Transcript the mid-term review powerpoint

Mid-term review
Physics of sound, part 1
What to focus on:
When a sound is created in the environment:
1. Something vibrates (our vocal cords, a machine, a musical instrument, as a result of friction when
we move, etc.)
2. The air (or sometimes another fluid medium like water) is disturbed by the vibration, resulting in
pressure waves
3. The pressure waves ‘wash over’ our ears (or the hearing apparatus of other creatures) and are interpreted as
sounds by our or the creature’s brain circuitry (more on this later)
Representation of sound as an oscillogram:
When the oscilogram line passes above the center line, it represents an increase in atmospheric pressure
(molecules compressed by the surge of the pressure wave)
Center line represents zero atmospheric pressure (no disturbance in the environment)
When the oscilogram line passes below the center line, it represents a decrease in atmospheric pressure
(molecules decompressing after the surge of the pressure wave)
Mid-term review
Physics of sound, part 1
The vertical axis of an oscillogram represents the amplitude or force of the pressure wave
Lower amplitude (softer) parts of the sound
Higher amplitude (louder) parts of the sound
(Beginning / earlier part of sound)
(End / later part of sound)
The horizontal axis of an oscillogram represents the timeline over which the pressure wave plays out
Mid-term review
Physics of hearing
Stage one:
-vibrating air
enters the ear
canal
-pressure waves
cause the
eardrum to
vibrate
sympathetically
2 cm
Sympathetic vibrations to
incoming pressure waves
Mid-term review
Physics of hearing
dB SPL
(different reading than
on mixers and in sound
editors, etc.)
Threshold of pain
Guard against hearing
loss!
SPL levels at many
concerts, in clubs and
when wearing
headphones can be
deceptively high and
cause temporary or
permanent damage to
parts of the outer,
middle or inner ear.
Mid-term review
Physics of hearing
Stage two:
-eardrum motion
causes the attached
Malleus (hammer) bone
to move, which acts as a
lever for the Incus (anvil)
and the Stapes (stirrup).
Stage three:
Mid-term review
Stage three:
-the stapes bone
vibrates onto the
oval window at
the entrance of
the cochlea
-the vibration
(amplified
because of the
leverage between
the 3 bones) causes
a liquid pressure
wave* to move thru
the cochlea
Physics of hearing
Mid-term review
Stage three:
-the stapes bone
vibrates onto the
oval window at
the entrance of
the cochlea
Physics of hearing
-the vibration
(amplified
because of the
leverage between
the 3 bones) causes
a liquid pressure
wave* to move thru
the cochlea
(Stage four)
Mid-term review
Stage four:
-in the basilar membrane of the
cochlea, tiny hair cells react to
the liquid pressure wave
-humans have around
20,000 of these cells,
each of which is
a slightly
different length,
and each of
which
resonates with
a different
frequency in our
hearing
range.
Physics of hearing
-each of the cells
connect with nerve
endings that then
transmit precise
frequency, spectrum
and amplitude
information
electrically to the
brain.
-spatial information
comes from the
different readings
each ear gets from the
environment.
Full auditory
perception, including
spectrum, then occurs
Mid-term review
Physics of hearing
It is important to
note that we do not
have ‘flat response’
hearing.
