When you convert Stereo to Mono, frequencies...

[zgmg]

I'm trying to understand what happens on a waveform level when converting a Stereo to Mono.

Frequencies cancel out and contribute to one another yes?

Is there anything bad about it for understanding overall Zero-Point Crossings? Versus per channel?

Ty.

[David]

Converting stereo to mono means summing two channels. Saying that "frequencies cancel out and contribute to one another" is an over simplification, but you could say that if you have two signals at the exact same frequency but out of phase in both channels, then they'd cancel each other, if you had two signals at the exact same frequency but in phase, they'd reinforce each other.

As for zero-crossing, judging from your other posts, I believe I know where you're going at, but reading the zero-crossing of the sum (mono) will give you ZERO information on the zero-crossing locations of the two original channels.

[zgmg]

Thanks David - yea and that's good for me. Zero information means no sound so that's what I'm looking for!

[David]

You misinterpreted what I said. Zero information, means you still don't know where the zero-crossings from the original left and right channels are by looking at where the zero crossings are on the mono sum of the channels.

Simple as math.
• If I tell you the sum of two numbers is 5, can you tell me whether one of the two numbers was zero or not? No.
• If I tell you the sum of two numbers is 0, can you now tell me if one of the two numbers was zero? No.

[zgmg]

Alright... so Zero-Crossings + Mono is the best way to do it?

[zgmg]

Also David, in light of your math thing, I think it'd actually be:

If I tell you the sum of two numbers is 0, can you tell me whether one of the two numbers was zero or not? No.
• If I tell you the sum of two numbers is 0, can you now tell me if one of the two numbers was zero? No.

So the point is the sum is 0, the numbers that make it so, anything. But the point is that there is no visual waveform nor sound that can clip at that moment.

[David]

Alright... so Zero-Crossings + Mono is the best way to do it?

To do what?

In regard to my math, I wanted to point out 2 things.

1) that a zero sum does not mean one of the numbers was a zero. (so a zero crossing on the mono sum does not mean one of the channels had a zero crossing)
2) that a non-zero sum does not mean that one of the number wasn't a zero. (so no zero crossing on the mono sum does not mean there wasn't a zero crossing on one of the channels)

[zgmg]

I get you. For my purposes, it works. I'm creating a virtual Zero-Crossing for a Stereo-to-Mono reference.

[David]

zgmg, I have no idea what you're talking about (there's no such thing as a "virtual zero crossing point", and I'm not sure what a "stereo to mono reference" is either) but as I already stated in another one of your threads, with all due respect I don't believe you understand what you're doing. If you wish to detail what that last post (a virtual zero-crossing for a stereo to mono reference) means I can detail further, but for now, your answers show me that you don't understand what I'm explaining.

[zgmg]

Stereo = 2 channels. Mono = 1 channel. Finding perfect chops is practical in Mono as you are isolating the Zero-Information crossings as chop locations. Using those locations, you can then go back to your Stereo source and make chops using Mono chop locations. Simple as that.

[ski]

Dude, you need a drink. Maybe several. Anything to get you to relax your obsession with zero! Perspective: most people get on with sampling and looping material, but you're stuck on zero. What's wrong with this picture?

What makes other people's samples of (random) material usable? They know how to deal with 'non-zero' crossings when they happen. And as I said in another of your posts on this subject, you need to take time to learn how to manufacture "zero" if it's not presented to you in the samples themselves. The tools are there, same ones everyone else uses.

Oh, and zero crossings? In many cases there simply is no such thing. Close to zero? Yes. But exact zero is more hypothetical than real. I'll leave it to you to figure out why.

[zgmg]

Heh, ski - I get it. I'm obsessed with perceivable perfection - making the absolute most out of what you have and nothing less.

Anyway, I'm very happy with this theory and I'm going to use it. I just wanted to share.

[ski]

Right. But when "zero" doesn't exist, and your theory falls down flat, what are you gonna do? You wanna know how many times I've encountered a stereo file where at my chop point, one channel was near zero and the other one was hangin' up in the air? Now, when perfection doesn't stare you in the face at your chop point, you can attempt to create it using the tools we've discussed in your threads before, but expecting to get perfection just because you want it is Quixotic. Hey, tryin' to give you some real world perspective here.

[David]

Stereo = 2 channels. Mono = 1 channel. Finding perfect chops is practical in Mono as you are isolating the Zero-Information crossings as chop locations. Using those locations, you can then go back to your Stereo source and make chops using Mono chop locations. Simple as that.

Sorry zgmg, what you say doesn't make any sense at all. I thought I explained that, but apparently I failed. I just explained to you that what you see as a zero crossing on the mono sum could well be the highest point on one channel and highest negative point on another (-100 + 100 = 0). So you'd end up chopping at the worst place to chop, the furthest from the zero crossings on both the left and right channel. So what you're looking at as a zero crossing could very well yield the nastiest, noisiest cut you could get.

Anyway, I'm very happy with this theory and I'm going to use it.

I'm sorry but your theory is completely flawed.

In any case, I think you misunderstand what a zero crossing is and why it may (or may not) be important to edit on a zero crossing in the first place.

[zgmg]

Apparently I don't understand at all. All I'm getting at is that if you combine a Stereo sample to Mono, you can better analyze the waveform in one dimension than two, and in doing such, much more rapidly find locations where the waveform exhibits new sound information. Using that location, you can apply it to your Stereo sample and create accurate chops. I'll leave it at that.

[ski]

All I'm getting at is that if you combine a Stereo sample to Mono, you can better analyze the waveform in one dimension than two, and in doing such...

You're ignoring both dimensions.

[zgmg]

What's it matter of the outcome is still 0?

[ski]

If your purpose is to make samples out of mono-ized stereo source material then indeed, your method will work just fine but only if L & R sum to zero at your loop point. However, that doesn't seem to be the case. You're using summing as an ersatz diagnostic tool to create stereo samples. And as David has correctly pointed out, X + Y = 0 with the big question being: what value is X, and what is Y?

The answer is that they can be anything, anything, anything at all, but most of all, not necessarily zero!!! So your method is flawed.

[zgmg]

Oh. I get that. Except I learn things don't click if they're followed by sound or music. It's only an issue if there's nothing following them. Even then, that's what fades are for.

[ski]

If there's nothing following your chop point then chances are that you'll be closer to zero than if there is sound occurring afterward. So what you're saying is the reverse of the norm.

Indeed, fades are for manufacturing a hit-zero mark. In this case, true zero (provided there's no DC offset in the sample). So you got that part right. But really, you're wasting your time pursuing this method. If you're looping stereo samples, deal with L and R as they actually exist.