Audio Engineering 201: "How to Sound"

SOUND 201 Large-scale Sound Reinforcement

Langton Labs August 18-19th, 2014

Michael Broxton Contact: [email protected]

mailto:[email protected]

WHAT IS IN THE CLASS?

• The physical behavior of sound & the sound field

• The human perception of sound

Today we will discuss the fundamental nature of sound.

Tomorrow we will discuss its reproduction using loudspeakers.

• Understanding loudspeaker specifications

• Moar volume: loudspeaker arrays

• Tuning the system: crossover, gain structure, delays, and EQ

• Tips for running live sound

THE SOUND WAVESound is our perception of a mechanical wave of pressure or

displacement traveling through a medium such as air.

Broadly speaking, a sound wave has three properties that carry information that determine its behavior.

THE SOUND WAVESound is our perception of a mechanical wave of pressure or

displacement traveling through a medium such as air.

Broadly speaking, a sound wave has three properties that carry information that determine its behavior.

THE SOUND WAVE

Image credit: Daniel Russel. Acoustics and Vibration Animations

Although sound waves coming from multiple sources sum together in a straight-forward manner…

http://www.acs.psu.edu/drussell/demos/superposition/superposition.html

THE SOUND WAVE




THE SOUND WAVE



…their summation can lead to complex patterns of constructive and destructive interference.


THE SOUND FIELDOf course, sound waves are not 1-dimensional.

They travel through space, and the evolve over time.

y

xClick on the image to try this demo online using Johannes Singler’s WaveWorkshop

http://www.singler.name/waveworkshop/#(ex:!((a:1,ph:0,s:(p:(x:0.35,y:0.67)),t:p)),l:0.1,t:Untitled,v:!((h:553,s:(ll:(x:0,y:-0.696319),ru:(x:3.07362,y:1)),st:1,t:r2,w:1002)))

http://www.singler.name/apps/waveworkshop/index.php?language=en

THE SOUND FIELDAn ideal point source radiates sound

equally in all directions as a spherical wavefront.

Click on the image to try this demo online using Johannes Singler’s WaveWorkshop

http://www.singler.name/waveworkshop/#(ex:!((a:1,ph:0,s:(p:(x:0.35,y:0.67)),t:p)),l:0.1,t:Untitled,v:!((h:553,s:(ll:(x:0,y:-0.696319),ru:(x:3.07362,y:1)),st:1,t:r2,w:1002)))


THE SOUND FIELDWhen wavefronts from two point sources (e.g. stereo

loudspeakers) interact, there is a pattern of interference.


http://www.singler.name/waveworkshop/#(ex:!((a:1,ph:0,s:(p:(x:0.4,y:-0.28)),t:p),(a:1,ph:0,s:(p:(x:0.39,y:0.51)),t:p)),l:0.1,ph:123.22,t:Untitled,v:!((h:545,s:(ll:(x:0,y:-0.696319),ru:(x:3.07362,y:1)),st:1,t:r2,w:994)))


THE SOUND FIELDNull zones of destructive interference span areas of

constructive interference where sound amplitude is doubled*.

Click on the image to try this demo online using Johannes Singler’s WaveWorkshop* - But perceived “loudness” is not doubled! We’ll cover that later…

http://www.singler.name/waveworkshop/#(ex:!((a:1,ph:0,s:(p:(x:0.4,y:-0.28)),t:p),(a:1,ph:0,s:(p:(x:0.39,y:0.51)),t:p)),l:0.1,ph:123.22,t:Untitled,v:!((h:545,s:(ll:(x:0,y:-0.696319),ru:(x:3.07362,y:1)),st:1,t:r2,w:994)))


THE SOUND FIELDThe spacing between the point sources effects the interference pattern.

As point sources get closer together, the null zones get farther apart, and vice versa.


When sources are really close together, they form a dipole radiator that acts like a directional point source.

Example: mid and high range driver in the same loudspeaker.

Example: two adjacent mid-range drivers in the same loudspeaker.

http://www.singler.name/waveworkshop/#(ex:!((a:1,ph:0,s:(p:(x:0.3,y:0.11)),t:p),(a:1,ph:0,s:(p:(x:0.31,y:-0.04)),t:p)),l:0.1,ph:1368.04,t:Untitled,v:!((h:545,s:(ll:(x:0,y:-0.696319),ru:(x:3.07362,y:1)),st:1,t:r2,w:994)))


http://www.singler.name/waveworkshop/#(ex:!((a:1,ph:0,s:(p:(x:0.3,y:0.11)),t:p),(a:1,ph:0,s:(p:(x:0.3,y:0.06)),t:p)),l:0.1,ph:1368.04,t:Untitled,v:!((h:541,s:(ll:(x:0,y:-0.696319),ru:(x:3.07362,y:1)),st:1,t:r2,w:990)))

SPEAKER CROSSOVERSThis pattern of interference must be considered

when designing loudspeakers with multiple drivers.

