs because organists get overtones in other ways and like them stable.

[ElectricDruid]

The Stompboxology "Mo' Tremolo" issue has the best discussion of bias tremolo that I know:

http://moosapotamus.net/stompboxology.html

I reckon you could do something similar with some solid state device with a nice soft tanh curve - a OTA input perhaps, or a differential pair. Then you'd need to literally shift the bias on the signal going in, but that should be the easy bit. The hard part will be stopping it from throbbing at you when you're not playing.

[anotherjim]

The hard part will be stopping it from throbbing at you...

...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger...must resist childish snigger....
Phew! I did it!

[ElectricDruid]

To have a go at actually answering the question in the thread title, it's the way that the signal goes from loud-but-clean to quiet-but-distorted. And it's heavily assymmetric and soft-clipped distortion, at that, so it's not too offensive.

Basically, it adds a harmonic variation along with the volume variation, and it does it in the opposite way to what we normally expect. Instead of "quiet=soft, fewer harmonics" and "loud=harsh, bright, lots of harmonics", it's subtly the reverse.

[Mark Hammer]

That's exactly what I was looking for, Tom.  Thanks so much for that!  Could I mimic it in the analog domain?  Not likely.

On the other hand, if one used complementary LFOs, perhaps there's a way that the harmonic content could be generated where you need it.  Alternatively, the harmonic content could be produced all the time and filtered out during the amplitude peaks.

[ElectricDruid]

That's exactly what I was looking for, Tom.  Thanks so much for that!

Paul said the same thing, but he was being very concise!

Could I mimic it in the analog domain?  Not likely.

No, I don't agree! I think if you use a similar method, you'll get a similar result. That's why I suggested soft-clipping functions like the input of an OTA or a differential pair. If you then use an LFO to give you a bias shift on a signal going into one of those stages, you'll get something that's created pretty much identically to a bias tremolo and should give much the same result.

[BubbaFet]

Does not a well-tuned EA Tremolo kinda-sorta mimic that?