I have an audio clip that, even when normalized to -1 dB, is still soft. How can I basically squish the waveform down to allow more room to add gain to bring the levels up to an acceptable point? I swear I’ve done this before, I just cant remember how!
The compressors will do that.
Effect > Compressor
http://manual.audacityteam.org/man/compressor.html
Normalize to -1 before and after you compress. That way the only thing changing will be the compression.
Mess with the Ratio. The higher, the stiffer and louder the sound.
Koz
Somebody recently got burned with this: Export your raw edit or performance as WAV (Microsoft) before you start messing with it. put the safety export in a separate place like its own folder or thumb drive.
Compression is one of the effects we can’t take out later. Once you say OK and close Audacity, the filter and the sound quality are burned into the show forever. Even Audacity Projects won’t help you. Projects save a lot of things, but UNDO isn’t one of them.
Koz
Thanks for the info, but Ive already run the compressor. I feel like there must be a way to do this, I hope Im explaining it properly. Basically I want to keep the same loudness, but shrink the waveform vertically, to basically rescale the gain so that I can reapply amplify and get the levels up much further. Am I just smoking crack here?
I hope Im explaining it properly.
Maybe not. There are two tools associated with custom volume increase. Compression changes more or less all the waves increasing the low ones and suppressing the louder ones.
Limiter is the other one. That one you can pick which of the loud peaks “wake up” the tool and only those get reduced.
Stop micromanaging the blue waves for a second. What’s the goal?
If it’s just dense, loud sound you might like Chris’s Compressor.
I use it by increasing the top option from 0.5 to 0.78 to simulate the volume management from the local FM station.
These are the three waves I get. 1 is original show, 2 is default 0.5 compression, and 3 is 0.78 compression.
It’s significantly louder and denser with much less volume variation between the performers and segments of the show.
Koz
I appreciate you taking the time to help me. Imagine that you have all the levels even and nicely compressed, and basically maxed out just before clipping. Yet the sound is still fairly quiet compared to what you would normally expect, for whatever reason. Is there any solution for this, or am I simply stuck with low volume material?
I have an audio clip that, even when normalized to -1 dB, is still soft.
Don’t be afraid of 0dB. 1dB louder won’t make that much difference but since you are trying to make it louder, you might as well “maximize” it.
Nothing bad happens if you get close to, or hit 0dB.* Bad things only happen if you try to go over 0dB.
Also try the [u]Limiter[/u] with the Soft Limit or Hard Limit option (not Hard or soft Clip). Make sure you’ve got the latest version of Audacity because the limiter was improved awhile back.
You can use input gain or make-up gain, or normalize after limiting. For example, you can limit to -6dB and then Normalize (or Amplify) for 0dB peaks. If you normalize to 0dB Audacity may show red for (potential) clipping, and if that makes you nervous you can normalize to -0.1dB, but that won’t change the wave shape.
\
- If you are recording analog you don’t want to hit 0dB because you don’t know if the analog signal was over 0dB and clipping your ADC… But once you’ve got a digital file, Audacity knows exactly what the peaks are and it knows exactly how much gain is needed to hit exactly 0dB without going over (that is without trying to go over and clipping).
…And, some people (audiophiles) worry about “inter-sample overs”. But, there is no information between the samples in the digital audio data, so the “over” would be in the reconstructed analog signal. And, there’s nothing limiting the reconstructed analog signal to 0dB. It’s only the digital data (and the digital side of the DAC) that has a hard limit, due to a fixed-limited number of bits.
Unless you’re planning to render the final result to an MP3…
The first answer is the right answer, but there are additional options.
To increase the ‘loudness’ when the file / track is already at (or almost) 0 dB, you have to use some form of “dynamic range compression”.
The way it works is to reduce the difference between the highest peaks and the low level sound (the “dynamic range” of the music) so that the overall level can be increased.
There are several kinds of effects that can do this, such as Compressors (short name for “dynamic range compressor”), Limiters, Multi-band Compressors and Waveshapers (distortion effects).
The Compressor effect in Audacity generally produces very gentle compression, and the default settings will probably not have much noticeable effect on music that is almost maxed out. The “Based on Peaks” setting is better if you want a stronger effect.
For strong effects set:
Threshold: low, but higher than the noise floor
Noise Floor: just a little above the noise level of the audio
Ratio: high
Attack / Release time: short
Make-up gain: enabled
Compress based on Peaks: enabled
If the track has noticeable “spikes” (transient peaks) in the waveform, then the Limiter effect in Audacity can bring these peaks down very effectively. The “Hard Limit” option is usually the most effective for the purpose of “maxing out” the loudness (Limiter - Audacity Manual).
The Compressor and Limiter can both be used. Use the Limiter first, then the Compressor, then optionally the Limiter again if you want to “max-out”.
Mix the tracks down to a New Track, then mute that track, then save the project with a new name “Save As”. That’s your backup copy.
Now save the project with another new name. This is your working copy. (ignore making backups at your own peril 
)
Now, in your working copy, mute all tracks except for the new “mix” track, and un-mute the mix track.
Normalize the track to 0dB.
“Maximize” the new track with limiting and compression (see previous post).
Check that the mix still sounds OK. depending on the material, it may be necessary to re-Eq, or adjust the mix in other ways.
Export.
Note that if you are having the tracks “professionally mastered”, you would usually leave all of this to the mastering engineer. It’s near impossible to correct overly-compressed audio without unwanted side effects (hence the importance of backups).
Yes, music commercially mastered can (usually) sound louder than what you can do at home, but sometimes it’s your player. Some players shift gears between, say, MP3 and WAV. This can make matching volumes an uphill battle.
I have personally been burned by Windows “helping me” when I wasn’t expecting it.
Koz