THE SOUND FIELDThe frequency of the sound also changes interference pattern and

spacing between the null zones.

Low frequencies push the null zones apart, but they also grow larger and more noticeable to the listener.


High frequencies can push the null zones so close together that they are so small as to not be noticeable at all.

Example: stereo subwoofers Example: stereo tweeters




THE SOUND FIELD

For mid/high frequencies we begin to perceive a stereo effect when the spacing between the sources is large enough, despite

the dense interference pattern.




THE SOUND FIELD

But sub-woofers are more tricky, though, because the pattern of interference is human-scale. We also do not tend to perceive

a stereo image at low frequencies.




SUB-WOOFER ARRAYSTo address this problem, it is best to place subs near each other so that

they appear as a single, coherent source.

A rear-delay or end-fire array, used in conjunction the the proper delays, can reduce the amount of sound energy emanating from the back and sides of the stage.

There are many clever ways to array sub-woofers to achieve advantageous cancellation effects. Often these are used to achieve a cardioid pattern of sound energy.

OTHER FACTORSThere are many other important factors that effect the sound field.

• Reflection… off walls, the ground. Plays a particularly important role indoors.

• Refraction… around objects, around people.

• Absorption… helps to attenuate reflections, or as the sound passes through humid air.

• Scattering… diffuses sound energy, spreading it out in a random manner.

These are topics that are important to consider in room acoustics. Another class!!

FOR MORE FUN EXAMPLESTry the Ripple Tank app (mac, iPad, or web)

LET’S TALK ABOUT REAL SOUNDS

They are the (linear) sum of many individual pure waveforms each with their own frequency and phase.

Add together

Final waveform

The real sounds we encounter in the world are complex.

DUAL DOMAINSThe Fourier transform is the mathematical tool we use to decompose a time signal into its frequency components, and vice versa.

It is quite useful, but no matter how complicated the time waveform, the Fourier transform only gives us information about the average power in each frequency over all time.

However we simultaneously perceive sound in terms of both time and frequency.

A MIXED REPRESENTATIONOur ear does a neat trick: it decomposes the sound into a mixed

time-frequency representation.

Time (linear)

Freq

uenc

y (lo

garit

hmic)

We can do this digitally (albeit imperfectly) using a tool called a spectrogram or a windowed or short-time Fourier transform (STFT).

!Using the spectrogram, we can see both spectral and temporal aspects

in the music in a way that is similar to how we hear it.

GET TO KNOW YOUR FREQUENCIES“Subs” “Mids" “Tops”

PERCEPTION OF FREQUENCY: PITCHThe human range of hearing is from about 20 Hz to 20KHz

Freq

uenc

y (lo

garit

hmic)

We perceive pitch logarithmically in relation to frequency. !

Each frequency doubling is perceived as an equal (perceptually linear) increase in pitch: i.e. one “octave.”

PERCEPTION OF SOUND INTENSITY: LOUDNESSNext let’s talk about another perceptual phenomenon: loudness.

Unit: dBu (voltage) or dBm (power)

Oscilloscope

Sound pressure

Mea

sure

men

t D

evice

Phen

omen

on

Sound Pressure Meter

Electricity Human Perception

Voltage, current, or power Acoustic Energy “Loudness”Unit: dB SPL

Your Ear

Unit: Phon

BRIEF DIGRESSION: THE DECIBELOur perception of loudness is also logarithmic.

What we perceive to be 2x as “loud” is actually 10x the acoustic energy intensity.

What we perceive to be 4x as “loud” is actually 100x the acoustic energy intensity.

What we perceive to be 8x as “loud” is actually 1000x the acoustic energy intensity.etc.

The decibel is a logarithmic measure relative to some reference level.

�� = ��

�� = �� µ�� (��)

Again, we use this because it is convenient, and matches our

perceptual experience.

For example: dB sound pressure level

or db SPL

BRIEF DIGRESSION: THE DECIBEL

BRIEF DIGRESSION: THE DECIBEL

Depending on where you are in the signal chain, you may find yourself using a different one of these scales.

But they are all compatible, inasmuch as a +10dB in one scale leads to a +10dB increase in the others! It is designed to be simple and intuitive.

