Loudspeaker Data Reliable, Comprehensive, Interpretable

Introduction Biography: 1977-1982 Study Electrical Engineering, TU Dresden 1982-1990 R&D Engineer VEB RFT, Leipzig, 1992-1993 Scholarship at the University Waterloo (Canada) 1993-1995 Harman International, USA 1995-1997 Consultancy 1997 Managing the KLIPPEL GmbH 2007 Professor for Electro-acoustics, TU Dresden Wolfgang Klippel Klippel GmbH My interests and experiences: electro-acustics, loudspeakers digital signal processing applied to audio psycho-acoustics and measurement techniques

Left Audio Channel Right Audio Channel Agenda Perception Final Audio Application (Room, Speaker, Listening Position, Stimulus) Audio-System (Transducer, DSP, Amplifier) Transducer (woofer, tweeter) 1. Perceptual and physical evaluation at the listening point perceptive modeling & sound quality assessment auralization techniques & systematic listening tests 2. Output-based evaluation of (active) audio systems holografic near field measurement of 3D sound output prediction of far field and room interaction nonlinear distortion at max. SPL 3. Comprehensive description of the passive transducer parameters (H(f), T/S, nonlinear, thermal) symptoms (THD, IMD, rub&buzz, power handling) 3

Left Audio Channel Right Audio Channel Objectives Perception Final Audio Application (Room, Speaker, Listening Position, Stimulus) Audio-System (Transducer, DSP, Amplifier) Transducer (woofer, tweeter) clear definition of sound quality in target application filling the gap between measurement and listening numerical evaluation of design choices meaningful transducer data for DSP and system design selection of optimal components maximal performance-cost ratio smooth communication between customer and supplier 4

Objective Methods for Assessing Loudspeakers Room Parameters Loudspeaker Parameters Parameter-Based Method e.g. T/S parameter, amplitude and phase response, nonlinear and thermal parameters Stimulus Loudspeaker- Room Model nonlinear Psychoacoustical Model nonlinear Sensations e.g. music, test signals Distortion Measurement e.g. THD, IMD, rub&buzz distortion Perceptual Quality Method e.g. predicted preference 5

Perceptual Evaluation of Signal Distortion reference signal distortion test signal Basic monaural processing Basic Auditory Sensations Ideal Conceptions stimulus test signal Binaural Processing Loudness Fluctuations Sharpness S Coloration V Spaciousness R Localization Perceived Defects DS DV DR + Overall Quality Loss reference signal distortion Basic monaural processing The basic auditory sensations are the dimensions of the perceptional space and describe the audibility of the distortion Perceived defects consider the ideal conceptions and the impact on quality 7

Auralization of Signal Distortion Input Signal Output Signal u (t ) s 0 e p(t) Linear Model Nonlinear Model Unpredictable Dynamics d lin (t) Linear Distortion d nlin (t) Nonlinear Distortion d irr (t) Irregular Distortion n(t) Noise Model OBJECTIVES: 1. Virtual enhancement or attenuation of distortion components 2. Systematic Listening Tests 3. Defining a value S DIS in db describing the distance to the audibilty threshold 10

Finding Audibility Thresholds histogram of the audibility thresholds of 55000 participants of a listening test at www.klippel.de weighted up and down method enhancement S DIS attenuation audibility threshold S DIS =-15 db low distortion 11

Subjective and Objective Evaluation Objective Subjective Engineering Evaluation Listening Test + Auralization Evaluation Marketing Management Perceptive Modeling S DIS Physical Data Distortion, Maximal Output Displacement, Temperature Audibility of distortion Perference, Evaluation of Design Choices Clues for Improvements Defining target specification Tuning to the market Performance/cost ratio 13

