Sound Design and Music Technology Past Paper PDF

SOUND DESIGN AND MUSIC TECHNOLOGY 3 / 10 / 2024 - CLASS 1 GENERAL INFORMATION Lectures: Thursday and Friday 12:30-14:00 (last class will be probably on December 20) Exam: multiple choice quiz (18 pt) + podcast (6 pt) + sonorization of a short movie (9 pt) — individual project Exam dates: 4 Feb a...

SOUND DESIGN AND MUSIC TECHNOLOGY 3 / 10 / 2024 - CLASS 1 GENERAL INFORMATION Lectures: Thursday and Friday 12:30-14:00 (last class will be probably on December 20) Exam: multiple choice quiz (18 pt) + podcast (6 pt) + sonorization of a short movie (9 pt) — individual project Exam dates: 4 Feb and 18 Feb 2025 at 10 am Sound is an amazing means to communicate information and to elicit strong emotions. Sound and music are currently a basic component of multimedia products and are integrated in the functionalities of many objects (for ex. we are able to recognize the specific sound of Microsoft Windows when we start our computer). Sounds need to be designed carefully and sound designers know how to choose and control sound parameters such as amplitude, pitch, or timbre to convey sensations, emotions, or functional information. Sound design in commonly used in film-making. In this case we are talking about foley sounds. Foley sounds = the reproduction of everyday sound effects that are added to films, videos and other media. Foley sound is named after sound-effects artist Jack Foley. Nowadays, digital technologies are very powerful tools for designing sounds. THE POWER OF SOUND When a film is shot, usually the video part and the audio part (dialogues and sound effects) are recorded separately. 4 / 10 / 2024 - CLASS 2 INTRODUCTION TO MUSICAL ACOUSTICS What is a sound? TWO DEFINITIONS OF SOUND 1. A mechanical disturbance from a state of equilibrium that propagates through an elastic material medium (external part of sound) → the material medium could be the air (for example, in the vacuum there’s no air so there’s no sound) 2. What is perceived by the ear → this definition takes into account what happens inside our body and our brain (internal part of sound) So sound is both a physical and a psychological phenomenon. 1 A DEFINITION OF ELASTICITY Elasticity is the ability of a deformed material body to return to its original shape and size when the forces causing the deformation are removed. A string can be in a state of equilibrium, but I can deform the string by overshooting it or by compressing it. Equilibrium → deformation → equilibrium → compression → equilibrium → this happens every time we plug the string of a guitar for example. We have sound only if we have a body that starts to oscillate about in the equilibrium point (whereas in the equilibrium point we don’t have any sound). The sound of an elastic body has to propagate through another elastic body → for example, the sound of a drum (which has an elastic membrane) propagates through the ear (which has an elastic membrane as well). Propagation = compression of layers of molecules through the space. The membrane that we have in our ears is similar to the one of the drum where the sound started. Electrical impulses are send to our brain and these are translated into a perception that we call sound. IMPULSIVE SOUND An impulsive sound is a sound of very short duration. Impact sounds are an example of impulsive sounds → they are generated by a short time interaction between 2 objects. The correlation between impact sounds and their sources is quite enduring (= duratura). A strong relation is present between the physical properties of a sound and what we perceive. What is the physical nature of any given recognizable sound as it comes through the air to our ears? What is the nature of perception and in what way does the human mind produce a distillation of those properties of the sound that are in some way interesting or important to it? How can be described the properties of a sound? Frequency (f) = in physics is the number of events per second (taps per second in the experiment). It’s measured in hertz. Period (T) = interval between two repetitive events (time between two repetitions in the experiment). It’s measured in seconds. There’s a mathematical relationship between frequency and period: if we have the frequency we can calculate the period (T = 1/f), and if we have the period we can calculate the frequency (f = 1/ T). INTERVAL BETWEEN TWO EVENTS To calculate the interval between two events (in a series of repeated events) we divide 1 second by the number of taps per second. Ex: if we have 5 taps per second → the time between taps is 1/5 seconds = 0.2 seconds = 200 ms HIGH RATE REPETITIONS What happens if we increase the number of repetitions per second of an impulsive sound? The period decreases. From a physical point of view these are just impulses at a fixed period, but if we change the