Spoon

If you have an amp (or other device) that accepts front and rear channel inputs and sums them to a pre-out jack for non-fading bass, anyone know specifically how that's done? Is it a simple summation or a weighted average? So, if you had two channels, X and Y, would the pre-out give you case #1 or #2?

{1} X + Y = [X + Y]
{2} X + Y = [(1/2)X + (1/2)Y]

elduderino

I don't know the way that mixer IC's work exactly. But since if you fade all the way one way or the other you can notice the volume difference on the non-fading sub out, I would assume that mixer IC's do #1. If they did #2 the delta would be less then 3dB and we wouldn't notive the diff.

Spoon

Maybe I'm just looking at this the wrong way, but since front and rear channels carry the same signal (with potentially differing amplitudes), Y = X and I probably should have written it (A)X + (B)X = (A+B)X. The point being that, if say you played a sine wave thru the left channel for example, and with the fader control dead center, case #1, X + X -> 2X, whereas case #2, X + X -> X. So, in this instance, would a summed pre-out have double the amplitude of an individual input signal or the same amplitude (allowing for all real world inefficiencies)?

elduderino

I would expect that if you summed two identical AC signals you would have 3dB more in level. I also suspect that acoustic engineers making preamp ICs for HU's probably took this into account... I think the answer to your question gets into the additive characteristics of AC and voltage and dB. But as I'm sure you know, it's not the best place to ask this kind of question : )

Rather than theorizing on someone else's execution, the best mode of answering is to use a sine wave test CD or generator and a DVM. Empirical testing beats hell out of wank...I mean theorizing. ; ) But I'm too busy todoay to test it, it'll have to be another day.

Spoon

Agreed, this is not really the proper "forum", but I couldn't refine my google search enough, so I figured I'd throw it out here and see if I could get anything useful. The key to my search was including the word "decibel", so thanks. While I haven't yet come across the answer to my question, I got enough information that the method employed is now somewhat irrelevant. For instance you raise a third possibility (and I'm sure there are plenty of others), where it could be X + Y = [X + Y] and the 3dB gain is attenuated. Or not, maybe there is a standard methodology, maybe there isn't, probably a host of implementations, I don't know, and since I have no way to measure it and no guarantee that all devices behave similarly, case closed (kinda).