Dynamic Range -- IMP Week 4

Hi folks -
Welcome to my latest assignment from week 4 of "Introduction to Music Production", an online course from Berklee College of Music via Coursera. We are really getting into the nitty-gritty!

My chosen topic this time is:

Explain Dynamic Range and the many ways producers manipulate dynamic range.

=== Lesson ===

Dynamic range, in acoustics, is the ratio between the volume -- the sound pressure level -- at which no sound can be heard, and the volume at which ear damage or pain occurs.
[Dynamic range can also refer to the theoretical limitations of a piece of equipment, but the equipment here is the human ear, so hearing threshold and pain threshold are appropriate parameters.]

To start talking about loudness of a sound, we have to start with the sound itself. 

Technical stuff

Sound in air consists of a sequence of compressions and matching rarifications. A sound with one thousand of these sequences per second has a frequency of 1000 CPS (cycles per second), also called 1000 Hz (hertz, named for physicist Heinrich Hertz). The frequency of the sound is what we hear as pitch, and contributes to timbre. The volume of the sound, however, is determined by the average* sound pressure level (SPL) of the wave compressions.  That is what we perceive as loudness. 

Note that loudness is how we perceive sound pressure. They are not quite the same thing. For one thing, the ear is more sensitive in some frequencies than in others. We hear best in the 1000-4000 Hz range, which is where the vocal overtones that let us distinguish speech sounds fall. Strong overtones, such as those generated by distortion, do not necessarily add physical volume (same SPL), but they increase the perceived volume, or loudness. 

In physics, we measure pressure in Pascals. The lowest sound pressure level (SPL) the human ear can hear is about 20 microPascals (μPa). In acoustics, we call this zero decibels, and use that as the baseline to define the human hearing dynamic range. 

The upper end of the human hearing dynamic range is the sound pressure which is actually painful, and can cause permanant damage to the ear with only a short exposure.  This is usually considered somewhere around 120-140 decibels. 

Decibels are logarithmic. A dynamic range of 0 to 100 dB would be from 20 to 100,000 microPascals. 

So how does the concept of dynamic range apply to sound recording? 

First, in any situation, there will be a certain amount of ambient noise. In even the quietest concert hall, there will be the sound of people breathing, moving, shifting their weight; in a cafe setting, there will be (hopefully quiet) conversations and noises from the food and drink. In a studio, there will be electronics noises. The "noise floor" essentially raises the bottom end of the dynamic range. A "noise ceiling" is also provided by the point at which the signal approaches distortion levels, or even just uncomfortable levels. 

Basically, to change the dynamic range, we can amplify or lower the loud parts, thus changing the "ceiling" of the dynamic range, or we can amplify or lower the soft parts, thus changing the "floor" of the dynamic range. 

Sometimes you need to change the level of a track, for instance when a vocalist varies the mic/mouth distance, or when the audience joins in on a chorus. But when would we want to change the overall dynamic range?

One example for which a reduced dynamic range might be desirable is narrating audiobooks. While a certain amount of dynamic variation adds interest to the storytelling, the volume must not vary so much as to make the words in a quiet passage hard to understand, or a shouted part unpleasant. 

"Mood music" is another genre which calls for a reduced dynamic range, as it is often geared towards a hypnotic smoothness. 

On the other hand, a movie may have a soundtrack designed to have a huge dynamic range; a film may have a whispered conversation in one part, a subliminal tone in another, and then try to shock you out of your seat with huge explosions. 

There is an unfortunate trend towards reduced dynamic range in the guise of increased loudness. If everything is loud, then the dynamic range is restricted to a boring monotone, just a monotone with unpleasant clipping. Here are some very good articles:

What Happened to Dynamic Range by Bob Speer

In Celebration of Dynamic Range by Matthew MacGlynn

*Footnote: To measure loudness, we want an average of the sound pressures of the cycles. However, the way math works, since compressions (positive number) are paired with rarifications (negative number), the pairs average to zero. To get around this, we have to first square the figures. This makes all the values positive numbers. Then we average the values, and take the square root. This gives us what is called the "root mean square" or RMS value.

=== Reflection ===

I hope I didn't get too technical on the "explain dynamic range" part. I found it fascinating, and had gone off on a tangent on exactly what a decibel is, before I decided that was really not appropriate and deleted it. But I do have definite geeky tendencies. 

Those articles on dynamic range were especially interesting. The first is a lament; the second is really more informative. But they both talk about a problem that I hadn't really been aware existed. 

Again, I wish I had more time to devote to this assignment. I need to spend a LOT more time playing with my DAW. 

Thank you for wading through this.