Publications

Seventy-four recent physiological and psychological studies revolving around human perception and its bandwidth were reviewed. The brain has ever only learned about the world through our five primary senses. With them, we receive a fraction of the information actually available, while we perceive far less still. A fraction of a fraction: The perceptual bandwidth. Conscious perception is furthermore influenced by long-term experience and learning, to an extent that it might be more accurately understood and studied as primarily a reach-out phenomenon. Considering hearing, time is found to be a determining factor on several planes. It is discussed how such sentient observations could be taken into account in pro audio, for instance when conducting subjective tests; and the term “slow listening” is devised.

Immersive monitoring systems contain a large number of loudspeakers, each strongly and differently influenced acoustically by the room boundaries. This creates uncertainty regarding the actual in-room frequency response, results in large acoustical differences between individual loudspeakers, and can reduce the transparency and reliability of monitoring on such system. The aim of this work is to demonstrate this influence and then study the importance of using multiple measurement microphone positions instead of one measurement microphone position when preforming in-room frequency response calibration and system alignment. An NHK 22.2 immersive monitoring setup demonstrates the variability typical of immersive layouts, and the influence individualized in-room monitoring loudspeaker equalization can bring. After this, various spatial distributions in the vicinity of the main listening location in 12 rooms are compared to the typical situation of using single microphone location at the listening location. When using equalization to remove response differences between loudspeakers, no significant differences were found between single point measurements and spatial averages for small spatial averaging displacements in rooms having low reverberation times.

In-room estimates of loudspeaker responses for professional use are typically taken either at one microphone location, replacing the listener with a microphone, or averaging in space, at multiple locations at and relatively close to the listening location. In this work, a number of listening rooms were measured using 18 precisely defined locations. In-space averaging in combination with frequency domain averaging can increase the stability of the frequency response estimate. However, the spatial averaging points used in taking a measurement should be chosen based on the room acoustics and on the application. Spatial averaging across a wide area may come with the risk of compromising the result at the main listening position.

Aki Mäkivirta, Jussi Väisänen, Ilpo Martikainen, Thomas Lund and Siamäk Naghian

May 2017, AES 142nd International Convention, Berlin, Germany

Complementary to precision microphones, creating an ideal point source monitoring speaker has long been considered the holy grail of loudspeaker design. Coaxial transducers unfortunately typically come with several design compromises, such as adding intermodulation distortion, giving rise to various sources of diffraction, and resulting in somewhat restricted maximum output performance or frequency response. In this paper, we review the history of coaxial transducer design, considerations for an ideal point source loudspeaker, discuss th performance of a minimum diffraction coaxial loudspeaker and describe novel designs where the bottlenecks of conventional coaxial transducers have been eliminated. In these, the coaxial element also forms an integral part of a compact, continuous waveguide, thereby further facilitating smooth off-axis dispersion.

Aki Mäkivirta and Thomas Lund

September 2016, AES 141st International Convention, Los Angeles, USA

In-room estimates of loudspeaker responses at the listening location are typically taken either at one microphone location, replacing the listener with a microphone, or averaging in space, at multiple microphone locations at and relatively close to the listening location. In-frequency averaging can attenuate the locality of the frequency response features in mid and high frequencies. In-space averaging extracts the common frequency response features visible in all the measurement positions. Spatial weighting combined with frequency domain averaging can increase the stability of the frequency response estimate for the features relevant for the subjective compensation of the sound color at the listening location. Spacing out the spatial average measurement points affects the nature of the spatial average and the focus on the frequency response features common to the measurement points. The spatial averaging points used in taking a measurement should be chosen based on the intention of the room equalization.

Aki Mäkivirta, Juha Holm, Juha Urhonen, Ilkka Rissanen, Jussi Väisänen

August 2015, Genelec Oy, Iisalmi, Finland

The statistical distribution of the sound event duration in a cinematic immersive audio track is considered as information to determine the short and long term sound pressure output capacities needed in loudspeakers and subwoofers. The statistical distribution of audio event duration in cinematic audio tracks is studied to evaluate the needed monitoring system capacity. Monitoring system capacity requirement is also related to the monitoring room size, the listening distance, and the room reverberation time. Impact of using bass management and selection bass management crossover frequency are not considered. An attempt is made to develop these factors into a design guideline enabling successful monitoring room design and selection of the acoustic and physical characteristics of the monitoring loudspeakers and subwoofers needed in the room of a certain physical size.

