Reproduction of recorded sound

I wrote this essay / monologue for a classroom session with some teenaged would-be audio engineers. Mike Eschman.

Physics - The Reproduction of Sound.

The Social Impact of Reproduced Sound (including Video).

Reproduced music is commonplace.

It's live music that is unusual today.

Even in a club, you hear the band through microphones, amplifers and speakers. With Dolby 5.1, the acoustics of a room can be overcome, so that the soundstage has no apparent relationship to the room.

A peculiar situation, because our notion of a soundstage is based, on first person accounts of live music. But our experience, to some degree, is fundamentally third person. That is why reviews in a publication, say, Consumer Reports, is more likely to influence the purchase of audio and video equipment for the home than listening test. Most word-of-mouth comes from reviews. That is a social endorsement of a third person "preference in selection", as opposed to a first person bias. Somehow, to some unknown degree, we have lost absolute faith in first hand experience. Disturbingly, that is occurring in tandem with growing isolation, as discussed in the essay on "Human Diversity".

Compounding all other problems, the notion that costs and prices should never stabilize, have put dead musicians in direct competition with the living. That could drive live musicians into a system of patronage, where the few decide what the many will hear. In the past, listening to music played by someone other than you and your friends, played a relatively minor role in daily life for most Americans. Now, it occupies a greater portion of our day, and we are more isolated from each other.

Automobiles make suburbs possible by encouraging mobility, which weakens neighborhoods.

I can't put numbers to that, but, I would be embarrassed for anyone who argued that automobiles discouraged mobility and were good for inner city neighborhoods. Orson Welle's film adaption of Booth Tarkington's "The Magnificent Ambersons" is a fine exposition on this view of recent social history.

What are the changes to the listening experience with reproduced sound?

(1) Room reverberation detracts from a recorded performance, but enhances a live performance. We want to hear a recording that sounds like the Hall the band is playing in, we do not wish to hear the room we are sitting in.

(2) Each listener added to the room changes the performance in some fundamental way. In a live performance, if the audience and the band share some bond, it can add a great deal of excitement, possibly even insight, into the performance. For most people, most of the time, at least in my experience, that is not typically the result of a home listening experience with more than ten participants. For working adults, most listening to recorded music is done one-on-one.

(3) Each listener must maintain a geometric relationship to the speakers in the room, even in multi- channel systems, in order to hear EXACTLY what was recorded. This is about the same risk as buying a seat in a new concert hall, but the frequency and duration of events where you hear DISTORTION of one type or another is probably most of your adult musical experience, because opportunities to listen to recorded music permeate our daily lives.

Within these limitations, a stunning range of human accomplishment is available to a huge worldwide audience with a "reasonable" fidelity to the original. Most of that audience exists because reproduced music exists.

What is "reasonable" fidelity?

(1) "Good" distortion is slow and sneaks up on you, and bad distortion leaps out at you.

(2) Any distortion at a frequency that was silent in the original sounds horrible. Example: A 60 cycle hum, or any other constant buzzing in the sound. Example : When a radio station starts to "break up".

(3) Reproducing the sound quality of voices and instruments is very important. Example : The Trumpet and the Clarinet must sound different from each other. If I can't tell one from the other, the reproduced sound is BAD.

That quality is known as "Timbre". It is related to the spectrum of sound produced by an instrument or voice. It can always be heard, even in a recording that is BAD. For any instrument or voice, the timbre is different when a player plays LOUD than when singing SOFTLY. Furthermore, if a recording of a player playing SOFTLY is played LOUDLY, the listener can still tell that the player is playing SOFTLY. If you listen to a LOUD passage at a low level, you still know they are playing really LOUD - you can hear it - so it works in both directions.

You should make a grid with all the combinations and label each possibility true or false.

Using instruments which display a video, that shows clearly WHEN frequencies change in volume, it is possible to see how an instrument or voice sends sound into a room. These SHAPES demonstrate that a single musical note is composed of MANY frequencies, and that each instrument, including voice, has a distinctive shape on the video screen, because the note must start at 0 and change until it reaches the desired note.

