Ever heard of a bass phase aligner?

[Angstrom]

Hey all, I was thinking of a bass mixing technique over the last few weeks/ months, but I'm not sure if a product exists like this or if it is even possible.

The problem:
Below about 250hz waveforms start to become more audibly susceptible to phase cancellation, due to the longer wavelength two " bass" waveforms with opposing polarity will sum to 0. Of course this happens to higher pitched sounds too, but due to the higher rate of change (hz) and because the higher signals are usually stereo, the issues are noticeable mainly in transients. In a group of bass waves ( a kick, a bass, the low end of a pad, the bottom of a snare, the low end of a tom" ) all of these are being summed, and not necessarily beneficially toward a more powerful sound. Usually we find that basses clash to become muddy and flabby. Additionally we have a tendency at frequencies below 100hz the tendency to "tune" kicks and drums to the key of the song, and may actually give more problems for a coherent bass.

So imagine we have a usual mix, a slice of time of one sample length where the Bassline fundamental is at 100hz and it's phase is at +0.5, and simultaneously the kick is at 101hz, and it's phase is currently at -0.5. So : Our output of these full ranging signals is 0!
Shouldn't it be 1, or -1. Two loud sounds, should not really cancel, that makes our mixing very difficult.

My concept
I was thinking about some kind of phase coherence device which takes a silent signal as the phase master, then applies it to all of the low band signal. I'm sure you are familiar with sync and softsync from synth oscillators. And I'm sure soft-syncing the incoming low end carriers to a silent modulator bank (a master phase at every frequency) would sound terrible. But I cant help but think that some method of unifying low end phase would potentially be a way to deliver much more powerful, solid, low end.

My not-really a solution
It would have to be an offline solution, but I was thinking of an FFT analysis of each stem, and then each FFT band is normalised before it was summed to produce a coherent phase low end


Thoughts? Have I finally lost a grip?

[Steve Glen]

Ducking your sounds correctly gets most of the problems out of the way and should be the main place where everyone starts. but beyond that, if you wanna get into the details....


This, to my ears, sounds like it works.
http://www.soundradix.com/products/pi

[Angstrom]

This, to my ears, sounds like it works.
http://www.soundradix.com/products/pi

thanks, that's exactly the sort of thing I was thinking of. I was sure I wasn't the first to have considered phase relationships between elements in a complex mix.

[Tarekith]

Steve, I understand your idea from an analysis standpoint, but I don't get your explanation for what your processing would actually do. How would you reconstruct the correct volumes in each band if everything is normalized during processing?

[Angstrom]

Steve, I understand your idea from an analysis standpoint, but I don't get your explanation for what your processing would actually do. How would you reconstruct the correct volumes in each band if everything is normalized during processing?

to be clear about what I meant by normalised, I don't mean "raise everything to max", I meant polarity. So if we have a bass wave of 100hz at +0.5 , and a second bass wave at -0.5 we would currently mix them together and get zero -0.5 +0.5 = 0

I was saying that by rotating the polarities of our incoming tracks (on a frequency band basis) to conform to each other we could get
+0.5 +0.5 = 1

[memes_33]

while this might address phase issues due to left/right alignment, it does not address the bigger problem of phase issues due to early reflections.

i use a subwoofer with an internal crossover, so anything below the crossover point is mono-ed and comes out of the subwoofer. this alleviates any phase issues of right vs. left. that is, unless, the left/right signals i'm sending are not phase-aligned, but i don't know why anyone would ever do something like that on purpose.

[Angstrom]

I'm not talking about left/right bass phase for mono compatibility, I'm talking about summing bass frequencies (potentially already mono) in the mix stage. And the phase misalignments between tracks and their consequences, which are usually solved with an ugly sledgehammer of "kill all the bass with EQ carving"

Picture three mono sine waves, we can pretend one is a kick drum one is a bass and one is the bottom of a pad. Ignoring the fact that most people "carve" these frequencies, and assuming a new aproach. For this example/ explanation - We have these three sine audio streams , A is 60hz, B is 61hz, C is 63 hz.
Now you know that a wave plot goes up and down on the Y axis, the volume goes up in the positive phase, then down, then negative phase increases and decreases. We see the familiar plot of a sine wave.

Now remember that if Sine A is currently at its loudest and sine C is also currently at its loudest, what would you expect to hear? Answer : two loud sines. But if sine C is currently in its negative phase, which is still "loud" remember, but inverse of sine A. What now?

Sine A=1
Sine C=-1

Add them together, and what do you get? Is it a double-loud signal. Or is it zero!

This is my point, that by phase aligning we can reclaim lost volume. It's highly speculative though, and my own efforts to produce one would doubtlessly just produce a multiband sync mod.

[nightscope]

[Angstrom]

Thats not what I'm talking about.