PERCEPTIVE LEVEL VS PHYSICAL LEVEL

 

 

THE ISSUE A quite important aspect of perception : the link between perceptive level and physical level, depending on frequency. Very useful when it comes to mixing and related activities.   *The unit of physical level is dB SPL (Sound Pressure Level). It measures the actual level of the sound wave, meaning the amplitude of the air vibration. It's merely a physical notion. *The unit of perceptive level is dB A. This one doesnt have anything to do with the sound wave. It corresponds to what a "normal" listener actually feels. The dB A scale is based on statistic studies. Several people got asked : "how loud is this ?" "how loud is this other sound" ? Based on multiple answers, the dB A scale was made.   The following diagram gives the relation between the two scales : dB A and dB SPL.     Let's look at the line labelled "20". It represents 20dBA, meaning 20 units of perceptive level. At 50Hz (very low frequency), 65dB SPL are needed to give an impression of loudness corresponding to 20dB A. At 1000Hz, to get the same perceptive level (20dB A), the sound pressure level only needs to be 18dB SPL. That means the ear is much more sensitive to 1000Hz sounds than to 50Hz sounds.

 

CONSEQUENCES ON MIXING This means a lot of things when mixing. A standard 16bit CD can handle physical levels from -96dB FS to 0dB FS. FS means "full scale". It's an electrical level, which will be translated linearly to sound pressure level coming from the speakers. It's also a physical level. Let's say I put in the mix a sound centered around 1000Hz. I want it to sound like 40 dB A at the output of the speakers, for a given volume setting of the amp. Then I will require this sound to have an electrical (physical) level of, let's say, -50dB FS (arbitrary value - it depends of the volume setting of the amp). Now, I want the same perceptive level coming out of the speakers, but for a sound centered around 50Hz (that's a deep bass). Then I'd need -10dB FS of electrical level to have that same 40 dB A at the output of the speakers. -10dB FS, that almost full scale ! It means that in a mix, very low frequencies will take practically all the available electrical level. On the other hand, 1000-4000Hz freq will take almost no level.

 

LISTENING LEVEL WHEN MIXING Interesting other consequence : when looking again at the diagram, one can see that the higher the perceptive level, the flatter the curves. It means that when you listen to a mix at low level, you will only hear the medium frequencies. (the low and high freqs will drastically fall down due to the bad sensitivity of the ear at those frequencies - see line labelled "0") When the overall level gets higher - for instance see line labelled "100" - the ear will get more sensitive to high and low frequencies. This means the mix will get much richer in basses and in highs. In practice, this is especially true for the basses.

 

WHAT WOULD THE WORLD SOUND LIKE WITHOUT THE EAR'S BUILT-IN CORRECTIONS ? Another interesting aspect of this diagram, though not very useful in audio production. Take the "20 dB A" line. Imagine applying a reverse EQ to an ambience - that's what we would constantly hear if our ear was not correcting the outside informations this way : - human voice as we hear it (mp3) - - human voice as we would hear it without the ear corrections (mp3) -   Back to main index