frequency, we move from a rhythm to a noise. We can describe rhythm with frequencies (slow or fast rhythm), but when we talk about sound we use the word pitch, because a sound can be high or low (high pitch/low pitch). - High frequency = high pitch 2 - Low frequency = low pitch If our impulse repetition rate is less than about 20 repetitions per second, we perceive separate impulses having slow to fast tempo. If our impulse repetition rate is roughly between 20 repetitions and 150 repetitions per second, we perceive a buzz. If our impulse repetition rate is above about 100 repetitions per second, we perceive a tone of progressively higher pitch. PITCH (= tono) Pitch is a perceptual property of sounds that makes it possibile to judge sounds as higher or lower. Pitch is related to the definition of sound as an internal thing because it is a perceptual property. As concern impulses, we saw that pitch is related to repetition rate: fast repetition/frequency = high pitch and slow repetition/frequency = low pitch. The pitch of a tuning fork is quite low and it’s the same for all the time we can hear the sound, it’s the volume that changes → it is composed by a serie of pitches. The sound of a recorded tuning fork is characterized by a shape that is repeated many times, so we can calculate the period and the frequency of this repetition (we know that frequency is related to the pitch). Repetition rate = number between two repetitions → the time between two repetitions is the interval between two waves. It is perceived by our ear as a pitch. A sound-wave can have two different shapes: sinusoidal wave (round shape) and triangular wave. FREQUENCY Frequency is the number of cycles or vibrations undergone (=subite) during one unit of time by a body in periodic motion. It is related to the physical nature of sound. Frequency is usually measured as number of cycles per second. The measurement unit of frequency in the International System of Units (SI) is the Hertz (abbreviated Hz) → 1 Hertz = 1 cycle per second. PERIOD The oscillations of a tuning fork are quite regular, in the sense that the duration and the shape of each oscillation is almost constant → we say that vibrational motion of a tuning fork is, with a good approximation, periodic. A time period (denoted by T) is the time taken to complete one cycle. Period T is inversely proportional to frequency f → when T increases f decreases and viceversa. This relation can be better represented by these math equations: 3 10 / 10 / 2024 - CLASS 3 ABOUT PITCH PERCEPTION Not all sounds have a pitch → pitched sounds are progressively perceived as unpitched (= stonati) when the duration becomes very short. The pitches of two simultaneous sound changes when their frequencies are very close. PITCH-FREQUENCY RELATION Pitch is related to frequency, even though the perceived pitch of a sound is not entirely explained by frequency → perceived pitch is also related to the duration of the sound. There are also some soun’ds that have a definite pitch even though they may lack a repetition rate at all (= potrebbero non avere alcuna frequenza di ripetizione), or they might possess several different repetition rates. From a musical point of view, pitches are organized in scales (like Do, Re, Mi, Fa, Sol, La, Si, Do), the relation between degrees of the scale and frequencies is conventional (culturally based → for example, English people use the first letters of the alphabet Do = C, La = A, etc). If we put two frequencies together we can perceive only one pitch; also the rhythm is related to the frequencies of the two sounds. So, the physical characteristics of sound are not directly related to what we perceive. 11 / 10 / 2024 - CLASS 4 WELL-TEMPERED TUNING Well-tempered tuning is the current tuning convention most used in the European music. The idea is that the pitch system is organized into blocks of 12 tones, called octaves. The frequency interval between two degrees of the scale is not constant. The frequencies of the degrees of each octave are related by a factor 2 (multiplied or divided by 2) → for example, the frequency of the C4 is 261.6, so the frequency of C5 is 523.25 while the frequency of C3 is 130.81. If we have the frequencies of an octave, we can then calculate all other frequencies by following this path. There is a relation between the name of the note (C or Do) and the oscillation frequency of that note. We use sequences of notes to generate a melody, that is the key component of a sound → thanks to this relation, we can also use also frequencies to compose melodies. MIDI KEY NUMBER In digital musical instruments, musical notes are conventionally identified by a number, called key number → every note on a keyboard is associated with a key number. For example: - the middle C (do) on the keyboard is identified by the number 60 - the A (la) with a frequency of 440 Hz is associated with the key number 69 The association between key numbers and notes is specified in the MIDI (Musical Instrument Digital Interface) standard, a protocol for transmitting information between digital music instruments. 