Weird Clipping Problem

[Syncretia]

None of my tracks are clipping yet the master track IS clipping. WTF?

This is happening in a couple of sections in my song - particularly with a clap and a snare. These two tracks are going nowhere near clipping yet every time they come in, the master track clips. If I turn them right down, the master doesn't clip, yet they are barely audible. I've tried putting upper ceiling limiters on the track but it makes no difference either. I'm really at a loss to figure out what is going on. I do have quite a few sends in my track but on the clap and the snare I've turned them right down and the return tracks are only sitting about 30% of the volume that would cause clipping. When I solo the clap however, there is no clipping at all. It only happens in the mix.

Am I wrong in saying that if no tracks are clipping, the master track should not be clipping? The master track is on 0db.

Could this be some glitch in Ableton's summing engine that accidentally doubles volumes when certain frequencies bump in to each other or something?

[masterblasterofdisaster]

"Am I wrong in saying that if no tracks are clipping, the master track should not be clipping?"

You are sometimes wrong in that assertion.

[Syncretia]

When would I be wrong? Under what conditions? How would I go about diagnosing the problem?

[Tarekith]

Think of it in this (grossly over-simplified) way. The master fader is the SUM of all the individual tracks in your project. So if you have one track peaking at 5, and one track peaking at 4, the master is not going to show you the value of the highest track (ie, 5), but both tracks combined (ie 9).

The actual maths are not at all like this, but you get the idea.

[Syncretia]

Hmm.....

I wouldn't have thought that that would be quite right but it appears that that is what is happening.

If this is the case, then why is that under most conditions, I can stack several instruments which are almost peaking but the master track doesn't clip?

Why is it only happening in this case? Is it perhaps something to do with a precise frequency that overlaps?

[Syncretia]

So if you have one track peaking at 5, and one track peaking at 4, the master is not going to show you the value of the highest track (ie, 5), but both tracks combined (ie 9).

I did a kind of test. I stuck an Analog on one track and made it almost clip. The master track was not clipping. I then duplicated the track. The sounds was a bit thicker, but no clipping on the master track. I then duplicated it a few more times. Not much change in the sound, but the master track still did not clip. In this case several tracks were at almost clipping level and were all playing at the same frequency. So, while what you said may be true in some cases, it's not true in this case.

It might be the case that all things being equal "if you have one track peaking at 5, and one track peaking at 4, the master [WILL SHOW] both tracks combined (ie 9)". It's the all things not being equal I want to understand. Under which cases does Ableton's engine Add sounds together, and under what circumstances does it Average sounds together? It appears that my mix has hit upon a point where it is adding them together instead of averaging them. What could cause this and how would I determine it?

I ended up having to turn the clap right down in the mix to avoid clipping.

[Tarekith]

Sure, a lot of it will depend on the frequency and phase of the audio on the tracks. You'll definitely get some places where the signal combine to form a hotter signal, and some places where they combine and cancel out.

[Syncretia]

I'm not getting this.

[Der_Makrophag]

Does this, maybe, have to do with the headroom calculations, Live uses to avoid clipping in Tracks but not in the Master Channel?

I also thought, that adding two identical waves, each one almost clipping has to result in clipping. Did you try the experiment with samples? These are usually more "phase-locked" than synthesizers. Choose 1 Sample and lay two copies of it exactly in sync with each other in arrangement.

[Syncretia]

Does this, maybe, have to do with the headroom calculations, Live uses to avoid clipping in Tracks but not in the Master Channel?

I don't know....

Did you try the experiment with samples?

I just tried it now with the sample Kamoni Hermitfrost.aif in Ableton's loops folder. It's a good sample because it almost clips. I put the master on 0db. I put the sample at just under -1db. No clipping on the master.

I duplicated the track - the sound was a bit more full but not louder and and no clipping.

I duplicated it up to 8 tracks - no clipping.

I tried putting a phaser on some of the tracks and inverting the phase - slightly audible effect but no clipping.

So, I must say the whole thing is quite baffling for me. What we can say is this:

-Under most circumstances, if 1 or more tracks are not clipping, and the master track is at 0db, there will be no clipping on the master track
-Under some unknown circumstances where there is more than 1 track, two signals will combine to make a "hotter" signal and cause the master to clip.
-According to Tarekith this is related to "frequency and phase"

But, I'm just not sure what the relation to "frequency and phase" is. I mean a can't reliably recreate a situation where I can create clipping on the master track.

Oh well.

[Syncretia]

Actually, I can't recreate the problem anywhere except for the one track that I mentioned.

I've never had this problem before.

In all the tracks I have ever mixed, I've never noticed the master track go red unless one of the other tracks went red. I'm not saying that this is not possible, I'm just saying that something different is happening on this one track I'm working on that is not happening on other tracks. It's the exception - not the rule that is bothering me.

[Tarekith]

Do you have a limiter or something on the master? Because otherwise this isn't making sense at all.

[yur2die4]

I'd suggest looking at that clap in Spectrum. See if there are any unnecessary frequencies pushing you off phase. From there you can EQ things out (don't need any bottom).

[Syncretia]

Do you have a limiter or something on the master?

Do you mean in my tests, or in the problematic track? I don't have a limiter on the master in either case. However, I have tried putting a limiter on the clap track with the problem and it doesn't fix it.

You can recreate my tests very easily. Just duplicate samples that are almost clipping. You will see that the master doesn't clip.

I'd suggest looking at that clap in Spectrum.

Will look at this.

[yur2die4]

Also, that clap could have a DC Offset issue. I don't think it is super likely but it could. It happens when for instance, you have an ultra low frequency being recorded, it moves the waveform in big sweeping fat waves. Any higher frequency information introduced to it uses that wave as the crossing point for the wave action. So if the wave is way off center, when it is recorded, it pushes all of your sound off center (this is waveform center, not panning center). Since it is away from the center, it also means that any amplitude of vibration happening can peak that much easier, since it is closer to the edge. It is not the best situation for optimum levels. Now also imagine that this could happen with pretty much any samples as they are summed.

Perhaps you could make an audio channel that records from the master. Then do a screencap of the waveform so you can see what is happening.

Since peaking can result in a clusterfuck of squiggles, you may also want an instance recorded that is Not clipping. And one last one where you mute all elements except bass, kick, and clap. Something like that

Or not haha. Sounds like a lot of work