I spent a lot of time yesterday working on my audio compression - that is audio dynamics, the use of the word compression that a record producer would assume, not digital data compression. My Wire Compression algorithm is unique and uses a volume tracer to compress rather than doing it purely mathematically. I wanted to experiment with settings. Here is my raw signal:
And here with 3.0:1 compression using SoundForge, a well established standard. This roughly means that anything over a threshold is squished by 3 times the 'height'. The result is then volume boosted so it gives a bit of a false impression of getting louder. It is really getting 'flatter' and more even:
Wire Compression doesn't really do that. When above a threshold it attenuates the signal by the inverse of a volume trace, so it has the effect of levelling off, but in a different way. I have some parameters for the volume tracer and the gate power. Here are some. This is 200,200 Gate 200; 400,200 Gate 200; 800,200 Gate 200:
The 800 (quick 'attack') squashed things well. Here are 400,400 Gate 200; 600,600 Gate 200; 800,800 Gate 200:
These look similar because the faster tail off means it will hug the signal as closely when leaving as it when it hits a peak.
As a result, I decided to add another parameter and separate the gate on and gate off slopes, so I can have a gate which activates quickly but more gradually turns off. I'm unsure if this is useful, but the parameter is easy to add and doesn't take up and more space or time.
In sound, Wire Compression sounds like the original wave, like all compressors tend to, but at extremes there can be a sort of fuzzy and almost analogue distortion which is quite pleasant. I concluded though that for general use, I'll use SoundForge for studio basics like vocals, but use mine for other things.
I rarely use compression, mainly for vocals because it's important that they are heard at about the same level and not drowned out, and for bass sounds because these will vary in loudness a lot depending on pitch. Just playing a note or two lower or higher can radically change the volume of a bass (because it is heavily filtered), and the volume of the bass can really change the overall balance in a song, so compression is useful here.
This took all morning. I produced a song in the afternoon, a funny like country-style stomp called The Misery's Hard To Take. I can only see flaws in it, but it's not too bad. The programming on the new compressors is done, so I can get back to the hard work of music.
