I was just looking at the Audacity Manual for my answer but I found it to be just a little too vague. Please bare with me as I tell a real quick story. That way you know my background in recording.
A long time ago, I started recording stuff to tape. I ran a professional sound board for a Church and recorded sermons to tape. The first thing I did was read the manuals. The tape decks all told me to set the max peak to 0db when using a normal bias tape. You went higher then that if you were using Cr02 or Metal tapes. I buy professional recorded music and I noticed that they get carried way with maxing out the meters. I know that people in recording debate about this. I don’t want this to turn into a debate.
I noticed that in the audacity manual they want you to set the max db to -6. Since I make my audio recordings for myself to use at work, around fairly loud things, I max out my recordings to 0.0db. I don’t hear any noticeable distortion doing this. Am I doing the wrong thing?
My second question is about the plug in: limiter. I originally thought that it just clips off the top parts of wave. Is this true or does it also compress the wave?
I don’t hear any noticeable distortion doing this (0dB). Am I doing the wrong thing?
In the tape world, the volume meter was arbitrarily chosen for the “knee” of the tape characteristic curve—where the transfer curve started to flatten and not follow the show. This is also the reason you should calibrate the tape machine for the tape type.
In digital, 0dB is the volume where the digital system runs out of numbers and stops dead. It just stops following the show. Nothing graceful about it. Once you clip off the tips of the waveforms because of data damage, they’re gone.
You can set the timeline to View > Show Clipping.
Audacity Clip Fix will try and reconstruct the waveforms for minor clipping based on wave history and wild-ass guessing. But that’s all it is. Clipping isn’t benign. It puts clicks, ticks, pops and harsh gritty trash in loud sounds depending on how bad the damage is.
-6dB to -10dB was chosen so most people can announce with good theater and expression without hitting 0dB. If you’re happy with your performances, then you can keep doing what you’re doing.
Effect > Limiter has several settings from Hard Clipping to Soft Limiting. Soft Limiting is a cousin to active compressors in that properly adjusted, you can’t hear them work. There is a Limiter in the Audiobook Mastering Suite of tools.
There is another tool Effect > Distortion > Leveler. That one works a good deal more like what you’re probably accustomed to. It takes a hatchet to the waveform peaks no matter what it sounds like. When Limiter came in, Leveler was going to be binned, but we suggested Leveler was a good tool to simulate air traffic controller, taxi radio and two-way communications distortion.
I wasn’t clear enough. I don’t record audio in real time at 0db. I do edits to get it there after the fact. I typically don’t run the wave into the clipper. I stop just short of that. I got confused because the versions of Audacity didn’t use to have a “red zone”. Then I started seeing people on youtube showing that you got to stop at -6. I started scratching my head.
We recommend the -6db for recording tom provide headroom.
As your final production step it is perfectly good practice to use Amplify of Normalize to adjust the audio level.
When Dominic Mazzoni first wrote Audacity he set the default level for Normalize to be -3 dB - we have subsequently reset that default to -1dB (personally I choose -2 dB).
Amplify wil always default to 0 dB each time you use it - while technically that is good and sound - there are some players that don’t play nice with sound maxed out at 0 dB - hence the little headroom we supply by default with Normalize.
The tape decks all told me to set the max peak to 0db when using a normal bias tape. You went higher then that if you were using Cr02 or Metal tapes.
It depends one the machine but if the machine has a CrO2 switch that should be taken care of automatically.
With tape you want a strong signal to overcome the tape noise. And as Koz says, tape as a “knee” and it tends to soft-clip (or “limit”) as you start saturating the tape, and the NAB tape equalization tends to further-soften the distortion. So with tape it’s common to go occasionally “into the red”.
Of course, digital has no “tape noise” and it has a HUGE dynamic range so there’s no need for a hot signal. Digital recording levels are not critical as long as you avoid clipping. It’s OK to peak below -6dB and it’s OK to get close to 0dB.
With live recording you need more headroom (safety margin) for unexpected peaks but if you’re digitizing records or tapes the analog levels are more predictable so you can get-away with less headroom, but it’s not necessary.
The Amplify or Normalize can be used to bring-up the volume after recording. Both are simply digital level adjustments. Audacity’s Amplify effect can be used to normalize, since it has the option of setting a “New Peak Amplitude”. The ]Normalize effect offers a couple of additional options.
I buy professional recorded music and I noticed that they get carried way with maxing out the meters. I know that people in recording debate about this.
My second question is about the plug in: limiter. I originally thought that it just clips off the top parts of wave. Is this true or does it also compress the wave?
The limiter is actually very-good! With most settings it uses “look-ahead” and it won’t hard-clip. But, it won’t help if you’ve already clipped your analog-to-digital converter during recording.
Limiting is a kind of fast dynamic compression. Hardware compressors are often sold as a “compressor/limiter”. Clipping is a bad kind of dynamic compression.