Jaakko Nisula, Juha Holm, and Aki Mäkivirta

August 2013, AES 51st International Conference, Helsinki, Finland

Walls of a loudspeaker enclosure exhibit vibrations and resonances, creating unwanted acoustic output. These vibrations are excited by the mechanical forces from the drivers and by the acoustic pressure inside the enclosure. Enclosure vibrations cannot be satisfactorily simulated using a mechanical model that only includes the mechanical conduction of vibration in the enclosure. Also acoustical phenomena have to be considered to get accurate results. In this paper we show how Finite Element Analysis using a coupled model of mechanics and acoustics can predict enclosure vibrations and their causes with good resolution.

Juha Holm, and Aki Mäkivirta

August 2013, AES 51st International Conference, Helsinki, Finland

Finite element method applied in frequency domain can show diffraction effects, but separation of the excitation from the diffracted wave fronts is difficult as the excitation dominates the pressure field. Time domain simulation overcomes these problems by enabling separation of the pressure fields in time. Time domain separation of the excitation and diffracted wave fronts enable not only time domain analysis but also specific frequency domain analysis of the diffracted sound. This paper demonstrates how time domain finite element analysis can be used to simulate diffraction in loudspeakers. The method is demonstrated in three geometries where the models are excited with one cycle of Gaussian windowed 5 kHz sine wave. The geometries demonstrate how rounded edges and a waveguide in front of the source can reduce diffraction.

Darren Rose

August 2013, Genelec Oy, Iisalmi, Finland

In 2002 the author made a comparison of the audio quality, audio quantity and cost of some commonly available power amplifier modules. This paper will investigate the hypothesis that today the best value can obtained using Class D amplifiers. This will be studied by comparing four topologies. The amplifiers have been measured in the same controlled conditions representing an application in an active loudspeaker. To make a fair cost comparison the amplifiers have been assembled on PCBs from components in-house rather than using complete ready power amplifier modules. In addition to the audio and cost criteria, energy efficiency will be considered.

Mika Rautiainen, Hannu Aska, Timo Ojala, Matti Hosio, Aki Mäkivirta and Niko Haatainen - MediaTeam Oulu, Department of Electrical and Information Engineering

September 2009, University of Oulu, Finland and Genelec Oy, Iisalmi, Finland

IP networks allow constructing versatile device configurations for multimedia streaming. However, the stochastic nature of the packet-switched data transmission may complicate IP-based implementations of some conventional applications such as analog wired transmission of synchronized multi-channel audio. This paper introduces a multimedia streaming system based on the synchronization of multiple playback clients as a ‘swarm’. The proposed ‘swarm synchronization’ mechanism is based on precise clock synchronization with the PTP protocol and adjusting the client-specific sampling rates according to the true playback rates of other clients.

Timo Hiekkanen, Aki Mäkivirta, Matti Karjalainen - Department of Signal Processing and Acoustics, Helsinki University of Technology, Espoo, Finland and Genelec Oy, Iisalmi, Finland

April 2009, J. Audio Eng. Soc., Vol. 57, No. 4

The precise location of a loudspeaker in a listening room is known to affect loudspeaker preference ratings. When multiple loudspeakers are compared, the evaluation is limited by the poor human auditory memory. To overcome these problems, a method to evaluate and compare loudspeakers using headphones is proposed. The method utilizes personal headrelated transfer functions in rendering the sound field recorded in a standard listening room with an artificial head. The equalization of circumaural headphones and the artifical-head responses for individual listeners are investigated. Formal listening tests are conducted to examine differences between the proposed binaural method and real loudspeakers in a standard listening room. Listening tests show that the virtualized loudspeakers can be nearly imperceptible from reality in many but not all cases.