Electrical Energy

�� = ��

�� = �� µ�� (��)

�� = ��

�� = ��

��

Unit: dBu

�� = ��

�� = ��

Voltage e.g. Mixers

Unit: dBm

Power e.g. Amplifiers

Acoustic Energy

Unit: db SPLPressure

Power Unit: db SWL

Loudness

A phon is equal to the sound pressure level (in db SPL) of an equivalently “loud” 1-KHz tone.

�� = ��-��

�� = ��

��

Beware… there are many decibel scales & reference levels!

CONTROLLING VOLUME OR “GAIN”


Oscilloscope

Sound pressure

Mea

sure

men

t D

evice

Phen

omen

on




Your Ear

Unit: Phon

LEARNING TO THINK IN TERMS OF DECIBELS

+3dB is: 2x the acoustic power

but only 1.23x as “loud”

+6dB is: 4x the acoustic power but only 1.5x as “loud”

+10dB is: 10x the acoustic power

and 2x as “loud”

Some implications to think about:

• A 1-2 dB change in volume is barely perceptible

• Doubling amplifier power does not double loudness

SOUND INTENSITY AND ACOUSTIC ATTENUATION


Oscilloscope

Sound pressure

Mea

sure

men

t D

evice

Phen

omen

on




Your Ear

Unit: Phon

SOUND ATTENUATION OVER DISTANCE

SOUND ATTENUATION OVER DISTANCE

PERCEPTION OF SOUND INTENSITY: “LOUDNESS”


Oscilloscope

Sound pressure

Mea

sure

men

t D

evice

Phen

omen

on




Your Ear

Unit: Phon

EQUAL-LOUDNESS CONTOURS

The equal loudness curves are psychoacoustic measurements of how humans perceive loudness.

THE RANGE OF HUMAN HEARING

They can help us to draw a boundary around the perceptible range of sounds.

THE RANGE OF HUMAN HEARING

Or the sounds that represent speech, or music.

As the sound engineer, it is your job to protect your audience from hearing damage.

• Use limiters to prevent transients from damaging hearing.

MEASURING NOISE EXPOSURE

Measures you should take:

• Measure the average noise “exposure” over the scale of minutes or hours.

Rough guidelines for dance music:• 90-100 dB SPL (average) is a good, relatively safe

volume early or late in the night.

• 100-110 dB SPL (average) is a good sustained level at the peak.

• Beyond this, you risk damaging ears and speakers.

PROTECT YOUR HEARING!

More info: https://www.etymotic.com/pdf/er_noise_exposure_whitepaper.pdf

Occupational Safety and Health Administration (OSHA) &

National Institute for Occupational Safety and Health (NIOSH)

https://www.etymotic.com/pdf/er_noise_exposure_whitepaper.pdf

PROTECT YOUR HEARING!Passive Protection Active

$300$10

$150

PART II

• The physical behavior of sound & the sound field

• The human perception of sound

• Understanding loudspeaker specifications

Yesterday we discussed the fundamental nature of sound.

Today we will discuss its reproduction using loudspeakers.

• Moar volume: loudspeaker arrays

• Tuning the system: crossover, gain structure, delays, and EQ

• Tips for running live sound

THE IDEALS VS. REALITY

The ideal loudspeaker would: !

… radiate sound like an ideal point or line source … !

… play music with a flat frequency response over all audible frequencies … !

… and get arbitrarily loud.

This is not physically possible!

THE IDEALS VS. REALITYFor starters, different frequencies of sound have different properties.

Low frequencies:

• Diffract more, reflect less

• Require a driver that can move a lot of air!

• Are highly omnidirectional

High frequencies:

• Diffract less, reflect more

• Requires a driver that can move very fast!

• Are highly directional

LOUDSPEAKER DIRECTIVITY

(at least down to the low frequencies)

A good loudspeaker has been optimized to produce roughly

equal acoustic power over a limited arc of angles.

Managing the pattern of sound dispersion is called pattern control and it is the key to understanding how multiple

loudspeakers interact.

LOUDSPEAKER DIRECTIVITYThe first consequence of this is that you should put people’s ears where

the speaker is producing the best possible sound.

LOUDSPEAKER DIRECTIVITY

As an aside: this has implications for where you place stereo loudspeakers, and the directions you point them.

ANATOMY OF A LOUDSPEAKERBass-reflex vent

Bass-reflex vents

Direct Radiating Woofer

Horn-loaded mid

Horn-loaded HF compression driver

THE IMPORTANCE OF THE ENCLOSURE

DIRECT RADIATING LOUDSPEAKERS

Direct Radiator Reflex Enclosure

HORN LOUDSPEAKERS

FOLDED HORN

OUR NEW SOUND SYSTEMLA400

LA460

LA215LF Subsystem: 1x 15-in, vented

HF Subsystem: 1x 2-in exit/3-in voice coil compression driver on Wave Guid Plate.