4 17 / 10 / 2024 - CLASS 5 TIME ANALYSIS Sound oscillations captured by a microphone and showed on a computer can be observed at different time scales (ex: we can have a second-time scale or lower time scales, like millisecond). The shape of these oscillations represent how the loudness of the sound changes overtime. AMPLITUDE ENVELOPE When we use a time scale of seconds, the shape that we see is called amplitude envelope → the amplitude envelope of an oscillating signal is a smooth curve outlining its extremes. For different sounds we can have different amplitude envelopes. ADSR MODEL One of the most used way to analyze an amplitude envelope is to divide this amplitude envelope in 4 phases, because it is modeled by a four pieces profile, called ADSR: 1. Attack → the higher part of the amplitude (attack time) 2. Decay → decrease of the amplitude (decay time) 3. Sustain → when the amplitude is more or less constant (sustain level) 4. Release (release time) Not all sounds have this 4 phases → some sounds can have just 2 phases. SOUND AND PRESSURE We know that the ear is a sort of pressure-measuring device in which the eardrum is alternately pressed inward and pulled outward in response to the oscillatory fluctuation of pressure above and below the normal atmospheric pressure in the room. When we have a compression, it means that there is high pressure, whereas if we have a rarefaction, it means that there is low pressure. 5 PRESSURE AMPLITUDE Pressure amplitude is a measure of the size of the variation in air pressure caused by a sound oscillation. The higher parts of the curve stand for high pressure, so where there’s a compression of the molecules; instead, the lower parts of the curve stand for low pressure, so where there’s a rarefaction of the molecules. The wave’s peak corresponds to the colliding molecules amass. In pure silence there is a (more or less) constant pressure, called atmospheric pressure → state of equilibrium. The compression of air into a smaller volume (compared to the equilibrium state) causes an increase in pressure. Similarly, the presence of more air (compared to the equilibrium state) within the same volume causes an increase in pressure. Pressure amplitude is a physical aspect of sound, but when we hear a sound we say that it could be more or less loud. LOUDNESS Loudness is the attribute of auditory (= uditiva) sensation in terms of which sounds can be ordered on a scale extending from quiet to loud. Loudness is related to pressure amplitude, even though the perceived loudness of a sound is not entirely explained by the pressure amplitude (so they’re not the same thing). Loudness is an internal processing of sound, so it is a perceptual domain. Loudness is related to pressure amplitude, but it depends also on other factors, like frequency, shape of oscillation, context, etc → for example, two sounds with the same amplitude can have two different loudness (and different frequencies); or two sounds can have the same amplitude and the same frequency, but they have different shapes and loudness. So, loudness depends also on shape. 6 Rather than being related to instantaneous pressure amplitude, loudness is more related to the oscillation pattern within a cycle → we need to introduce parameter to express pressure amplitude. 24 / 10 / 2024 - CLASS 6 RMS AMPLITUDE Different parameters can describe how the amplitude changes along a cycle: Instantaneous amplitude refers to the amplitude of a signal at a specific moment in time → not related to loudness perception Peak amplitude refers to the maximum amplitude in a cycle → not very related to loudness because sound with the same peak amplitude can be perceived with different loudness Mean amplitude refers to the mean amplitude in a cycle → almost useless as it is (almost) zero because we have both positive and negative values Root Mean Square (RMS) Amplitude is defined by the following equation (not required for the exam): The Root Mean Square is calculated in 3 steps: 1. The instantaneous amplitudes (red line) are squared (blue line) 2. The mean of the squared amplitudes in a given time interval T (ex: in a cycle) is calculated 3. Finally, the square root of the mean is calculated from the mean of the squared amplitude (green dotted line) For a sinusoidal sound (and only in this case) ARMS = 0.707 x peak amplitude → for other sounds, we need to calculate it. RMS amplitude is a sort of mean of the instantaneous amplitude of a sound → this value is useful because it is more related to loudness, in comparison to peak amplitude. For example, if you double the peak amplitude, the RMS increases