Sound Quality Metrics Research Articles

Investigation of perceived annoyance in indoor workplaces using short and ultra-short sound samples

Typically, metrics crucial for assessing sound quality are derived from high-quality audio samples obtained in controlled environments. However, due to the sensitive nature of workplace settings, recording audio samples on-site becomes impractical. Privacy emerges as a significant concern for all when collecting audio samples. Consequently, best practice would require that all analyses must occur either locally on a device used to record audio or in the cloud, utilizing a purely mathematical approach. The paper aims to formulate methodologies for determining psycho-acoustic sound quality metrics without the need to analyze audio data directly. This can be achieved by collecting and analyzing data in small batches. The variation of these metrics over these smaller samples can serve as an indicator of sound quality. Initiating the determination of values for psycho-acoustic sound quality metrics such as loudness (N), sharpness (S), roughness (R) and fluctuation strength (FS) in short and ultra-short batches, followed by monitoring their variation over an extended period (e.g., 5 minutes), provides additional insight into the sound characteristics of the space. Using the obtained psycho-acoustic sound quality metrics, we will explore the perceived annoyance among occupants through indoor soundscape assessments.

ECMA-418-2 roughness, a challenging implementation

Over the past decade, psychoacoustics has grown significantly and sound quality metrics have become an easy way to include the human factor in NVH analysis. Most of these metrics have been defined by standards (ISO, DIN, ANSI or ECMA), which should promote their use in industrial applications. MOSQITO is an open-source project providing implementations of sound quality metrics in Python(e.g. loudness ISO 532-1, sharpness DIN 45692, SII ANSI S3.5) along with complete documentations and validation procedures. In 2020, the European Computer Manufacturers Association released the ECMA-418-2 standard, revised in 2022, describing new loudness, roughness and tonality implementations based on a common hearing model. During the implementation of the roughness metric in MOSQITO, it appeared that some implementation instructions are subject to interpretation, and some information was missing in the standard to allow for a valid computation of this metric. Moreover, the standard does not define some validation cases, which can limit the spread of its metrics. This paper highlights these difficulties for each section of the calculation process, and proposes new steps that would aid the user's understanding and implementation of the standard.

Sound prediction of an electric car using numerical simulation and sound synthesis techniques

The automotive industry's transition to electrification raises new challenges to maintain a comfortable acoustics environment both inside and outside the car. This paper presents an innovative way to make an early prediction of the overall sound of a vehicle at driver's ear by mixing noise sources that can be simulated and noise sources measured at component level. This hybrid approach combining Test and Simulation data helps define acoustic and vibration specifications at component level and take decisions on design strategies. For instance, the mounting method of the Electric motor to the body can be decided using a combination of Electric motor numerical simulation and Rolling noise sound synthesis techniques. Realistic signals could be generated and compared with sound quality metrics to choose the best vibration isolation strategy.

Sound quality characteristics of reciprocating hermetic refrigeration compressors

In recent years, sound quality (SQ) has become an increasingly important factor in the home appliance industry, enhancing the customer experience and comfort. Manufacturers that prioritize SQ in their designs can gain a competitive advantage in the market and increase customer satisfaction. The compressor is one of the main sources of noise in a refrigerator. While manufacturers typically provide refrigerator sound quality and noise labels, research that specifically evaluates the compressor sound quality metrics at a component level is not common. Therefore, this study aims to isolate the compressor from the refrigerator to investigate SQ metrics for hermetic refrigeration compressors. To achieve this, compressors with different capacities are operated and recorded at different working speeds for different thermal conditions. Objective parameters are calculated and tested using time domain signals and analyzed with spectral analysis methods. In addition to the objective parameters, subjective results are obtained through jury tests consisting of 75 people of different genders and age ranges. By considering objective parameters of sound quality as input and sound quality evaluation values as output, a regression model is established. It is found that Loudness, Sharpness, and Degree of Modulation play an important role in compressor sound quality evaluations consider.

Whistle or squeal? Distinguish and evaluate two typical abnormal sounds of collaborative robots

The whistle and squeal are two typical abnormal sounds among the radiated noise of running collaborative robots, indicating potential faults and causing discomfort through auditory perception. This study investigated whether psychoacoustic parameters for sound quality could distinguish the whistle and squeal and evaluate their discomfort. A total of 20 abnormal sound stimuli, including 6 whistles and 14 squeals, were obtained from various collaborative robots. The cepstrum characteristics and sound quality metrics (i.e., roughness, sharpness, fluctuation strength and tonality) of all stimuli were analyzed by controlling the equal loudness. Twenty-four participants rated the 'discomfort of abnormal sound in the stimuli' on an 11-point scale. The results show that the modulation frequencies of the fundamental frequency of the whistles were 2.9 Hz to 5.5 Hz, whereas the squeals had no obvious modulation characteristics. The modulation frequencies of the fundamental frequency were highly correlated (with a correlation coefficient of 0.89) to the discomfort responses for whistles, whereas tonality was significantly correlated (with a correlation coefficient of 0.82) to the discomfort responses for squeals. The linear models using single dependent variables of modulation frequency in cepstrum and tonality had determination coefficients of 0.77 and 0.66, respectively, in evaluating the discomfort of whistles and squeals.

