lopsided waveforms

[BRIEFCASEMANX]

Usually when i record waveforms they are usually skewed towards the top a little, so when i normalize it makes it so that the top of the waveform is almost clipping but not quite, but there still some space on the bottom of the waveform. I think this is creating a problem with loudness in the mastering phase. I am not able to even close to match the loudness of commercially recorded cd's even when compressed. Could this be part of the problem? I can move the waveform down in the effects<amplify menu using dc bias adjust 20% or so, then normalize, but would this cause any problems anywhere? I have the adjust for dc option checked in the settings menu, but it doesn't seem to help. When looking at commerically recorded wavs they seem to be ~equal on both ends of the waveform.

unrelated: do you guys know of any good DX analog tube/tape distortion plug-ins? What are the ones to stay away from?

 

[ocnor]

Instead of Normalizing try using the Hard Limiter.

 

[dobro]

Sure, and after he hard limits, he'll want to normalize, at which point...

I'd try the DC bias adjust and then normalize. Then I'd compare that approach to the first approach, where you just normalized without the DC adjust bias. Listen. Crank it. Play it quietly. My guess is that taking the DC bias adjust approach will give you what you want without any deterioriation in the sound quality of the track.

 

[Jim Y]

Yes, DC bias adjust is the way.
But the real culprit is the soundcard. By recording with a DC offset, you are throwing away some quality since the full resolution of the analog-digital converter is not available. It will clip before all the bits are used.

Some sounds do and should show some offset, but this will be more curved - I'm thinking of the positive rise and fall you see on a kick drum sample for instance. This is the initial forward pressure of the drum skin. Although actually it's a low frequency componant - should you correct this with DC offset? I've read of someone who can hear if the kick drum mic was out of phase - in which case the recording has a negative curve offset. I keep meaning to do an A-B test to find out for myself, but so many more things to do...

Get a decent recording card and you won't ever have to worry about offset.

 

[dobro]

"Get a decent recording card and you won't ever have to worry about offset."

I'm willing to believe that, but I've had this experience with my one and only soundcard (not a bad soundcard, I think): sometimes (not often) I've recorded something and got DC bias. Most of the time it's okay. Which suggests in this case, the problem has more to do with the way I'm working the mic or something like that, right?

 

[Jim Y]

Maybe what you're plugging into it AND the card is at fault. Signal inputs normally go through capacitors to "de-couple" any DC on the external equipment from the internal circuits. Similarly, outputs also often have capacitors to stop the equipments internal DC from feeding out. Cheap capacitors are "leaky" and let some DC through giving this offset. It will be worse with a soundcard microphone input as this has some gain which can amplify the offset which might only be a few millivolts in the first place and not show on line-in recording.
Balanced circuits often have no de-coupling, but as the amplifiers work on the difference between hot and cold wire signals, the average level is zero. A fault causing an inbalance can cause an offset too.

I've never seen any significant offset with my Audiophile2496. This is always fed from my mixer so the connection is constant, whatever I'm recording.

I have seen offset in .wav samples provided on CD, but I think they were pretty old products.

 

[lpdeluxe]

Try Options/Device Properties/Wave In; check "Adjust To Zero DC When Recording."

 

[BRIEFCASEMANX]

In the Options<Device Properties menu I check the box "adjust to zero dc for recording"checking the DC offset box does not make any difference. I tried recording something with it checked and then with it unchecked and it looked exactly the same. I have a Soundblaster Audigy 2. I thought all soundcards had offset in order for you to use a certain type of mic (electret or something).

I don't understand what the bottom and top peaks of the wave really are.

UPDATE: I guess some analog synths do this according to some other people online. i plugged in a microphone and it seemed about centered. the problem is that I guess the analog synth seems to make skew a mixed waveform to the top somewhat. Rectifying it with limiting seems like it would make sense if the limiter acts the same way with the top of the waveform as it does on the bottom part but in reverse.

 

[Jim Y]

Maybe we're on the wrong track here.
DC offset appears even if you record silence. Does it?

All natural waveforms will have some asymmetry about the centre-line. This is due the waveform being a composite of all the individual sounds added together - sometimes this produces a strong positive (top) or negative (bottom) peak. These are annoying as they prevent normalising beyond these odd peaks, but they are completely natural.
If you imagine recording a simple sine tone - a nice smooth wave - and say if on a positive peak there was a short click sound added - then the recording will show a top spike at that point but not at the bottom. Proving that you cannot expect complete symmetry.

Synths can indeed produce offset waveforms - pulse waves for instance, but natural sounds are not consistant enough to appear like that.

Hard limiting can force a level boost with asymmetric recordings. Before the Hard limit was available in CEP, I used to zoom in on those odd peaks one by one and apply the Amplify with the 3db cut preset - doing this to just these peaky cycles had an inaudible effect, but allowed the Normalise to go 3db higher! I believe this tactic has a more natural result than Hard Limiting, but not sure why!

Creative cards use a TRS mini jack for the mic input - the ring contact is 2.5volt DC phantom power for condensor headset mikes. However, this DC voltage should be removed from the signal by the pre-amps input capacitor.