Ok, I have gotten a number of questions from people asking to explain a few of the basic processes that we use. Normalization (or peak normalization), Compression, and Hard Limiting. So I put together a video that goes through these processes. Here are the basics before you jump into the video (or just jump in because I talk about the basics in the video):
So Peak Normalization (or Normalization for short) is simply like turning up the volume knob. All this process does is scan the audio, find the loudest peak, then it raises the entire selection of audio (or lowers if you set it that way) so that the loudest peak hits the specified level. Your audio will have one single loudest peak. Let’s say this peak is at -6dB. If you normalize the audio to -1dB, it will raise the level of the entire file by 5dB. This does NOT change the dynamics of the audio, it ONLY changes the amplitude (or gain or volume or level…depending on who you talk to). The dynamics are left intact.
Compression is a bit different. Compression works by altering the dynamic range of the audio. There are two settings that we will talk about: Threshold and Ratio. Threshold is the level at which the process will begin taking place. For standard compression, the Threshold is set and all audio ABOVE the threshold will be processed. This means if the threshold is set at -12dB, all audio ABOVE -12dB will be processed. Now the Ratio is basically the amount of compression (or squashing) that happens to the audio. The Ratio is read 3:1 or 4:1 or 10:1 or 100:1 (three to one, four to one…etc.). This is the input level to the output level. So a Ratio of 3:1 means for every 3dB of input, you get 1dB of output. So, for example, if you have a 6dB peak, the compressor will reduce that peak to 2dB. Now when you combine those two settings, you can control HOW MUCH the audio is compressed and WHEN the compression starts. With a threshold of -10dB, the process will compress the audio above -10dB only. So if the audio doesn’t get to -10dB it is not affected by the compressor. The audio ABOVE -10dB will be squashed so that every 3dB of input will only result in an final output of 1dB.
Let’s put that all together with the above specs. Ready for some math?? You have a bit of audio that has an average peak level of -4dB. You have your compressor set with a threshold of -10dB and a ratio of 3:1. So every peak above -10dB (-4dB IS above -10dB) gets squashed. If a peak (before compression) hits -4dB, then AFTER (the output) the compression it will be reduced so that it only hits -8dB. This is because the peak that hits -4dB is 6dB ABOVE -10dB. That 6dB is compressed at a ration of 3:1, so 6dB input would result in 2dB output. 2dB above -10dB is -8dB. Got it??
Now hard limiting is basically compression with a VERY high ratio…like infinity:1. This means that no matter what the input level is, it doesn’t raise the output level. Hard Limiters have quite a few different settings, but in Adobe Audition, you only need to set the Maximum Amplitude (ceiling) to achieve the effect you are looking for. If I set my Max Amplitude to -6dB, then all audio that originally went over -6dB will be squashed so that the peaks STOP at -6dB. So if I have an average peak level of -4dB, and I set my Hard Limiter with a Max Amplitude of -6dB, it will result in all peaks being squashed down to -6dB. This results in more of a chopping effect because as soon as the peak hits -6dB it is squashed. Too much hard limiting (or any process really) will result in audible distortion and abnormality. Hard limiting is to be used sparingly. I never use it for typical VO…just audiobook mastering.
Well there you have the text version. Normalization raises the level to a specified point and doesn’t affect dynamic, Compression changes dynamic, and Hard Limiting changes dynamic as well, but in a different way than compression.
That’s it for now! Check out the video for a demonstration of each process.