Timo Hiekkanen, Aki Mäkivirta, Matti Karjalainen - Department of Signal Processing and Acoustics, Helsinki University of Technology, Espoo, Finland and Genelec Oy, Iisalmi, Finland

May 17-20th 2008, AES 124th Convention, Amsterdam, The Netherlands

The precise location of a loudspeaker in a listening room is known to affect loudspeaker preference ratings. When multiple loudspeakers are compared, the evaluation is limited by the poor human auditory memory. To overcome these problems, a method to evaluate and compare loudspeakers using headphones is proposed. The method utilizes personal headrelated transfer functions in rendering the sound field recorded in a standard listening room with an artificial head. The equalization of circumaural headphones and the artifical-head responses for individual listeners are investigated. Formal listening tests are conducted to examine differences between the proposed binaural method and real loudspeakers in a standard listening room. Listening tests show that the virtualized loudspeakers can be nearly imperceptible from reality in many but not all cases.

Andrew Goldberg

2005, Proceedings of The Institute of Acoustics, Vol 27

A listening test system designed to measure the threshold of audibility of the decay time of low frequency resonances is described. The system employs the Parameter Estimation by Sequential Testing (PEST) technique and the listening test is conducted on calibrated headphones to remove factors associated with the listening environment. Program signal, replay level, and resonance frequency are believed to influence decay time threshold. A trial listening test shows that the system reveals realistic results but the temporal resonance modelling filter requires some adjustment to remove audible non-modal cues. Transducer limitations still affect the test at low frequencies and high replay levels. Factors for a future large-scale listening test are refined. Early indications are that temporal decay thresholds rise with reduced frequency and SPL.

Andrew Goldberg, Aki Mäkivirta and Ari Varla

October 2004, The Institute of Acoustics Reproduced Sound 20 Conference, Oxford UK / November 2004, AES 117th Convention, San Francisco, CA, USA

In professional audio applications, small loudspeakers are often mounted on or near (within the loudspeaker’s near field region) large solid surfaces, such as mixing consoles, desktops and work surfaces. In approximately two-thirds of loudspeakers mounted in such a fashion, the magnitude response is compromised in a predictable and systematic way. An upward deviation of peak value 5.0 dB ± 1.5 dB centred on 141 Hz ± 31 Hz was observable in approximately 80% of the cases studied. An additional Room Response Control in active loudspeakers is proposed to compensate for this aberration. A statistical analysis of 89 near-field loudspeakers helps define the correction filter, and quantifies the effectiveness of the fixed filter design. Use of the proposed filter in an automated response optimisation algorithm for in-situ response equalisation is demonstrated.

Andrew Goldberg and Aki Mäkivirta

May 8-11th 2004, AES 116th Convention, Berlin Germany

We compare the room response controls available in active loudspeakers to a third-octave graphical equaliser. The room response controls are set using an automated optimisation method presented in earlier AES publications. A third-octave ISO frequency constant-Q graphic equaliser is set to minimise the least squares deviation from linear within the passband in a smoothed acoustical response. The resulting equalisation performance of the two methods is compared using objective metrics, to show how these standard room response equalising methods perform. For all loudspeaker models pooled together, the room response controls improve the RMS deviation from a linear response from 6.1 dB to 4.7 dB (improvement 22%), whereas graphic equalisation improves the RMS deviation to 1.8 dB (improvement 70%). Both equalisation techniques achieve a similar improvement in the broadband balance, which has been shown to affect a subjective lack of colouration in sound systems. The optimisation time for a graphic equaliser is up to 48 times longer compared to that for active loudspeaker room response controls.

Matti Karjalainen, Poju Antsalo, Aki Mäkivirta and Vesa Välimäki

May 8-11th 2004, AES 116th Convention, Berlin Germany

Modal equalization has recently been of research interest in order to improve sound reproduction in rooms that have excessively strong modes at low frequencies. Instead of acoustic treatment by expensive and space-reserving absorbing structures, modal equalization is based on DSP affecting the electric-to-acoustic reproduction chain. Several DSP-based techniques for modal equalization have been proposed recently and tested in performance. From a perceptual point of view, however, no clear picture on the importance of controlled temporal decay has been shown, although it is known that towards the lowest frequencies the human hearing becomes increasingly insensitive to temporal details. In the present study we conducted listening tests where only a single synthetic mode with increased decay time but magnitude-equalized response was used to find the JND threshold of excessive decay time. The main conclusion is that at typical listening levels and down to 100 Hz the modal decay time T60 is allowed to increase from about 0.3 seconds by 0.1 to 0.4 seconds, while at 50 Hz even decay times of up to two seconds do not make a noticeable difference.