Example : When a trumpet plays the musical note A = 440 hertz, a note that changes in intensity as a pressure wave in a room, exactly 440 times a minute, the trumpet must reproduce all the frequencies from 0 up to 440. As it moves towards it's stable frequency, some intermediate frequencies have no enharmonic relationship to that frequency, but others do.

What that means : The molecules in the materials from which an instrument is made, including voice, have geometric properties, one of these being length. When the length of the space between the pressure waves in a sound field, and the length of the molecules in the material have some specific numerical relationships to each other (don't worry about what those are, it's well documented) then the instrument adds a coloration to the sound, because while most of the frequencies being run die immediately, these resonate with the instrument and the musician's body cavities, if you're a wind player. So these make up part of the sound for an extended time period and cause multiple frequencies to shape the pressure waves making up the sound field.

Musical timbre depends on the relative loudness of these "partials", when the target frequency is reached.

Phase relationships have an important role in determining the sound of instruments, and the human voice.

What is a Phase relationship?

In this case, it is the time each partial took to reach it's stable volume level, on the way to the target frequency.

If all the partials rise to the same volume at the same rate very quickly, our instruments will display pressure waves that have straight sides and a flat top. This is a harsh, raspy sound.

It is most closely approached by plucked strings (guitar) and struck strings (piano).

But pianos and guitars can sound very smooth and expressive, not harsh or raspy?

Let's take a look at bowed strings - violins.

The motion of a bowed string follows a sawtooth pattern. The bow drags the string along for a short distance, then the string slips back, only to be picked up again and carried along with the motion of the bow. This is abrupt, and should sound BAD. But it can sound quite beautiful. How can this be? Bowed instruments incorporate a type of amplifer that smears the time domain in some beneficial way. This amplifier is called a SOUNDBOARD.

Partials that are close to the resonances in the soundboard are made louder than other frequencies, even frequencies right next to a partial frequency. A soundboard can also make a partial SOFTER, rather than louder.

Think about making an energy conservation argument that might explain this. For bonus points, specify an experiment to test your hypothesis.

Partials affect timbre in unexpected ways. Closed organ pipes have a "hollow" sound because only odd harmonic partials are present.

Vowel sounds can be distinguished and perceived regardless of pitch because they get their sound from the partials in our personal soundboards, the human vocal cords. Even in whispered speech, where ALL FREQUENCIES are present, vowels are clearly distinguishable.

If you sing the "ah" phrase - a vowel phrase - for as long as you can, two minutes? more? you will hear the sound begin to lose its "ah-ness".

So a vowel sound looses its vowelness over time, if sustained.

If that's so, then whispering "ah", it should "disappear" into the "hash" of ALL FREQUENCIES? If so, if not ...

You can play SOFT clarinet sound LOUD on a radio, and STILL TELL, the player played softly? This is due to loudness being a function of the distribution of volume among the partials.

Example : Loud trumpet has proportionally more energy in upper partials than soft trumpet.

This type of distribution, volume across the partials in a sound, is a type of "spectral energy distribution".

In all brass instruments, partials tend to rise late and fall early. More so for a trumpet than a tuba - i.e. more pronounced in higher register brasses.

The upper partials oscillate rapidly - this is called "brilliance".

Draw a "brilliant" spectral energy distribution that accounts for partial volume as a function of time. Hint : color coded partials vary volume over time.

In woodwinds partials rise and fall at the same time.

Flute and string players produce a small burst of noise just before a musical note sounds. This noise is important for the quality of sound the player rises too.

Bells and gongs produce partials which share no harmonic (i.e. the numerical relationships we talked about earlier) relationship. The harmonic series defines the relationships between fundamental frequencies and their related partials. This behavior can be experimentally deduced using vibrating strings.

Here's a real mystery : Bells and gongs don't exhibit consonance and dissonance when you play a chord sequence on them. But they are able to reproduce melody, as they reproduce pitch.

Write me with an experiment we can try in "the lab".

I'll videotape it and send it back to you.

mike eschman, etc ... http://www.etc-edu.com "Not just an afterthought ...