of 6 decibels. RMS is related to peak amplitude, but not always → it depends also on the shape of the sound (for example, the RMS amplitude of a sinusoidal sound is different from the RMS amplitude of a squared sound even if they have the same peak amplitude). Talking about loudness: peak amplitude is not related to the perception of loudness, but loudness depends on the RMS amplitude of a sound (for example, if we have 2 sounds with the same peak amplitude, but different RMS amplitude, the sound with a higher RMS amplitude will have a higher loudness). Sounds with a higher RMS amplitude are also perceived as higher in loudness (even if there are also other aspects to consider) 7  Sounds with a lower RMS amplitude are also perceived as lower in loudness DECIBEL (dB) The pressure amplitude (peak or RMS) can be numerically expressed by means of a measurement unit called deciBel (dB) → the decibel is not a quantity of sound, but it expresses a relationship between any two quantities. Decibel is a measurement unit for pressure amplitude. For example, a car causes a sound 20dB more (+20dB) than a bicycle or that the sound of a bee is 6dB less (-6dB) than that of a bicycle. NOTICE → I can’t simply say that the pressure amplitude of a sound is 20dB, without specifying in reference to what. In some context, the reference sound is conventionally known and we can omit to specify it → for example, in the digital domain, as the sound recorded in a computer, it is common to measure the amplitude of a sound in reference to the maximum amplitude that can be measured with the finite number of digits (usually 16, 32 or 64 bits) of a computer → deciBel Full Scale (dBFS). If we have an RMS = 0 dB it means that is the highest/largest sound that can be measured by this computer or if we have an RMS = -9.0 dB it means that this sound is 9 dB below the maximum. AMPLITUDE/DECIBAL GRAPH In mathematical terms, this graph corresponds to this equation → SOUND PRESSURE LEVEL By definition, the deciBel expresses a relationship between 2 signals. However, in many applications it is useful to express the amplitude of a sound by means of absolute values. When a sound propagates through air, it causes variations/oscillations in the air pressure level. Since the unit of measurement for pressure in the International System is Pascal, the amplitude of these variations can also be measured in Pascal. When referring to environmental sounds, it is preferred to measure the amplitude of pressure variations in dBSPL (in this case the value is called Sound Pressure Level). The Sound Pressure Level (SPL) is calculated by using a standardized reference amplitude A0 → the standardized reference amplitude A0 used i n calculating SPLs is equals to 1/3,530,000,000th of atmospheric pressure and corresponds to 0.00002 Pa (Pascal is the measurement unit for pressure). This value has been chosen because it is the threshold of audibility of a sinusoidal sound with f = 1000 Hz (the lowest sound that we can perceive by our ear). 8 The decibel in the digital domain is called decibel full scale, while the decibel used to measure sounds in the environment is called SPL. 7 / 11 / 2024 - CLASS 9 OTHER REFERENCE LEVELS (not required for the exam) Besides dBFS and dBSPL, other reference levels are used, depending on the context → for example, the VU meter is a device used to display sound amplitude in analog audio equipments, such as audio amplifiers. The reference level of VU meter is usually equally to 1.228 Volts RMS (Volts are used because analog audio is an electric signal). Digital workstations usually allow the user to choose among several options (we should look carefully at the display or read the manual). The most used unit is deciBel LUFS (Loudness Unit Full Scale). HEARING DOMAIN A sound with a frequency oscillation = 1k Hz and a sound pressure level = 0 dB is right on the hearing threshold. The hearing threshold is an equal loudness curve → all the sounds that are on this curve are perceived with the same loudness. 9 EQUAL-LOUDNESS CURVES The equal-loudness curves describes how loudness of sinusoidal signals depend on frequency and SPL → loudness is flatter at higher SPL (we hear the bass better when we turn the volume up). Example: two sounds with different frequency and SPL that are on the same equal loudness curve. PHON Phon is a unit of loudness level → the loudness level of a sound is a subjective measure, rather than an objective one. The loudness level of sounds, in phons, is equal to the sound-pressure level, in decibels, of a sinusoidal sound of 1000Hz, that is perceived at the same loudness level. For example, all sounds on the red line have a loudness level of 50 phons, therefore they will be perceived with (about) the same loudness. Examples: A is a sinusoidal sound with a frequency f = 40Hz, a sound pressure level SPL = 100dB