Compression is a word that’s tossed around but what the heck does it mean? The best example of compression is radio station broadcasting. Radio stations play music in their format (AC, CCM, R&B, Jazz, etc.). Within ANY format, you have songs that are soft and those that are loud. However, you don’t have to constantly adjust your volume knob because of compression. Compression reduces the overall dynamic range by boosting the soft sounds and lowering the loud ones.
Before I can venture into the wide world of compression, there are many other facets of sound that first have to be explained. These will explain why compression needs to be used.
The first definition you need to know is noise floor. The noise floor is the amount of ambient noise in the environment and the residual electronic noise present in the sound system. Ambient noise includes wind, traffic, people talking, people moving around, any natural noise in the environment. Rain is also ambient noise because rain makes sound.
The noise floor is measured in decibels. A concert venue would have a louder noise floor than say a church sanctuary just because of the nature of the environment.
The dynamic range is the next important term. The dynamic range is the difference, in decibels, between the loudest and the quietest portions of an event. Let’s say that range is 130 dB to 30 dB. Subtract the two numbers and you get a dynamic range of 100 dB.
Dynamic range can be affected by the noise floor. If the noise floor is louder than the quietest sound, then the dynamic range is the difference between the loudest sound and the noise floor. This might look like; 130 dB – 40 dB. That dynamic range is now only 90 dB. The noise floor is usually used as the value for determining dynamic range because it’s the easiest to measure.
If noise floor is the bottom, you also have to look at the upper limits of the dynamic range. I said 130 dB was the loudest volume. The nominal level is the average electronic line level. In easy to understand terms, the nominal level is the average sound level. The average sound level for example, might be 100 dB.
The difference between the nominal level and the loudest level is known as the amount of headroom. Again, we do the math with our current example and subtract 130 dB – 100 dB. That’s 30 dB’s of headroom.
Headroom, nominal level, noise floor, what does it all mean?
The amount of headroom available effects how much of a peak is allowable. Here we get into some practical application. A sound system capable of only 120 dB would create distorted sounds if the event’s dB level jumped above the 120 dB mark.
Let’s mark a few values of an example system.
- Noise Floor: 35 dB
- Quietest Part: 30 dB
- Nominal Level: 95 dB
- Loudest Part: 120 dB
- Headroom: (120 dB – 95 db) = 25 dB.
- Dynamic range: 35 dB to 120 dB
Note the dynamic range uses the noise floor because it’s louder than the quietest part. Also, the dynamic range includes the headroom.
Using this example, let’s place it in an environment with a sub-par sound system that can only handle 110 dB.
The problem in the above example is the noise floor is louder than the quiet portions and the sound system can only take 110 dB even though you’ll be pumping out an occasional 120 dB. Without compression, parts of the event will be inaudible and other parts will create distortion in the sound system.
Using compression, we are going to raise the soft parts and lower the loud parts. In audio terms, we are going to decrease the dynamic range. Compression works by applying a ratio of sound reduction to the input signal. For example, a 2:1 ratio would be for every 2 dB input, only 1 dB would be output.
Our range of sounds goes from 30 dB to 120 dB. We need to boost the low and cut the high. Therefore, we need to change from a 90 dB range to a (110-35) 75 dB range. 90 divided by 75 equals 1.2. Therefore, we want to apply a compression ratio of 1.2 to 1. This will compress the sound into the allowable range. However, when we use compression in this manner, we are compressing all of the audio signals.
It’s better to compress only what is necessary because compression can effect the quality of the sound.
We don’t need to compress everything. We can just work on the problem areas.
Our quietest level is 30 dB but with a 35 dB noise floor, we need to boost the volume by +5 dB. That changes the other levels. The nominal level becomes 100 dB and the loudest part is now 125 dB. We’ve just fixed the noise floor issue but what about the higher volumes?
We need to compress the high value down to 110 dB. Compressors allow for specification of a threshold, therefore, we mark the threshold at the 100 dB nominal level. From here, we apply the compression. Doing this, we only compress the levels that need compression. You might decide to set the threshold at 105 dB. This means the compression won’t kick in until 105 dB but would then require a higher compression ratio.
Depending on the equipment you have, you might be able to apply compression on a channel-by-channel basis. At the very least, you should have a limiter/compressor unit for the overall house levels.
Limiters (limiter/compressor units) are similar to compressors in that they compress the audio signal. The difference is that limiters can compressor at a much higher ratio such at 10:1. You can specify that above a certain threshold, a large compression ratio is implemented. This way, whether a sound peaks at, say, 140 dB or 190 dB, the output doesn’t knock people out of their seats.
With all that being said…
Using compression correctly means first knowing about noise floor, dynamic range, nominal level, and the limitations of your equipment. It’s also helpful to know the acceptable levels of sound for your environment. Having all this knowledge, you should be able to apply compression when and where it is needed.
If you are interested in the dark side of compression, check out this great post on the dark side of compression by JB.
There is more to compression regarding using it on a channel-by-channel basis such as if you want to restrain a sound to only occupy a certain range. But that’s another story…