Andrew Goldberg and Aki Mäkivirta

November 2003, The Institute of Acoustics Reproduced Sound 19 Conference, Oxford, UK

This paper presents a novel method for robust automatic selection of the optimal in-situ acoustical frequency response within a discrete-valued set of responses offered by room response controls on active loudspeakers. The rationale of the room response controls for the active loudspeakers is described. The frequency response, calculated from the acquired impulse response, is used as the input for the optimisation algorithm to select the most favourable combination of room response controls. The optimisation algorithm is described and the performance of the algorithm is statistically analysed and discussed. It improves the acoustical similarity between loudspeakers in one space and performs robustly and systematically in widely varying acoustical environments. The efficiency and performance of the algorithm is discussed. The algorithm dramatically improves the speed of optimisation compared to an exhaustive search. This algorithm has been implemented and is currently in active use by specialist loudspeaker system calibrators who set up and tune studios and listening rooms.

Andrew Goldberg

September 2003, IBC Conference, Amsterdam, Netherlands

A modern audio production facility must be able to supply productions in a large number of different formats. The change from mono & stereo to multi-channel reproduction has created many problems, both in converting existing production facilities to multichannel format and when building new installations. This paper examines the LFE channel across different encoding formats and presents a method to reproduce it, using a bass management system, through a single monitoring system.

Andrew Goldberg and Aki Mäkivirta

May 2003, AES 23rd International Conference, Copenhagen, Denmark

This paper presents a novel method for automatically selecting the optimal in-situ acoustical frequency response of active loudspeakers within a discrete-valued set of responses offered by room response controls on active loudspeakers. The rationale of the room response controls for the active loudspeakers is explained. The frequency response, calculated from the acquired impulse response, is used as the input for the optimisation algorithm to select the most favourable combination of room response controls. The optimisation algorithm is described. The performance of the algorithm is analysed and discussed. This algorithm has been implemented and is currently in active use by specialist loudspeaker system calibrators who set up and tune studios and listening rooms.

Matti Karjalainen, Poju Antsalo, and Aki Mäkivirta

May 2003, AES 23rd International Conference, Copenhagen, Denmark

The low-frequency behavior of sound reproduction in listening rooms is often problematic due to long-ringing modes that are difficult and expensive to control by acoustic means. Modal equalization has been proposed recently to correct the low-frequency problems by shortening the decay times of problematic modes through modification of transfer function poles. While the previous methods were based on the estimation of isolated modes and their parameters, the new method proposed here is a technique to change the time-domain response more directly. It is an advanced windowing technique where the temporal shaping of a given impulse response can be done in a frequency-dependent manner. The method is compared with previous modal equalization techniques.

Poju Antsalo, Matti Karjalainen, Aki Mäkivirta and Vesa Välimäki

March 2003, AES 114th International Convention, Amsterdam, The Netherlands

Modal equalization of low-frequency room modes has recently been proposed as a method to improve sound reproduction in spaces where modal decay time is too long. Modal equalization is achieved by signal processing reducing the pole radii of problematic modes in the overall transfer function. In this paper, we compare the performance of two proposed methods for designing modal equalizers. Comparison includes a preliminary subjective listening test indicating a possible marginal improvement by modal equalization over conventional magnitude equalization.

Andrew Goldberg and Aki Mäkivirta

March 2003, AES 114th Convention, Amsterdam, Netherlands

This paper presents a novel method for robust automatic selection of optimal in-situ acoustical frequency response within a discrete-valued set of responses offered by room response controls on an active loudspeaker. A frequency response measurement is used as the input data for the algorithm. The rationale of the room response control system is described. The response controls are described for each supported loudspeaker type. The optimisation algorithm is described. Examples of the optimisation process are given. The efficiency and performance of the algorithm are discussed. The algorithm dramatically improves the speed of optimisation compared to an exhaustive search. It improves the acoustical similarity between loudspeakers in one space and performs robustly and systematically in widely varying acoustical environments. The algorithm is currently in active use by specialists who set up and tune studios and listening rooms.

Matti Karjalainen, Paulo A. A. Esquef, Poju Antsalo, Aki Mäkivirta, and Vesa Välimäki

December 2002, Journal of the Audio Engineering Society, Vol. 50, No. 12.

