Sound recording and playback Angelo Farina 14122012 Sound

Sound recording and playback • Angelo Farina 14/12/2012 Sound recording and playback 1

Physical nature of sound Origin of Sound: Thermofluidodynamic phenomenon: Particle velocity and variable density of medium (air) Human body can detect sound (p and v) with: ears, but also skin, chest, stomach The trasducers should detect the same quantities 14/12/2012 Sound recording and playback 2

Transducers: microphones From acoustic pressure and particle velocity to physical (electrical) quantities Output signal: voltage (Volts), current (Amperes) or charge (Coulombs) Pressure microphones (related with acoustic pressure) Velocity microphones (related with particle velocity) Hybrid microphones (a proper combination of both quantities) From omnidirectional (100 p and 0 v) to figure of eight (0 p and 100 v) 14/12/2012 Sound recording and playback 3

Microphone directivity patterns Omnidirectional (100, 0) Subcardioid (75, 25) Hypercardioid (25, 75) 14/12/2012 Cardioid (50, 50) Figure-of-Eight (0, 100) Sound recording and playback 4

Microphones Variable pattern microphone: Neumann U 89 i variable-pattern microphone 14/12/2012 Sound recording and playback 5

Cables The weak microphone signal (few m. V/Pa) has to be amplified and transmitted by means of cables Signal contamination can occur inside the cable, if not properly shielded (balanced) with two oppositepolarity signals Balanced audio cables with XLR connectors (3 pins) 14/12/2012 Sound recording and playback 6

Preamplifiers They should simply amplify the signals, but often they also process the signals: • linearly (band pass frequencies), for phantom power supply to mics • non-linearly (compression, harmonic distortion) it should be avoided during room acoustics measurements. 2 -channels tube microphone preamplifier 14/12/2012 Sound recording and playback 7

ADC (Analog to Digital Converter) Conceptually a black box connecting with two wires: • Analog input (sound signal) • Digital output (serial digital interface) Two different types of ADCs: 1. PCM converters (Pulse Code Modulation – CD, DAT, DVD) 2. Bitstream converters (DSD, Direct Stream Digital, also called single-bit, employed in SACD). 14/12/2012 Sound recording and playback 8

ADC (Analog to Digital Converter) 2 PCM Converters A master clock defines with high precision the instants at which the analog signal has to be “sampled” (Shannon theorem) Applications Resolutions CDs 44100 Hz 48000 Hz 96000 Hz 192000 Hz DAT, DVD Video DVD audio HD rec. Special soundcards 14/12/2012 Sound recording and playback 9

ADC (Analog to Digital Converter) 3 Low-pass filtering must be applied before entering the ADC, otherwise the signal will be aliased Example: pure tone of 35 k. Hz Sample rate 48 k. Hz, k. Hz Nyquist freq. 24 k. Hz, difference = 11 After digital conversion will be 13 k. Hz (i. e. 24 -11 k. Hz) Solution: low-pass antialiasing filters, oversampling 14/12/2012 Sound recording and playback 10

ADC (Analog to Digital Converter) 4 Also vertical axis (amplitude) is discretized. Typical resolutions: 16 bits; 20 bits; (24 bits) Example: maximum voltage +5 V Discretization with 16 bit (32767 steps) means 90 d. B Discretization with 20 bit (524272 steps) means 114 d. B High-end, 2 -channels ADC unit (24 bits, 192 k. Hz, Firewire interface) 14/12/2012 Sound recording and playback 11

ADC (Analog to Digital Converter) 5 Bitstream Converters • The idea arises from oversampling: increasing sample frequency it would be possible to increase amplitude resolution (bits) • Sample rate: 2. 88 MHz and 1 bit resolution: dividing to 2 for 6 time is equivalent to CD audio sample rate, but only with 7 bits! • In order to enhance high freq. resolution, a proper noise shaping of high order is required, suitable for static (non transient) signals. • Below 88 Hz the Bitstream converters outperform PCM conv. 14/12/2012 Sound recording and playback 12

ADC (Analog to Digital Converter) 6 Bitstream Converters • The Bitstream converters are widely employed with SACD system (Super Audio CD), co-developed by Sony and Philips. • However they are much more expensive that 24 bit 96 k. Hz PCM a low-cost multichannel USB-2 soundcard, equipped with 2 microphone preamps 14/12/2012 Sound recording and playback 13