and a loudness level L = 80phon B is a sinusoidal sound with a frequency f = 1000Hz, a sound pressure level SPL = 60dB and a loudness level L = 60phon C is a sinusoidal sound with a frequency f = 3500Hz, a sound pressure level SPL = 98dB and a loudness level L = 110phon The sound C will be perceived louder than the other sounds, because it has the highest loudness level, while sound A has the highest amplitude (RMS), because it has the highest SPL. LOUDNESS UNIT FULL SCALE (LUFS) To display the level of a sound, digital workstations usually allow the user to choose among several options (we have to look carefully at the display or read the manual). The most used unit is deciBel LUFS = Loudness Unit Full Scale. As in the dBFS, the reference level is the Full Scale amplitude, BUT differently from dBFS, the sound is beforehand (= in anticipo) modified (equalized) according to the equal loudness. 8 / 11 / 2024 - CLASS 10 FINE TIME ANALYSIS OF THE TUNING FORK The final part of the decay phase of the tuning fork sound shows a smooth, sinuous line with a well-defined repetition time of about 7.5 ms. Whereas, the first part of the tuning fork sound shows a different behavior → we can observe the superposition of: - a bigger oscillation with a time duration of about 7.5 ms (133 oscillations per second) - a smaller oscillation of about 0.8 ms (1250 oscillations per second) 10 The shape of the first part of the tuning fork sound can be obtained by superposing 2 simple oscillations. SUPERPOSITION OF SINUSOIDS FINE TIME ANALYSIS OF THE GUITAR We can observe the superposition of - a longer oscillation pattern with a time duration of about 12 ms (83 repetitions per second) - shorter (and more or less big) oscillations of different time durations OSCILLATIONS OF A GLOCKENSPIEL We can see repetitions at more than one time-scale. The folds shape is not sinusoidal → it is “sharper”. So we need another approach to analyze the characteristics of complex sounds (such as the guitar or the glockenspiel) → frequency analysis. PRISM AND LIGHT SPECTRUM The prism allows us to see the primary colors composing the white light → we would need something similar to see the simple frequencies composing of a complex sound. 11 FREQUENCY ANALYSIS (SPECTRUM) OF THE GLOCKENSPIEL Each vertical peak corresponds to a specific frequency composing the sound. The glockenspiel sound contains more than one frequency → these frequencies are called partials, because they are the elementary parts that constitute a complex sound. There are 4 main partials: - 651Hz, -40dB - 1799Hz, -50dB - 3531Hz, -27dB - 5856Hz, -52dB The graph of a sound is also called spectrum, by analogy with light → the spectrum of a sound tells us which are the primary frequencies (partials) that compose that sound. DECIBEL FULL SCALE If we look at the vertical axis of the spectrum, we can see some negative values → these values in decibel are not the SPL, otherwise these sounds would be largely below the hearing threshold. Actually, SPL is mostly used to measure the pressure amplitude of a sound traveling on air (→ pressure is about air molecules). When sound is recorded on a computer (or other digital device), the pressure oscillations are first converted into electrical signals, and later into numbers. When we deal with numbers inside a computer, it is most common the use as reference level (A0) the maximum number that can be used to represent the sound amplitude → the resulting unit of measurement is called decibel Full Scale (dBFS). LOGARITHMIC SCALE If we look at the horizontal axis of the spectrum, we can see that ticks and grid are not uniformly distributed. In a linear scale, as in a normal tape measure, equal distances correspond to equal differences between the values → BUT in the horizontal axis of our spectrum, this is not true. On the contrary, equal distances correspond to equal ratio between values → if a scale has this property, it is called a logarithmic scale. Depending on what we want to display, we can choose one or the other scale. FREQUENCY ANALYSIS OF THE TUNING FORK In this graph there are 2 main partials: - 133Hz, -13dB - 1250Hz, -22dB And one smaller oscillation (not visible in the time analysis) → 266Hz, -34dB FREQUENCY ANALYSIS OF THE GUITAR In this graph there are many partials: - 83Hz, -48dB 12 - 166Hz, -28dB - 249Hz, -45dB - 332Hz, -36dB - … SPECTROGRAM (GLOCKENSPIEL) The spectrum is a static representation (as the photo of a radiography) → the spectrogram (aka sonogram), instead, shows how the frequency content of a sound changes over time. Each partial has a different frequency, amplitude (color and line tickness) and duration. There are 4 main partial (from below to above): - 651Hz, -40dB - 1799Hz, -50dB - 3531Hz, -27dB - 5856Hz, -52dB The first partial (5856Hz, -52dB) has the longest duration, while the second (3531Hz, -27dB ) one is