Psychoacoustics plays a very important role in our everyday lives. We are not necessarily affected by what we hear so much as how our minds interpret what we hear. For instance, right now you might think you are sitting in a perfectly silent environment. But listen closer: the whirr of your computer fan, the gentle hum of the air conditioner, your neighbors’ blarring all kinds of intolerable pop songs. We can notice all kinds of ambient noise when prompted, but often our minds just let them go unperceived. This is a good thing because it helps us not be disturbed by all the frivolous noise out there. Our minds filter out things for us so that we don’t get bothered by them unnecessarily
As professionals, amateurs, or hobbyists in the audio realm, we have to be more acquainted with psychoacoustic phenomena than the average Joe. I have been discussing the sub-bass portion of the audible spectrum, which is the most demanding register in terms of its share of the power spectrum, and it brings up an important psychoacoustic phenomenon called masking.
From Sweetwater Sound’s wonderful Word For the Day dictionary:
When sounds that contain similar frequencies are played simultaneously, the weaker sound tends to have those overlapping frequencies covered – ‘masked’ – by the frequencies from the stronger sound (especially in a dense mix). The frequencies of the weaker sound are still there; they are just not discernable over the more dominant sound with the same frequencies.
This is explicitly why it is important not to have too much information in the sub-bass region especially. The sub-bass is often an unusable portion of the audible spectrum, yet putting too much of it in a mix, perhaps such that some of it is audible, can cause it to mask neighboring frequencies in the bass register. This can lead to muddy, indistinct low end as the sub-bass masks frequencies in the bass section.
This becomes even more of an issue in digital audio due to encoding algorithms. The designers of audio codecs, notably MP3s, use masking as a way of excising “unnecessary” portions of audio. They have processes set up that detect masked frequencies and eliminate them from the mix. These algorithms are necessarily imperfect since no single metric could feasibly fit all recorded music.
If you look at a spectrometer for a full spectrum mix, you can see that the sub-bass portion generally reads extremely loud even though you can’t hear most of it. This means that, to an algorithm searching for masking phenomena, the sub-bass would read as the stronger sound, and despite being largely inaudible and unusable, the algorithm would preserve it at the expense of other more important, but less spectrally powerful portions of the audio spectrum.
As we’ll further explore next week, the entire bass region (including the sub-bass) is a relatively small region in terms of frequency bandwidth, so neighboring frequencies are very dependent on each other down there. Masking can occur at any portion of the audio spectrum, but it is especially important in the bass region.