The red lines in the
chart indicate the
level at which a
frequency must be
played in order to
sound to us as if it
is at the level in the
center column
(threshold-20-4060-80-100 phons or
psychoacoustic
equivalent levels)
In the high range (5.5- 20 kHz)
increasing amounts of boost
must made to be heard at an
equal level (ex: ~ 30 dB SPL =
20 phons at 17 kHz)
In the low range (20 - 100 Hz)
large amounts of boost must
made to be heard at an equal
level (ex: ~ 60 dB SPL = 20
phons at 40 Hz)
In the mid to upper-mid range (400 Hz-5 kHz), frequencies are either left at the same level
(ex: 20 dB SPL = 20 phons at 500 Hz) or reduced (ex: ~12 dB SPL = 20 phons at 3-4 kHz)
[This is why I so often give notes about having to reduce mid-range frequencies in mixes]
Mid-term review
Physics of sound, part 2
Representation of sound as a spectrogram / sonogram: sound analyzed as our ears do in Stage 4:
Like oscillograms, the sonogram L-R axis shows
time placement of spectral elements:
Sonogram height shows frequency of spectral elements:
20kHz
10kHz
5kHz
1kHz
500Hz
200Hz
20Hz
Sonogram A
In this sound, we see strong energy in the 100-300 Hz range,which is
where the fundamentals of voices and many musical instruments sound
Sonogram colors show intensity of spectral elements:
Softest
Just above,in the 300-750 Hz range, are medium intensity spectral
frequencies that might be higher fundamentals or strong harmonics
The weakest energy is in the 1-20 kHz range - very little there
Loudest
Mid-term review
Physics of sound, part 2
Like oscillograms, the sonogram L-R axis shows
time placement of spectral elements:
Sonogram height shows frequency of spectral elements:
20kHz
10kHz
5kHz
1kHz
500Hz
200Hz
20Hz
Sonogram B
In this sound, the energy below 250 Hz range,is fairly weak
The strongest energy is between 250 Hz - 4kHz, and it is more ‘grainy’
and less sustained than in Sonogram A
In the top (4-20 kHz) range is medium intensity energy, which means the
sound has a lot of very high frequencies
Sonogram colors show intensity of spectral elements:
Softest
Loudest
Mid-term review
Physics of sound, part 2
Like oscillograms, the sonogram L-R axis shows
time placement of spectral elements:
Sonogram height shows frequency of spectral elements:
20kHz
10kHz
5kHz
1kHz
500Hz
200Hz
20Hz
Sonogram C
In this sound, the energy below 450 Hz range,is strong and sustained
The next strongest area is between 500 Hz - 5kHz, possibly higher
components of a complex mix below
In the top (5-20 kHz) range is very weak energy, which indicates a faint
high-frequency ‘sheen’ to the sound.
Sonogram colors show intensity of spectral elements:
Softest
Loudest
Mid-term review
Physics of sound, part 2
Like oscillograms, the sonogram L-R axis shows
time placement of spectral elements:
Sonogram height shows frequency of spectral elements:
20kHz
10kHz
5kHz
1kHz
500Hz
200Hz
20Hz
Sonogram D
Sonogram colors show intensity of spectral elements:
In this sound, the energy is completely evenly distributed all across the
spectrum - something that rarely happens in natural sounds
Softest
Loudest
Mid-term review
Physics of sound, part 2
Pitched and unpitched sounds:
Certain sounds, like the notes produced by musical instruments, have what is referred to as pitch
Pitches are more neatly organized compared to other sounds, consisting of stable frequencies that reinforce one
another because of being mathematically related in a simple way.
Again, a zoomed-out oscillogram does not tell us whether a sound is pitched or unpitched:
Sound A
All we know is that the first sound has sharp attack, and
then a consistent decay…
Sound B
…and that the second sound has several peaks before
decaying
Mid-term review
Physics of sound, part 1
But a look at the sonograms for the two sounds immediately reveals which is a stable, organized pitch, and which
isn’t:
Sonogram A:
Harmonic
multiples of
the fundamental
Fundamental
frequency
Sonogram B:
Thick clusters
of partials not
reinforcing any
particular frequency
Mid-term review
Physics of sound, part 2
Pitched sounds have
ordered harmonics, only
cover specific parts of spectrum
5x, 6x, 7x, 8x, 9x, 10x, 11x, etc.
Unpitched sounds (e.g. ‘hits’, hiss in this file)
cover full stretches of spectrum
4th harm. 4 x 450 = 1.8 kHz
3rd harm. 3 x 450 = 1350 Hz
2nd harm. 2 x 450 = 900 Hz
1st harmonic ~450 Hz
(freq2)
Mid-term review
Physics of sound, part 2
Parts of a zoomed-in oscillogram (not the attack) can give us clues about whether the sound is pitched or nonpitched:
(A regularly repeating wave cycle such as this one will generate a stable pitch)
(This irregular waveform, which contains little of no repetition, will not generate a stable pitch)
Mid-term review
Physics of sound, part 2
Pitched sounds can also be stable (e.g. the examples looked at so far) or unstable (the violin note-gliss-vibrato note below):
Mid-term review
Physics of sound, part 2
Human speech combines pitched and unpitched sounds:
The hard consonants
(unvoiced) have
complex spectra and
aren’t organized like
pitches
The vowels and soft
consonants (voiced) are
much louder & have
pitch-like organizations,
often gliding
Sonogram of the same excerpt
Th
f
e
“The fourth anniversary of the nuclear misunderstanding…”
th
our
Oscillogram
of a speech
excerpt -->
v
an-
ni-
s
er-
th
a
ry
of
c
e
nu
s
le ar mi
s
under
Mid-term review
Physics of sound, part 2
Slower / lower frequency voiced sounds - less compressed in oscillogram
Higher / faster frequency unvoiced sounds - more compressed in oscillogram
…which is confirmed by the sonogram of the same excerpt:
Mid-term review
Physics of sound, part 2
Spatial behavior / perception of sound
Especially in indoor environments, the sound we hear is not only the direct vibrations made by the object, person or
process, but a number of echoes and reverberations laid on top of the direct sound, the nature of which is generally
referred to as (room) acoustics.
