I mainly use Audacity to record voice overs for video tutorials. Given I’m completely rubbish at reading a script, I record, each sentence, or parts of a sentence, into individual clips. However, if I re voice a section, I may get the mic placement in a slightly different position and the clips can then sound a little different to each other.
One thing I would like to do it normalise each clip separately to the same level, or indeed apply other effects to the individual clips separately rather than the entire track. I could do this manually, but given I have a number of recordings, each with many, many clips within them, then this would be very time consuming.
There’s a trade off between quality and effort.
For “professional” quality, you have to take the time and effort to set up correctly every time, and the rule of thumb is “10 hours editing per 1 hour of audio” (can be a bit less or much more).
The special “RMS Normalization” should match the volumes. (Regular peak normalization does not work as well). Do that separately for each recording/clip.
Since RMS Normalization (at -20dB) usually pushes the peaks into clipping/distortion, the limiter is used to “tame” the peaks (and to bring the peaks into spec for the ACX audiobook requirements).
And since you’re not trying to meet ACX requirements, you can follow-up with regular (peak) normalization. That will bring up the volume (RMS and peak) by a couple more dB.
The LevelSpeech2.NY is, er, horrid! The effect is very odd, not quite mechanical but it gives a non-human like quality as you lose all tone an inflection.
I must admit, I am not an audio professional and do not have as sound studio, but I have tried to do what I can with what I have. The Mic is a Blue Yeti and even though I seemingly get this position as good as I can each time, even every small variance in my head position can cause noticeable changes (at least to my ears). Perhaps this is also a natural change in my state of being from one day to the next…
The Audiobook Mastering link is very useful, however. Whilst I am not needing to hit a specific specification, it’s not bad to have something in mind when you produce these clips.
@steve, you recommend still manually selecting each clip and running the same procedure each time, rather than against the whole track (as is outlined in the guide)?
Effectively, what I have just started doing is removing background noise using the Noise gate (which is pretty good) on the whole track, applying a bit of compression, then manually adjusting the peak levels, and then normalising a peak value. I was also just looking at Macros and was wondering if they could be used to cycle through the clips individually.
I would recommend taking more care about your microphone setup.
The distance between microphone and mouth does not only affect the volume, it also affects the timbre of the voice, and the perceived acoustics of the room. Even if you match the volume perfectly, there will still be a noticeable difference in the sound of the voice.
If you are able to leave a microphone set up with a pop-shield in place, then it shouldn’t be too difficult to get the same mic position each time.
The default settings are very aggressive (for “fixing” very bad recordings).
If the original sound is not too bad, then you need to use much “softer” settings. As a starting point, try settings at 50, 0.2, 0.5, 0.8.
For even more gentle effect, try 10, 0.2, 0.5, 0.9.
Yep - pop shield is there, and even setup using the same method as was indicated in the article @DVDdoung pointed to (using my hand has a guide). When I talk, my head move around and I can’t really fix that in place. I have tried but it then feels unnatural talking. I guess it just me then
The software settings are better, although I still wonder should this been looked at on a clip by clip basis or rather the entire track. Obviously, if I re-voice something later, I need to some match that clip into the reset of the edited track that I had before.
Of course, it was @DVDdoug that suggested the clip-by-clip editing previous, not you @steve, so sorry about that.