Discrete-time analysis and modeling of reverberant and resonating systems has many applications in audio and acoustics. The methodology of ARMA modeling by pole–zero filters for measured impulse responses was investigated. In addition to an overview of the standard AR and ARMA techniques, a spectral zooming technique is proposed, which is useful for resolving very closely positioned modes and high-density modal clusters.

Application cases related to the analysis and modeling of room responses, loudspeaker–room equalization, and the estimation of parameters for musical instrument modeling are studied.

Aki Mäkivirta and Jan Abildgaard Pedersen

July/August 2002, J. Audio Eng. Soc., Vol. 50, No. 7/8.

In part I of two papers the requirements for low-frequency sound reproduction were investigated by the variation of lower cutoff frequency and slope and by the introduction of different levels of amplitude ripple and group delay ripple in the passband of a high-performance sound reproduction system. Listening tests were performed at three different sound pressure levels using both loudspeakers in an anechoic chamber and headphones in an audiometric booth.

Two reproduction setups were used to confirm that equal results of the listening tests could be obtained in the two cases when proper equalization was implemented. It is described how DSP was used to generate stimuli and perform equalization of the two reproduction setups. The shape and magnitude of amplitude and group delay ripple were derived from room simulations of an IEC 268-13 sized room with varying reverberation time. Proper equalization included the introduction of head-related transfer functions in the signal path to the headphones. This ensured that the sound pressures at the ear drums were very similar in two cases: a person sitting in front of the loudspeakers in the anechoic chamber and a person wearing headphones in the experimental booth. Level calibration was performed on both setups using pink noise. The nonlinearities measured in the physical loudspeakers were introduced into the signal path to the headphones using a nonlinearity simulator program.

Matti Karjalainen, Paulo A. A. Esquef, Poju Antsalo, Aki Mäkivirta, and Vesa Välimäki

May 2002, AES 112th International Convention, Munich, Germany

Discrete-time analysis and modeling of reverberant and resonating systems has many applications in audio and acoustics. In a recent paper (AES110, Preprint 5290) we formulated techniques for the estimation of modal decay parameters from noisy response measurements, targeting to systems such as room reverberation and modal decay as well as musical instrument modeling. In this paper we extend the methodology to AR and ARMA modeling of measured responses by all-pole and pole-zero filters. In addition to an overview of standard techniques we propose a spectral zooming technique that is useful for resolving very closely positioned modes and high-density modal clusters.

Sensitivity to background noise is also studied. Application cases are taken from analysis and modeling of room responses, loudspeaker-room equalization, and estimation of parameters for musical instrument modeling.

Darren Rose

May 2002, AES 112th Convention, Munich, Germany

Modern commercially available, compact, low power audio power amplifiers are mostly designed around one of three main technologies. These are integrated circuit class AB, thick film hybrid class AB, and switch mode power amplifier modules. The decision to use a particular technology is not only based on idealised performance specifications, but also on the performance under realistic operating conditions, and cost-to-performance considerations. In this study, the performance of each amplifier technology is studied in ideal and realistic operating conditions with two amplifier designs for each technology category. Regulated and unregulated power supplies are used, in combination with ideal resistive and real-life complex impedance loudspeaker loads. For a fixed nominal supply voltage, the value of the different technologies with regard to noise, distortion and continuous output power is discussed. This results in an analysis of the cost effectiveness, or value, of currently competing technologies for high quality, low power, compact audio power amplifiers.

Aki Mäkivirta and Christophe Anet

September 2001, AES 111th Convention, New York, USA

The in-situ responses of a total of 372 loudspeakers in 164 professional monitoring rooms around the world have been measured after acoustical calibration. All measured rooms have been equipped with factory calibrated three way monitors and acoustically calibrated with standardized apparatus. The results provide a thorough understanding of typical monitoring conditions for stereo and multi-channel rooms, distribution in room parameters and quality of reproduced audio. Results are compared to current standards and recommendations.

Aki Mäkivirta, Poju Antsalo, Matti Karjalainen and Vesa Välimäki

September 2001, AES 111th Convention, New York, USA

In a room with strong low-frequency modes the control of excessively long decays is problematic or impossible with conventional passive means. In this paper we present a systematic methodology for active modal equalization able to correct the modal decay behavior of a loudspeaker-room system. Two methods of modal equalization are proposed. The first method modifies the primary sound such that modal decays are controlled. The second method uses separate primary and secondary radiators and controls modal decays with sound fed into the secon-dary radiator. Case studies of the first method of implementation are presented.