the shortest. SYNTHESIS OF A GLOCKENSPIEL SOUND The info from the spectrogram can be used as a sort of recipes to generate or to design a sound → each partial has a different decay time, as observed in the spectrogram and the resulting sound changes over time. 21 / 11 / 2024 - CLASS 13 SPECTROGRAM (TUNING FORK) The spectrogram shows how the frequency content of a sound changes over time. Each partial has a different frequency, amplitude (color and line thickness) and duration. SPECTROGRAM (GUITAR) The spectrogram shows how the frequency content of a sound changes over time. Each partial has a different frequency, amplitude (color and line thickness) and duration. In this sound, the partials follow a clear pattern → the frequency difference between two successive partials is constant. In other words, the frequency of each partials is an integer multiple (multiplo intero) of the same value → in this specific case 83Hz. PARTIALS AND HARMONICS If the frequencies of the partials are integer multiple (= multipli interi) of the same value f0, as in this guitar sound where f0 = 83Hz, the partials are called harmonic partials or simply harmonics → the value f0 is called fundamental frequency. Many musical instruments (piano, violin, flute, singing voice…) 13 produce tones with a fundamental frequency and a variable number of harmonics. A BASIC CLASSIFICATION OF SOUNDS Simple periodic sound → sinusoid Complex periodic sound Non periodic sound SIMPLE PERIODIC SOUND Simple periodic sound is just a frequency → its spectrum is characterized by only one main peak, with harmonic partials. If we sum 5 simple sounds (sinusoids) we can obtain a complex sound (for example, the sound of the organ) → a complex periodic sound can be obtained by summing 5 simple periodic sounds. COMPLEX PERIODIC SOUND A complex periodic sound is defined as complex because it has a different shape from a simple sinusoid and as periodic because its shape repeats itself after a while. In general, each complex periodic sound can be decomposed into a series of sinusoidal partials, each one with a frequency that is an integer multiple of some value → that set of partials is called spectrum. N.B.: periodicity is a necessary condition. 14 DECOMPOSITION OF THE ELECTRONIC ORGAN SOUND The sound of the organ seen before can be obtained as sum of the following 5 sinusoids: SPECTRUM OF THE ELECTRONIC ORGAN The spectrum shows the frequency and the amplitude of every sinusoid into which the sound can be decomposed (5 partials in this case). The first frequency (220Hz) represents the fundamental frequency, while other frequencies are the harmonics. NON PERIODIC WAVES The amplitude of a non-periodic wave can be determined, but not the period and frequency. SPECTRUM OF A NON PERIODIC WAVE The concept of spectrum can also be applied to non-periodic waves → in this case, there are many peaks so there will be non- harmonic partials and it is not possible to identify a fundamental frequency. FOURIER TRANSFORM Nowadays, the most widely used method of obtaining the frequency analysis of a sound is based on the Fourier transform → it is a mathematical transform = a method/an equation to transform a function defined on one domain (time in this case) into another function usually defined in another domain (frequency in this case). As we usually analyze sounds that are finite (in time) and digitalized (by a computer), what is a actually used is the Fast Fourier Transform (FFT): 15 FFT PARAMETRES Window size → it corresponds to N in the FFT equation and it is the number of samples (= campioni) analyzed: if it’s greater it means that we are analyzing a lengthier sound excerpt (= estratto sonoro più lungo) Window axis → we can choose between logarithmic or linear scale Window function → window shape WINDOWING Some disturbing elements in the frequency analysis are due to the ends of the sound (where it is cutted) → to reduce these disturbs, the sound is usually windowed. Different window shapes (= window function) give different results (not required for the exam): 22 / 11 / 2024 - CLASS 14 SPECTROGRAM PARAMETRES TIME-FREQUENCY TRADE-OFF The resolution along the frequency axis increases if the size window N increases The resolution along the time axis (how many frames are calculated) decreases if the size window N increases It is necessary to choose a trade-off (= compromesso) between resolution in frequency and in time. 16 SPECTRUM OF A COMPLEX PERIODIC WAVE SPECTRAL ENVELOPE A spectral envelope is a curve in the frequency-amplitude plane, that represents how the amplitude (or energy) of the signal is distributed over frequency. N.B.: we don’t have to confuse this with the spectral amplitude defined in previous lessons. SPECTRAL FORMANTS Often, the spectral envelope has a shape with peaks and troughs (= depressioni) related to acoustic resonances and anti- resonances → the portion of the spectral envelope with a maximum between two anti-resonances is called formant. PARTIALS = peaks you can find in the spectrum → SPECTRAL ENVELOPE = imaginary line that envelopes the peaks of the partials → FORMANTS are related to the spectral envelope and we can find them in the envelope → PARTIALS ≠ FORMANTS. TIMBRE Timbre is the attribute of auditory sensation which enables a listener to judge that two nonidentical sounds, similarly presented and having the same loudness and pitch, are dissimilar → for example, it is the difference in sound between a guitar and a piano playing the same note at the same volume. Pitch, loudness and timbre are all perceptual properties → 3 most important dimensions of a sound. Timbre depends on: - Its wave form, which varies with the number of partials, or harmonics, that are present, their frequencies and their relative amplitudes - Its spectral envelope - How amplitude changes over time → its amplitude envelope EXAMPLE: GLOCKENSPIEL AND LA 17 First line of the spectrogram from below = fundamental frequency (pitch is related basically to the fundamental frequency) 12 / 12 / 2024 - CLASS 19 WAVES PROPAGATION The propagation rate (v) is the speed at which the wave travels through a medium (speed of a sound) → it depends on the medium (and its physical properties such as its density, elasticity and temperature). WAVELENGHT The wavelength of a periodic wave is the distance between corresponding points (for example, two consecutive compression points) of two consecutive repetitions → the space covered by the wave in the time of a period. The wavelength depends on the propagation speed (rate) of the sound: - if the oscillation frequency (f) increases, the wavelength will be shorter - if the propagation rate (v) increases, the wavelength will be longer REFLECTION, ABSORPTION, TRANSMISSION When the sound propagates in a closed environment and reaches the wall 1. A part of the wave bounces/reflect 2. Another part is transmitted outside the wall 3. Another is absorbed by the wall and converted in heat ABSORPTION COEFFICIENT Most materials absorb and reflect sound to some degree → the sound absorption coefficient is the portion of energy that is absorbed (a value = 1 means full absorption) REVERBERATION When sound hits a surface and is reflected, it bounces off the surface and it is perceived as reverberation or echo, depending on its interaction with other live surfaces in the vicinity. Depending on the path followed by the sound wave, we can have: - Direct sound = sound wave that comes from sound source directly to the listener (the path is the shortest, therefore is the first one we receive) - Early reflections = sound waves we receive just after one bouncing 18 - Late reflections = sound waves that as last IMPULSE RESPONSE REVERBERATION TIME Reverberation time, or decay time, is the time a sound takes to decrease 60 dB-SPL after its steady-state sound level has stopped. 13 / 12 / 2024 - CLASS 20 NORMAL MODE The normal mode of an oscillating system is a pattern of motion in which all parts of the system move sinusoidally with the same frequency. Usually, an oscillating system is characterized by many different normal modes → therefore, the most general motion of a system is a superposition of its normal modes. The motion described by the normal modes takes place at fixed frequencies → these fixed frequencies are known as the natural frequencies or resonant frequencies of the oscillating system. WAVE SUPERPOSITION When two or more vibrations meet, their respective effects add up → the resulting intensity depends on the relative phases of the vibrations: In-phase → the resulting intensity increases Out-of-phase opposition → the resulting intensity decreases Effects of great physical and musical importance can derive from the interference: - standing waves - beats - combination sounds 19 STANDING WAVES Standing waves are a phenomenon resulting from the superposition of sound waves of the same frequency, which propagate in opposite directions → they are characterized by nodes and antonodes (or loops) that maintain their position in space unchanged. There is a limited number of standing waves. STANDING WAVES ON A STRING Tension and density can change from string to string Each normal mode is characterized by a standing wave of different length, therefore it oscillates at a different frequency, according to the table SYMPATHETIC VIBRATION Sympathetic vibration (or sympathetic resonance) is a phenomenon wherein a previously still vibratory body begins to vibrate in response to external vibrations. FREQUENCY RESPONSE It measures the amplitude (and phase) of the output of an oscillating system (a string, a musical instrument, a glass, a table, a room, etc) that vibrates in response to stimuli of different frequencies (usually from 20 to 20,000 Hz). 20