LF Subsystem: 1x 15-in, vented

HF Subsystem: 1x 1.4-in exit/1.75-in voice coil compression driver on constant directivity horn.

MF Subsystem: 1x 8-in cone, horn-loaded

LF Subsystem: 12-in woofer, bent bass horn



LA128

LA128z

Coverage angle (+/- 6dB): 180°



Coverage angle (+/- 6dB): 90° Conical

Coverage angle: 60° x 45°

OUR NEW SOUND SYSTEMLA400

Power handling: 500W @ 8 Ω Freq response: 45-250 Hz

Sensitivity: 107 dB SPL/W @ 1m

LA215

Power handling: 600W @ 8 Ω Freq response: 69 Hz - 18 KHz


LA460Power handling:

full range: 500W @ 8 Ω bi-amp (LF/MF): 500W @ 8 Ω

bi-amp (HF): 150W @ 8 Ω

Freq response: 62 Hz - 20 KHz

Sensitivity: full range: 97 dB SPL

bi-amp (LF/MF): 97 dB SPL bi-amp (HF): 108 dB SPL

LA128

Power handling: 1600W @ 4ΩFreq response: 31-200 Hz


LA128z

Power handling: 2000W @ 4ΩFreq response: 31-200 Hz


LOUDSPEAKER ARRAYS

Line Source ArrayPoint Source Array

In order to create high sound pressure levels over a large areas, you need to array many loudspeakers together.

LOUDSPEAKER ARRAYS

Image credits: McCarthey, Bob. Meyer Design Reference for Sound Reproduction. 1998

LOUDSPEAKER ARRAYS

Image credits: McCarthey, Bob. Meyer Design Reference for Sound Reproduction. 1998

ANOTHER FUN LOUDSPEAKER SYSTEM

20 x 12" Mid-bass drivers in 10 cabinets (2500 Watts / cabinet)

18 x 21" drivers w/ 6” voice coil 15mm excursion, & neodymium magnets (4000W / 2 drivers)

12” and 2” horn-loaded drivers, “cat’s eyes” horn

flare

SYSTEM PROCESSOR

SYSTEM PROCESSOR

1. Configure crossover frequencies

2. Add driver alignment delays, polarity, and EQ

3. Calibrate the gain structure & set the limiters

4. Rough balancing of frequency response

5. Careful system EQ

CONFIGURE CROSSOVER FREQUENCIES

SYSTEM PROCESSOR






DRIVER ALIGNMENT DELAYS, POLARITY AND EQ

Bi-amplified loudspeaker Speaker system with flown tops

Sub

Top

2m

5m

5.4m

Delay the subs by: 0.4m (1.2ms)

SYSTEM PROCESSOR






CALIBRATE GAIN STRUCTURE & SET LIMITERSAs a general rule, use an amplifier delivering 1.5x - 2x the speaker's average ("RMS") power rating.

Note: amplifiers are a fixed-gain device. the knob of the front of the amplifier attenuates the input, rather than “turning up” the output.

http://www.doctorproaudio.com/doctor/temas/powerhandling.htm

CALIBRATE GAIN STRUCTURE & SET LIMITERS

Setting gain structure involves two steps:1. adjust levels so that all parts of the signal chain clip at the same time.2. use limiters to prevent the amplifiers from clipping

SYSTEM PROCESSOR






SYSTEM PROCESSOR






POWERING THE SOUND SYSTEMThere are a few things to consider here:

• Power to the mixer, stage monitors, laptops, etc. should be a single independent power circuit.

• Each amplifier will take somewhere between 5-15A on a normal 120V AC line. AC circuits are typically 20-30A. Plan accordingly!

• Make sure your extension cables are rated for the power you are delivering.

• Avoid power strips. Plug amps directly into the line, using splitters if necessary.

TROUBLESHOOTING

Blown fuse / tripped breaker

Unexpected resonance in a speaker

Clipping

A dead speaker or amp

If this happens: Do this:

Generator runs out of gas Hang your head in shame. :)

Reset breakers. Replace fuse.

Adjust crossover LPF

Turn down the mixer

Depends on the situation…

Late DJ Have laptop or DJ iPod at the ready

Mic feedback during live act Reduce gain, twiddle EQ

THANKS!Michael Broxton

Contact: [email protected]

mailto:[email protected]