Left Audio Channel Right Audio Channel Agenda Perception Final Audio Application (Room, Speaker, Listening Position, Stimulus) Audio-System (Transducer, DSP, Amplifier) 1. Perceptual and physical evaluation at the listening point Transducer (woofer, tweeter) perceptive modeling & sound quality assessment auralization techniques & systematic listening tests 2. Output-based evaluation of (active) audio systems holografic near field measurement of 3D sound output prediction of far field and room interaction nonlinear distortion at max. SPL 3. Comprehensive description of the passive transducer parameters (H(f), T/S, nonlinear, thermal) symptoms (THD, IMD, rub&buzz, power handling) 14

O U T 1 O U T 2 I C P 1 MIC1 LINE1 P U S H P U S H LINE2 MIC2 I C P 2 P W R I 0 O U T 1 O U T 2 I C P 1 MIC1 LINE1 P U S H P U S H LINE2 MIC2 I C P 2 P W R I 0 Evaluation of the Audio Product Measurement in Target Application Measurement under Standard Condition considering room, distance, ambient noise and other conditions (Standard) living room transfer of the audio system Anechoic room Suppressing the influence of acoustical environment Definition of target performance as perceived by final user Physical characteristics (comprehensive, simple to interpret, comparable, reproducible) Auralization/Listening Test Perceptual Evaluation Loudspeaker Development 15

Characteristics defined by IEC 60268-5 1. Impedance (rated value, Z(f)-curve, Qts, Vas) 2. Input voltage (rated noise, short + long term maximal) 3. Input power (rated noise, short + long term maximal) 4. Frequency characteristics (rated range, fs, fvent) 5. SPL in stated band, sensitivity for 1 W 6. SPL response for voltage, H(f), effec. freq. range 7. Output (acoustic) power, efficiency 8. Directivity (pattern, rad. angle, index, coverage) 9. Amplitude nonlinearity (THD, IMD) The scope of this standard is limited to passive loudspeaker systems! 16

Active Loudspeaker Systems digital audio stream Properties of the black box depend on control parameters and stimulus Black box No access to internal states drivers Sound Field Near Field Far Field Evaluation is based on evaluation of acoustical output control parameters (e.g. attenuation) 17

IEC 60268-5 applicable to Active Systems? can be applied, need modification, not applicable 1. Impedance (rated value, Z(f)-curve, Qts, Vas) 2. Input voltage (rated noise, short + long term maximal) 3. Input power (rated noise, short + long term maximal) 4. Frequency characteristics (fs, fvent) 5. SPL in stated band, Sensitivity for 1 W 6. SPL response for voltage input, H(f), effec. freq. range, 7. Output (acoustic) power, efficiency 8. Directivity (pattern, rad. angle, index, coverage) 9. Amplitude nonlinearity (THD, IMD) 18

Modern Audio Systems New Requirements: Audio systems become active no access to the electrical terminals of the transducer digital signal processing dedicated to the transducer amplifiers with more capabilities Audio systems become portable main axis of radiation, sweet point and position of the listener are not defined battery powered Audio systems become personal (hand-hold devices) listener is in the near field of the source Audio systems become smaller, lighter using green transducer technologies (efficient, nonlinear) 19

Integration of DSP, power amplification and electro-acoustical conversion amplifiers DSP Nonlinear components protection Tweeter Digital audio input Equalizer Limiter X-over protection linearization Midrange Gain Control protection linearization Woofer drivers Control input Smart technologies (DSP) saves hardware resources and energy more acoustical output at reduced weight, size and cost Green Speaker Technology 20

New Standards required for Evaluating Active and Passive Loudspeaker Systems Applicable to active and passive systems (prototypes, final and competitive products) Describing the radiated direct sound at any point within the listening area (including near field) Consideration of room-loudspeaker interaction Assessment of maximal acoustical output Irregular loudspeaker defects (rub, buzz, leakage, particles, loose connections) Comprehensive set of data (low redundancy, easy interpretation) Bridging QC and R&D Bridging perceptive and physical evaluation Currently discussed in standard committees 21