The time it takes for reverberated sound to die down to
below the threshold of hearing is called the reverb decay,
and is given in seconds or milliseconds
[SWP; OPCH]
Concert halls and other listening environments are / were designed to create a pleasing (but not too confusing)
reverberation, in the 1.5 - 2.0 second range:
Estherhazy Palace Hall, Austria, where most of the classical
composer Haydn’s works were performed in his lifetime.
Reverb time: 1.2 seconds
Grosser Musikvereinsaal, Vienna,
a major European concert hall.
Reverb time: 2.0 seconds
Powell Hall, St. Louis
Reverb time: 2.2 seconds
Mid-term review
Physics of sound, part 2
Recording studios, on the other hand, are designed with shorter reverb times, because this allows for more control
of recorded signals.
But contrary to what many people think, a completely ‘dead’ studio acoustic (i.e., reverb time < 0.3 sec) was only
ever considered the ideal in the late 70s and 80s.
Mid-term review
Physics of sound, part 2
Recording studios, on the other hand, are designed with shorter reverb times, because this allows for more control
of recorded signals.
But contrary to what many people think, a completely ‘dead’ studio acoustic (i.e., reverb time < 0.3 sec) was only
ever considered the ideal in the late 70s and 80s.
Today, most studio control rooms have reverb times between 0.5 - 0.8 sec, while recording rooms vary from
0.3 sec (e.g., booths for recording voiceovers) to rooms that can be ‘opened up’ to over 1.0 secs when carpeting
and other acoustic baffling is removed.
Mid-term review
Physics of sound, part 2
Arenas and other large concert venues not designed for sound present the challenge of very long reverb times
(>3-4 seconds) that can confuse listeners with too much overlap.
This is usually dealt with by providing as much direct sound to as many parts of the venue as possible.
Speaker clusters for a
stadium rock concert
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
Mid-term review
MICROPHONES
Recording the Voice
Mid-term review
MICROPHONES
Vowels and Consonants
Mid-term review
MICROPHONES
Frequencies
16 - 400 Hz
% Contibution to speech
intelligibility
7.2
400 - 800 Hz
14.4
800 - 1600 Hz
22.2
1600 - 3200 Hz
32.8
3200 - 20,000 Hz
23.4
Vowels
Consonants
Mid-term review
DIGITAL AUDIO RECORDING
INPUT ANALOG
SIGNAL
ANTI-DISTORTION
FILTERING
SAMPLE AND
HOLD
(The voltage stream from
a microphone, the line
level stream from
playback devices or
keyboards passing
through a mixer, etc.)
(Required to eliminate
aliasing - more on this in
a moment)
(Linked to sampling rate
and resolution)
(The stream of numbers
created according to the
sampling rate & resolution)
[Anti-aliasing low-pass filters used relative to the 1/2 sampling rate,
to eliminate problematic high frequencies]
Cutoff
Frequency
(Amplitude)
(All frequencies above 1/2 sampling
rate must be eliminated to avoid
distortion.