M. Karjalainen, P. Antsalo, A. Mäkivirta, T. Peltonen, V. Välimäki

May 2001, AES 110th Convention Amsterdam, The Netherlands

Estimation of modal decay parameters from noisy measurements of reverberant and resonating systems is a common problem in audio and acoustics, e.g., in room and concert hall measurements or musical instrument modeling. In this paper, reliable methods to estimate the initial response level, decay rate and noise floor level from noisy measurement data are studied and compared. A new method, based on nonlinear optimization of a model for exponential decay plus stationary noise floor, is presented. Comparison with traditional decay parameter estimation techniques using simulated measurement data shows that the proposed method outperforms in accuracy and robustness, especially in extreme SNR conditions. Three cases of practical applications of the method are demostrated.

Aki Mäkivirta and Jan Abildgaard Pedersen

September 2000, AES 109th Convention Los Angeles, USA

We present a method to change amplitude and delay responses in subjective listening test material independently of each other. This is necessary in subjective listening experiments to apply modern statistical methods treating simultaneously several statistical variables. A case study of producing audio test material with this method is presented. This is related to an experiment where the audibility of the amplitude response variation and the delay response variation are studied at low frequencies based on data obtained from impulse responses of a room.

Aki Mäkivirta and Jan Abildgaard Pedersen

February 2000, AES 108th International Convention, Paris, France

Lower cut-off frequency & slope, amplitude ripple and group delay ripple have been examined. Shape and magnitude of both amplitude and group delay ripple were calculated from room simulations. Listening tests are to be performed using both loudspeakers in an anechoic chamber and headphones, which were equalized in an attempt to obtain equal results.

A. Varla, A. Mäkivirta, I. Martikainen, M. Pilchner, R.Schoustal, C. Anet

April 1999, AES 16th International Conference Rovaniemi, Finland

This paper presents an overview of the practical aims, methods and problems of designing monitoring rooms for multichannel audio. The main problem areas encountered in practical multichannel audio monitoring room are described with a case study of a practical high quality monitoring room and with scale model measurements. The monitoring room studied does not follow entirely the ITU-R BS 775-1 international recommendation on speaker placement. The scale model is also used to investigate the effect of structural modifications to the room suggested in literature as methods to achieve better performance in multichannel audio reproduction. The design principles and methodology applied in the industry are reviewed, and some suggestions are developed to guide the monitoring room designer.

Ilpo Martikainen, Ari Varla, Topi Partanen

March 1989, AES 86th International Convention, Hamburg, Germany

This report describes the design of an active control room monitor for SPLs in excess of 130 dB (linear). To ensure good sound quality a new direct radiating midrange system was developed with an acoustic loading technique which allows 103 dB/W sensitivity and 1 kW peak power handling. Power amplifiers and driver unit protection logic are also presented.

Ilpo Martikainen, Ari Varla and Matti Otala

March 1983, AES 73rd International Convention, Eindhoven, The Netherlands

Based on an analysis of the equivalent circuit of a multiway loudspeaker, the possibility of large drive currents is predicted for a class of non-sinusoidal band- and amplitude-limited signals. The current builds up as coherent sum of two parts; charging of the driver reactance and simultaneous current drain by several drivers.

The input current of three commercial loudspeaker systems was measured using a signal derived from on the analysis. The results show that a loudspeaker may draw currents three to six times larger than those calculable from the rated speaker impedance. This indicates that certain generally accepted power amplifier design criteria should be reconsidered.

Ilpo Martikainen and Ari Varla

March 1982, AES 71st International Convention, Montreux, Switzerland

This experimental work deals with the loudspeaker-amplifier interface. The instantaneous input voItage/current relationships of loudspeakers are measured with low-pass filtered square wave at different levels. These preliminary measurement results allow us to assume that power amplifiers intended to drive loudspeakers should be designed to supply several times the current which could be expected directly from the system nominal impedance.

Ilpo Martikainen

February 1980, AES 65th International Convention, London, United Kingdom

A general design of monitoring loudspeakers for broadcasting use is presented. Requirements are desired from practical listening conditions. Amplifier integration, crossover shapes and driver bandwidths are discussed. These integrated systems with acoustic output powers from 20 mW to 1 W are synthesized and practical results presented.