Small Signal Performance Specifications for Active and Passive Loudspeaker Systems new On-Axis Sound Pressure at reference distance r ref =1m SPL(f) response Phase response (group delay (f) response) Directivity a) single-value characteristics sound power response P a (f) directivity index D i (f) b) 2D far-field data pressure distribution p(θ, ) on a spherical surface at large distance from the source p(θ, ) (balloon plot, beam pattern) c) 3D near/far-field data sound pressure p(θ,, r ) at any point r in the space beyond the sound source (spherical wave expansion) 22

azimutal angle 2D far-field data 4.1 khz at distance r=4m 6.1 khz at distance r=4m SPL 90 270 on-axis 180 0 90 frequency Balloon Plot -90 Distance r >> dimensions d of the loudspeaker Distance r >> wavelength Beam Pattern 23

Complete 3D Information Required Sound Pressure at 7.6 khz In the following application the listerner is closely located to the source: personal audio equipment (smart phone) multimedia (tablet, notebook) studio-monitor car audio loudspeaker Near Field far field data are less important 24

Holografic Measurement of the radiated direct sound in the complete 3D space 1. Scanning the sound pressure in the near field of the source Loudspeaker microphone 2. Expansion in spherical waves monopol z dipols r quadropols φ p out ( r,,, ) N n (2) m cn, m( ) hn ( kr) Yn (, ) n 0 m n Coefficients Hankel function Spherical Harmonics 3. Results: - frequency dependent set of coefficients - point r 0 of expansion - radius r s of validity (scanning surface) - order N of expansion c n, m( ) 25

Practical Realization near field scanning by using robotics Example of a Near-Field Scanner scanning in various coordinates (cylindrical, spherical, cartesian) Objectives of the robotics: 1. Acoustical properties transparent, low noises 2. Flexible scanning grid scanning close to the source accurate positioning on multiple layers 2 π half-space (driver in baffle) 4π full-space (compact sources) 3. High-Speed measurement simultanous positioning in 3 coordinates multiple channel aquisition (mic array) 4. Wide range of application from smart phone to line array heavy systems (> 500 kg) slim system (> 4 m) cost effective, portable 26

Evaluation of a Notebook Application of Near-field Acoustical Holography 1. Measurement of the sound pressure distribution 3. Extrapolation of the sound pressure at any point 2. Expansion outside the into scanning spherical surface waves far field near field r r 0 r s scanning surface close to the source 27

Line Arrays in Professional Audio Application of Near-field Acoustical Holography a 1) Near-field Measurement large dimension a of box anechoic room is too small for far field condition: distance r >> dimension a distance r >> wave length λ 2) Comprehensive 3D data coefficients of spherical wave expansion direct sound in near and far field (any distance) no redundancy (angular resolution at minimal data size) more information provided by GLL files 3) Input for Numerical Simulation Tools superposition of wave expansions design and evaluation of line arrays room interaction 28

Large Signal Performance Specifications for Active and Passive Loudspeaker Systems Maximal SPL max at reference point (1 m, on-axis), in rated frequency range Effective frequency range (Upper and lower limits f lower,l < f < f upper,l ) Compression of fundamental component (thermal and nonlinear effect) Harmonic distortion (Equivalent input distortion) Intermodulation distortion (IMD, MTD) Impulsive distortion (PHD, CHD) indicating rub&buzz, loose particles Modulated noise (MOD) indicating air leakage Durability verified in accelerated life test 29

Condition for Large Signal Measurements new IEC TC100 standard project Output-based Evaluation of audio systems gain and u max depend on measurement setup! ~ test signal* (chirp in rated frequency range) gain? maximal input value u max? u(t) optical analogue wavefile wireless active audio system (no access to internal states) sound field drivers Output* SPL MAX r e evaluation point* *conditions defined by manufacturer control parameters* 30