Digital Signal Processing Waveform editors sampled waveform displayed as amplitude vs time (time domain) 14/12/2012 Sound recording and playback 14

Recording/playback methods • • • Mono followed by amplitude panning (stereo or surround) Stereo (ORTF on 2 standard loudspeakers at +/- 30°) Discrete ITU 5. 1 (from different 5 -mikes layouts) Full 3 D Ambisonics 1 st order (decoding the B-format signal) 2 D Ambisonics 3 rd order (from Mark Poletti’s circular array microphone) Wave Field Synthesis (from the circular array of Soundfield microphones) Hybrid methods (Ambiophonics) HOA (high order Ambisonics, Eigenmike) SPS (Spatial PCM Sampling) 14/12/2012 Sound recording and playback 15

Traditional “surround” – amplitude panning 5 ch. 1 ch. 5 ch. M I X 5 ch. Surround panner • Each mono track is positioned artificially, by means of “pairwise” or “advanced” amplitide panning 14/12/2012 Sound recording and playback 16

Panning laws • Instead of just feeding a single loudspeaker with each mono track, it is advantageous to “pan” each signal ver two or more loudspeakers “Pairwise Panning” with constant Peter Craven’s Ambisonics panning power- each signal is always sent at 5 th order – all the loudspeakers are always fed to just 2 loudspeakers 14/12/2012 Sound recording and playback 17

Panning laws • An alternative display of the panning laws is by charting the polar patterns of a number of “virtual microphones”, each feeding the corresponding loudspeaker “Pairwise Panning” 14/12/2012 Peter Craven’s panning Sound recording and playback 18

ORTF Stereo 60° 2 Microphones 2 Loudspeakers Playback occurs over a pair of loudspeakers, in the standard configuration at angles of +/- 30°, each being fed by the signal of the corresponding microphone 14/12/2012 Sound recording and playback 19

Binaural (Stereo Dipole) Original 2 -channels recording of the signals coming from N sources d. Nr N … 3 d. Nl xr Cross-talk canceller d 2 r 2 20° xl d 2 l 1 d 1 r d 1 l Reproduction occurs over 2 loudspeakers angled at +/- 10°, being fed through a “cross-talk cancellation” digital filtering system 14/12/2012 Sound recording and playback 20

Binaural (Stereo Dipole#2) 14/12/2012 Sound recording and playback 21

Binaural (Stereo Dipole#3) hll hrl fll frl hlr hrr flr frr 14/12/2012 Sound recording and playback 22

Binaural (Dual Stereo Dipole) advantages: Scheme n n n 3 D sound reproduction Rotating of the head The cross-talk filters could equalise also the loudspeakers disadvantages: n Subwoofer 14/12/2012 n Sound recording and playback Low frequencies Coloration outside the “sweet spot” 23

Binaural (Dual Stereo Dipole#2) Frontal Rear Quested 2108 monitors Quested F 11 P monitors 14/12/2012 Sound recording and playback 24

“Discrete” microphone arrays • Many different geometrical layouts were proposed – each microphone feeds the corresponding loudspeaker 14/12/2012 Sound recording and playback 25

“Discrete” microphone arrays • In a discrete system, each microphone feeds just the corresponding loudspeaker: Microfoni 14/12/2012 Altoparlanti Sound recording and playback 26

“Discrete” microphone arrays • Williams MMA • INA-5 Schema del sistema microfonico Williams MMA C : Cardioide, 0° L, R : Cardioide, ± 40° LS, RS : Cardioide, ± 120° 14/12/2012 Schema del sistema microfonico INA-5 C : Cardioide, 0° L, R : Cardioide, ± 90° LS, RS : Cardioide, ± 150° Sound recording and playback 27

ITU 5. 1 surround OCT Schematic of the setup C : Cardioid, 0° L, R : Super Cardioid, ± 90° LS, RS : Cardioid, ± 180° 14/12/2012 Sound recording and playback 28

“Discrete” microphone arrays Why such a limited success? • Little control during recording and during post-processing • Fixed angle of covering, which cannot be adjusted after the recording is done • It makes it difficult to add to the recording the sound track of separately-recorded voices or instruments • Currently available directivity patterns of first-order microphones do not correspond with the ideal, asymmetrical patterns implemented by optimal panning laws. 14/12/2012 Sound recording and playback 29