E.g. 25k Hz frequency distorts at 18
kHz
at 44.1 kHz sampling rate:)
(-3 k)
ANALOG TO DIGITAL
CONVERSION
(Frequency)
(On hard drive, in live
RAM memory or
using other storage
media)
(The crucual process of remaking the waveform from the
stream of numbered samples, the
quality and accuracy of which is
linked to the sampling rate,
resolution and other factors)
DATA
STORAGE
DIGITAL TO ANALOG
CONVERSION
19
kHz
(The resulting analog
electrical waveform)
(+3 k)
25
kHz
(The sound we
hear through
speakers or
headphones)
Mid-term review
DIGITAL AUDIO RECORDING
…can be thought of as a series of ‘snapshots’ that the digital recorder takes of
the sound wave…
Sampling rate = # of digital ‘snapshots’
taken of the sound every second):
Sample…
[ProTools LE sessions (i.e. recording
& editing) .can only be in 44.1 kHz or
48 kHz, however…
Common sampling rates:
44.1 kHz (CD audio)
(sound wave)
48 kHz (Video & DAT)
[…when bouncing
(rendering) audio,
ProTools LE offers a
wide variety of
sampling rates,
from ‘lo-fi’ 8kHz
to ultra-high-end
192 kHz]
96 kHz (high-end digital recording)
… & Hold
(Stream of numbers according to sampling rate & digital resolution)
Mid-term review
DIGITAL AUDIO RECORDING
Sampling resolution = number range available for representing amplitude
Because digital memory is built on switch-like binary bits, the resolution of
each sample taken is determined by the bit rate of the digital recording:
0 1
…1 bit provides a resolution of 2 binary numbers (0 or 1)
0 1 0 1 0 1 0 1
…4 bits provide a resolution of 16 binary numbers (0-0-0-0; 0-0-0-1; 0-0-1-1; etc.)
0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
…8 bits (still heard in some digital samples on the Internet and in older video
games) provide a resolution of 256 binary numbers…
0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
…16 bits (standard for CDs and most digital sound) provide a resolution of 65,536 binary numbers…
0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
…24 bits (now the industry standard for high fidelity sound recording & reproduction)
provide a resolution of 16,777,216 binary numbers.
[ProTools LE sessions.and rendering
/bouncing can be in 8, 16 or 24-bit
resolution]
In-line effects
Mid-term review
Signal processing
Definition:
An effect through which the signal on a track or mixer channel passes directly en route to the output
The ‘raw’ signal is transformed by the effect, and cannot be heard as is unless the in-line effect is bypassed
Signal
In-line effect
(bypass -->)
Output
In-line effects
Mid-term review
Signal processing
In some situations (recording studios, film shoots) external
effects units are still used. These are of two basic types:
Rack mounted effects devices,
patched into a patch bay and routed
through the mixing board
Pedal or ‘stomp box’ external effects are patched
directly from a guitar, keyboard or other electric
instrument to a direct box, and then into the mixing
console via patched connections. Pedals often require
external noise reduction devices to make the
processed signal they output usable in a mix.
But most effects used today are sub-routines
within digital audio editors called plug-ins
Mid-term review
In-line effects
Signal processing
Equalizers (EQ)
Manipulation of the spectral color of a sound through phase cancelation (cut) or reinforcement (boost)
within certain frequency bands
EQ types:
Rolloff / low-pass, high-pass filters
Peak / dip (notch filter)
Shelving
Key points:
Pro engineers and mixers always advise using EQ to CUT rather than BOOST frequencies.
Eqs are not designed to act as effective preamps, which is what they do when they boost certain frequency
bands. Usually you boost conservatively -- a few dB or so. But extreme EQ can sometimes work -experiment!
Because of the way human hearing works, mid-range frequencies (which our ears are hypersensitive to)
almost always have to be cut to avoid a ‘crowded’ feeling in mixes.
Some examples of EQ use in music:
To establish clear ‘frequency areas’ in the mix that don’t conflict and don’t mask other sounds
To bring out the most characteristic part of a sound so that its presence in the mix is there without being
overbearing, e.g.
Electric piano <--> Electric guitar;
Bass drum <--> Lower register bass
Low toms <--> Upper register bass;
Voice(s) <--> Synth, saxophone
In-line effects
Mid-term review
Signal processing
Variable Gain Amplifiers
1. Compressors
Definition: An automatic volume (a.k.a. ‘gain’, ‘level’, ‘amplitude’) control.
Compressor/Limiters are the only signal processors generally used in recording, as a safeguard against distortion
Works by reducing gain by a certain ratio when the signal goes above a set threshold
Examples:
A compressor with a ratio of 2:1 and a threshold of -10 dB will reduce a -2 dB signal at the input to
-6 dB at the output (1/2 x 8 dB over threshold = 4 dB gain)
-2dB
-2dB
-6dB
-6dB
-10dB
-10dB
-14dB
-14dB
No compression:
-2 dB = -2 dB
With compression (2:1 ratio & -10dB threshold):
the -8 dB of gain is halved by the compressor (reduced by -4 dB)
Compressed output is -6 dB
In-line effects
Mid-term review
Signal processing
Variable Gain Amplifiers
2. Limiters
Definition: heavy compression (10:1 ratio or greater; ProTools’ default is 100:1).
Severely cuts gain above the set threshold..