Interpretation and Benefits of SPL MAX Example as specified by a manufacturer: SPL MAX, =108 db for 60 Hz < f < 3kHz (default test chirp, 1m on-axis) SPL MAX is a single-valued characteristic describing the limit of the acoustical output (but not sound quality of the system) is rated by the manufacturer considering target application depends on rated conditions (working range, reference point, stimulus,...) can be generated by the audio system without damage provides a fast way for adjusting the input level of any stimulus during measurements 31

db - [V] (rms) Definition of SPL MAX SPL MAX is the mean short-term SPL in the rated frequency range generated by a sinusoidal chirp at the reference point. 120 100 SPL MAX Fund. mean (0.06 to 3 khz) THD Fundamental 80 60 40 rated frequency range 20 0-20 20 50 100 200 500 1k 2k 5k 10k 20k Frequency [Hz] f l f u 32

db - [V] (rms) Short-Term Compression reveals mechanical nonlinearities only (no voice coil heating ) system excited by a chirp (T=1 s) generating SPLmax at the evalution point 125 120 115 110 105 100 95 90 85 80 75 70 65 60 linear prediction short-term fundamental (1 s) KLIPPEL 20 50 200 500 2k 20k Frequency [Hz] 33

db - [V] (rms) Long-Term Compression reveals effects of mechanical nonlinearities and voice coil heating system excited by a chirp (T=1 min) generating SPLmax at the evalution point 125 120 115 110 105 100 95 90 85 80 75 70 65 60 KLIPPEL linear prediction long-term fundamental (1 min) 20 50 200 500 2k 20k Frequency [Hz] 34

db - [V] (rms) Harmonic Distortion system excited by a chirp (T=1s) generating SPLmax at the evalution point 120 Fundamental Fund. mean (0.07 to 20 khz) 100 3rd Harmonic 80 2nd Harmonic 60 THD 40 20 0-20 20 50 100 200 500 1k 2k 5k 10k 20k Frequency [Hz] 35

db - [V] (rms) Higher-Order Distortion for assessing rub&buzz and other irregular loudspeaker defects system excited by a chirp (T=1s) generating SPLmax at the evalution point 120 100 Fundamental Fund. mean (0.07 to 20 khz) 80 60 PHD limit (-40dB) 40 20 0 Absolute PHD peak value of higher-order distortion -20 20 50 100 200 500 1k 2k 5k 10k 20k Frequency [Hz] 36

Interfaces between Signal Processing, Electronics, Transducer, Acoustical Environment Example: Active Loudspeaker System enclosure software HP transducer x x horn sound field Digital input DSP BP LP amplifiers x x x x x mechanical measurement acoustical measurement electrical measurement 39

How to Specify the Optimal Transducer? Parameters give a comprehensive set of data!! Should be transformed into parameters 1. Parameters (independent of stimuli) Acoustical transfer functions (from near-field holography) Mechanical transfer functions (from laser scanning) Small signal parameter T/S Large signal parameters (thermal, nonlinear) 2. Stimulus-based Characteristics Maximal SPL Nonlinear distortion (THD, IMD, XDC) Symptoms of irregular defects (rub, buzz, leakage,...) Coil temperature, compression, Pmax 40

Important Transducer Parameters 1. Linear parameters of Motor and Suspension T/S parameters (R e, M ms, R ms,...), lambda (), electr. impedance 2. Nonlinear lumped parameters of motor and suspension Bl(x), L e (x), K ms (x), L e (i) 3. Thermal parameters thermal resistances R tv, R tm and capacities C tv, C tm 4. Linear distributed mechanical parameters mechanical transfer function H x (r c,j ), cone geometry z(r), AAL response 5. Sound pressure responses (transducer in infinite baffle) spherical wave expansion, on-axis response, directivity 6. Mechanical or acoustical load Mechanical Admittance Y(j ) of the coil 41

Left Audio Channel Right Audio Channel Conclusions Perception Final Audio Application (Room, Speaker, Listening Position, Stimulus) Audio-System (Transducer, DSP, Amplifier) Transducer (woofer, tweeter) 1. Development of modern audio system requires different kinds of models, characteristics and measurement techniques 2. Perceived sound quality depends on the final audio application, perceptual processing, training and expectations of the customer 3. Output-based evaluation of active audio-systems is under discussion (join the IEC or AES standard groups) 4. Parameters (independent on the stimulus) play an important role in tranducer design and system integration 49

Thank you! 50