Waveforms not centered on the 0 line

[kinnylandrum]

I have a recent project on DP 8 that, when I bounce it, the waveform of the bounce looks funny. It sounds OK, but when I look at it on Peak or DP, it looks like the waveforms are not centered on the 0 (zero) line. In fact the right channel looks worse than the left. This also happens even if I record the track in real time onto another DP track.

In the old days, I would sometimes see this phenomenon with some samples, and I used something called Remove DC offset in either Peak or Sound Designer (shows how far back I'm going). So I tried that, but it just made the parts of the wave that WERE on the line now go below it, and it led to a pop or thump when the file reached the end. If I just tried it on the parts that were off the line, it meant a pop when going to or from other parts of the waveform.

Does anyone know why this happened? How can I avoid it? And should I even care?

Thanks for any insight anyone may have.

[stubbsonic]

You should definitely care. Anything that is adding DC (digital or not) should be tracked down.

This happens when you bounce or bus no matter what plug-ins, etc?

As an aside, when I record trombone- with the mic fairly close, it's a pretty lop-sided waveform. The trombone doesn't have much of a below-zero excursion.

Hope someone can track down this gremlin. I sure wouldn't want to ship a mix with a bunch of DC.

[kinnylandrum]

Yes it happens whether I bounce or record to a bus. But it appears to only happen on this project.

[kinnylandrum]

Believe I think I've figured this out. It seems to be a function of using Wavelore's Ether-Phone, a very good theremin plug-in. It seems any output from this plug-in is only "above the line", that is only positive. I don't know why this is, but I found I could put DP's plugin DC Notch on it and it would make the output stay in the middle. So that's what I'm going to do. In fact, I'll be using this Theremin plugin a lot in a movie I'm doing, so I will continue to use the DC Notch on that VI.

[Prime Mover]

I actually haven't had much luck with the DC Notch plugin, but an even better solution, and I recommend doing it anyway for all tracks, is to throw a high-pass filter on it, just below the frequency you want the instrument to occupy. All the DC Notch plugin is is a super low high-pass filter, at like 1Hz or something. DC offset is often cured by this. I have a tube preamp that puts out a little more positive than negative, and throwing a nice high-pass usually takes care of that.

Thing is, low-frequency energy is EVERYWHERE. Typical room ambience shows some, some circuitry, bodily movements, foot tapping, etc. A lot of it may be well below human hearing range, but when it comes to compression or any dynamic processing, it still adds up. There are typically only three instruments that should extend into the bass range: kick drum, bass, and keyboards. Of those, only kick drum should extend into the sub-bass range, and keyboards typically get chopped well above 100Hz when playing with a bass instrument. I throw a high-pass filter on ALL my tracks, even down to kick drum, and set the frequency where I want the instrument to start sitting in the mix. For kick it'll be around 20Hz, for bass, it'll be around 40Hz, for piano, I could well go into 120Hz to get it out of the bass's way... and then it all goes up from there. And while you're doing that, you're taming DC offset automatically.

That said, there are a few different types of asymmetrical waves, some of which will be tamed by high-pass filters, and some which won't. I'm not an expert on which will or won't. Some are natural, others are caused by equipment. Most of the time, when you really get down to it, the center is still at 0, but one pole will be steeper than the other. This is typical of brass or high SPL instruments, as well as some more gentle equipment issues. What you want to watch out for are waves which really actually DO center around something other than 0. That is a sure sign of some bizarre signal problem.

[philbrown]

Thanks for this thread. I have had the same question recently, and like Eric I found the DC notch plug-in to have little or no effect. I'll try the high pass filter trick- as a modular synth guy I should have thought of that!

Side note to kinnylandrum: I still listen to and am fascinated by the the Paul Gauguin song (Jimmy Webb) you turned me on to. I make it about 3/4 through chord charting it and then hit the wall. One of these days...

[stubbsonic]

Thanks for the high-pass tip. I will try that on some trombone tracks to see if it centers things a bit.

EDIT: Just tried it. I used MW EQ and the 36 dB HP. I tried at a low frequency which had insignificant effect. It wasn't until I moved up into 100-200Hz that things centered more. I guess I could also try re-applying the same high-pass freq to get a steeper cut-off. Still it's very useful to know.

[kinnylandrum]

Actually I have found that DC Notch solves my problem completely. It seems this Ether-Phone VI from Wavelore, while it is a really great Theremin VI, had ALL the energy from the output above the zero line. Putting DC Notch clears it right up.

BTW Phil, I'm glad you like the Paul Gaughin song. There's a new CD/DVD combo from 1988 with Jimmy Webb and Glen Campbell called In Session, from a Canadian TV show that's pretty good.

I have to return the compliment to you for turning me onto to Jellyfish, now one of my absolute favorite bands. In fact, do you know there are two "new" Jellyfish albums out. One is a package of the live Bogart's show. Most of that is on the Fan Club box set which I have so far been unsuccessful at buying for a reasonable price, so I might get the CD. And there's a release of all the backing tracks from their two albums, called Stack-A-Tracks. Their tracks are so great it would be interesting to hear it, but I don't think I want to spend money on what is essentially a really good karaoke disk.

[kgdrum]

Jellyfish is a great band! I love Spilt Milk & Bellybutton both of these albums are awesome!

[FMiguelez]

But what are the consequences of having these "un-centered" wave forms?
How do they affect a mix?
Is it really important to center them?

I've never had them or even seen them, so I don't know.

[MIDI Life Crisis]

It would seem on the surface that the audio would be lopsided in the mix. No?

[kinnylandrum]

Yes it is a little lopsided, but you don't really see it on the meters, you only really see it once the file is made and you look at it on a waveform editor, Peak in my case. It does mean that the wave is stronger above the line (not necessarily on one side of the stereo spread) so if you were to raise its level it would probably distort sooner. All I know in my case it it was exclusively with Ether-Phone, and putting DC Notch on it conquered it (and made the average level of my bounces as per Peak about 1 to 2 db softer).

[Prime Mover]

But what are the consequences of having these "un-centered" wave forms?
How do they affect a mix?
Is it really important to center them?

I've never had them or even seen them, so I don't know.

If anything, it makes mastering more difficult. Think about it, even though a wave might only be 2/3rds full energy, it may reach the peak of the raster on one pole, any compressor or limiter would see it as being such, even though the perceived loudness would be less. Even though it might be completely normalized to 0, it might be -6dB in perceived loudness! It's a waste of loudness. Not so destructive to the sound of a mix, but bad for getting a nice hot master.

If I remember correctly, the steepness of the high-pass filter effects the result of the re-zeroing. Try a really steep curve. I think the DC notch is an incredibly steep curve at 1Hz (maybe 10Hz).

[FMiguelez]

That makes sense.

Thanks, Eric

[stubbsonic]

I'm going out on that limb that is beyond-my-knowledge-sphere... so I should perhaps phrase this as a musing question:

With the nature of a loudspeaker being a somewhat "springy" medium, and assuming the power amp would send the DC forward to the speakers, would sending it signal with a DC offset be essentially asking it to generate an accurate wave whilst pushed or pulled off of it's resting position? And might this create some harmonic distortion?

I understand that speakers do this all the time, i.e., generate other content while also generating bass frequencies-- but I would imagine that the linearity of fidelity decreases at the extremes of speaker's range of motion.