Much more obvious to the ear than compression.
Often used as a ‘safety’ to prevent clipping above a certain limit, for broadcast signals, or for dynamically
limited media like cassettes.
Also used to reduce pops & clicks,
Example:A compressor with a ratio of 100:1 and a threshold of -10 dB will reduce a -2 dB signal at the
input to -10.08 dB at the output (1/100 x 8 dB over threshold = 0.08 gain )
-2dB
-2dB
-6dB
-6dB
-10dB
-10dB
-14dB
-14dB
No limiting:
-2 dB = -2 dB
With severe limiting (100:1 ratio & -10dB threshold):
the -8 dB of gain is reduced to 1/100th of its size by the limiter
Compressed output is -10.08 dB
In-line effects
Mid-term review
Signal processing
Variable Gain Amplifiers
3. Gates
Definition: does the opposite of a limiter. When the signal falls below a set threshold, the level is abruptly
cut to a specified level
.
Used to eliminate unwanted low level sounds in speech, music and other recordings.(e.g. to eliminate
headphone or other leakage on music tracks).
Also used for special effects like gated reverb
-2dB
-2dB
-6dB
-6dB
-10dB
-10dB
-14dB
-14dB
With no limiting, the signal
decays at a natural rate
With severe limiting (100:1 ratio & -10dB threshold the gain is
reduced dramatically (to 1/100th of the original) by the gate when
the threshold is reached.
In-line effects
Mid-term review
Signal processing
Variable Gain Amplifiers
4. Expanders
Definition: does the opposite of a compressor. When the signal falls below a set threshold, the level is cut
according to the set ratio.
Also used to eliminate unwanted low level sounds in speech, music and other recordings, but somewhat
more gently.
Also used for special effects like gated reverb
-2dB
-2dB
-6dB
-6dB
-10dB
-10dB
-14dB
-14dB
With no expander applied, the
signal decays at a natural rate
With an expander in-line using a 3:1 ratio, the signal decays
three times more rapidly than the original once the threshold
is reached.
In-line effects
Mid-term review
Signal processing
Variable Gain Amplifiers
5. De-essers
Definition: a limiter that is only applied within a set
certain frequency range and at a set amplitude
threshold.
All frequencies in the range are sharply attenuated
(ProTools default is 100:1).
Normal use is to reduce the harshness of hard
consonants (‘s’, ‘t’, ‘tch’, etc.)
Typical frequency range for this purpose is 6-7 kHz
and up.
In the settings shown, as soon as a consonant above
7 kHz reaches 0 dB, sharp limiting will occur to
those frequencies only.
Send/Return effects
Mid-term review
Signal processing
Definition:
An effect added to the direct signal on a track or mixer channel via an auxiliary send/ return prior to the output
The amount of effect can be controlled by adjusting the auxiliary send and / or return levels on the mixer
Output
Signal
Send effect
Send/Return effects
Mid-term review
Signal processing
Reverb
Definition: a device or plug-in that simulates
acoustic reverberation, a series of complex sound
reflections varying with the size, proportions, and
reflective surfaces of an interior space.
Digital reverbs simulate acoustic spaces by
providing programmable parmeters such as room
size, room type, decay, etc.
Usual use is as a send/return effect, but can also be
used as an in-line effect if RAM memory permits.
Playing with the mix between direct signal and
reverb then becomes a consideration.
Send/Return effects
Mid-term review
Signal processing
Delay
Definition: a device or plug-in that delays the input
signal by a set number of milliseconds.
In addition to the delay itself, a feedback control
feeds the delayed signal back into itself to varying
degrees (expressed as percentages) and generates
diminishing or infinite repeats of the delay
Most digital delays also include a filter to make the
repeated or single delays sound ‘duller’, similar to
a real acoustic echo.
ProTools’ delay also includes modulation, a cyclic
pitch variation of the delayed signal used to create
chorus and flange effects.
The delay repetitions can also be synced to a tempo
for musical purposes.
Send/Return effects
Mid-term review
Signal processing
Delay
Uses:.
A short delay (less than 40 ms) acts a bit like an
EQ, in that it reinforces or cancels certain
frequencies due to phase differences it causes.
Slap echo is obtained with slightly longer delay
times (40 to 170 ms), and adds a type of ‘ambience’
to the effected sound.
Medium (340 - 680) and longer delays (680 ms +)
are generally used for musical and special effects.
The repetitions of the sound are distinctly audible,
and tempo sync considerations come into play.