Aerodynamic and aeroacoustic design of electric ducted fans

This paper describes the aerodynamic and aeroacoustic design of an electric ducted fan (EDF). A prototype EDF was designed and built with three different rotor-stator sets to operate at design flow coefficients ϕd=[0.60,0.75,0.90]. Computational Fluid Dynamics (CFD) simulations show that changing the design flow coefficient changes the proportions of endwall and blade profile loss, with the ϕd=0.75 design providing the optimum balance. CFD predictions of aerodynamic performance are compared with experimental measurements and show good agreement. Data collected from full annulus Unsteady Reynolds-Averaged Navier Stokes (URANS) CFD simulations is used to predict tonal noise radiation using a numerical solution to Farassat's Formulation 1A, and compare favourably to acoustic measurements in an anechoic chamber. All three EDF iterations show similar levels of broadband noise radiation, although tonal noise radiation decreases with increasing design flow coefficient and results in more favourable Sound Quality Metrics performance. Design flow coefficient is shown to have a first-order impact on both the aerodynamic and aeroacoustic behaviour of the EDF, and it is proposed that an optimal design, encompassing both aerodynamic performance and reduced acoustic impact, can be achieved at design flow coefficients between ϕd=0.75 and ϕd=0.90.

EEG measurements-based study for evaluating acoustic human perception: A pilot study

Sound quality analysis and sound design are well-known human-centered strategies to evaluate the subjective perception of noise and to design machines and environments with pleasant and comfortable acoustic signatures. The subjective acoustic perception is conventionally measured by means of sound quality metrics determined through a correlation process with jury test results. The exploitation of electroencephalogram (EEG) measurements during the jury test for the registration of the brain activity in response to the acoustic stimuli presented to the jurors can allow us to estimate the jurors’ perception directly from their physiological response. This study presents results from the application of an EEG wearable device to investigate changes in the EEG frequency domain at different acoustic stimuli. Forty-three participants were recruited, and the EEG signals were recorded using the wearable sensor. The analysis of power spectral densities (PSDs) was performed to investigate features correlated to acoustic sensation induced by audio stimuli. Statistically significant differences were found between the three audio stimuli. The results bring to the conclusion that wearable sensors could be used for EEG acquisition applied to acoustic perception evaluation.

Perception of noise from unmanned aircraft systems: Efficacy of metrics for indoor and outdoor listener positions.

This paper presents the results of a listening experiment designed to assess annoyance and perceived loudness (PL) for several unmanned aircraft system (UAS) operations, with the listener simulated in indoor and outdoor positions. This research investigated (i) how participant responses change depending on UAS operation, (ii) which broadband metrics are most suitable for representing annoyance and PL, (iii) differences in noise level required to result in equal participant responses to different operations, and (iv) which sound quality metrics (SQMs) are significant for UAS noise perception. Results indicate annoyance and PL responses were greatest for landing operations with flyovers being the least annoying or loud. LAeq, LASmax, and loudness (N5) were the strongest predictors in representing annoyance. Offset analysis predicted small differences in annoyance responses between flyovers and other operations, but also indicated that flyovers would require an increase to LASmax of 3.3 to 6.3 dB compared to other operations to achieve equal PL. Loudness was the most significant SQM, with minor contributions from impulsivity for annoyance and PL when outside, and tonality for PL when indoors. These findings contribute to the understanding of UAS noise perception for the development of metrics and assessment methods accounting for the characteristics of UAS operations.

Contextual event-based sound quality metrics

Sound quality metrics are often used to analyze complex sound scenarios, such as soundscape applications. Sound quality can also affect the health and well-being of people in a given environment. Therefore, it is of the utmost importance that the definition of good sound quality in a particular context is as precise as possible. In this aspect, psychoacoustic indicators such as maximum, average, or percentile values of psychoacoustic parameters are usually used in a linear combination to create a metric based on listening experiment results. However, this linear approach may disregard many complex nonlinear parameter dependencies that could be helpful for a better description of human perception. Instead of using summary statistics, the importance to perception of individual events in terms of their duration, magnitude, and frequency should be considered. For example, individual tonal events of a certain duration may have a strong influence on the perceived quality, even with low average tonality and percentile values. And the low-frequency tonal content of music may be perceived as an important aspect of a high-quality sound event, whereas the same amount of low-frequency tonality would be not accepted for a technical sound. Interview evidence from people immersed in various soundscapes appears to indicate significant "contextual weightings" involving both time and frequency, which might be applied to or evolved from current psychoacoustic metrics according to contextual classes. In this paper we provide an overview and some practical guidance for developing contextual event-based sound quality metrics.

Motor and bearing multi-fault diagnosis with sound quality metrics

The occurrence of multi-faults is more critical as compared to single faults, and it creates unfavourable working conditions. The number of contact-based sensors and their placement plays a crucial role in diagnosing the multi-faults. The use of non-contact-based acoustic sensors could be a cost-effective potential solution to this problem. In this work, two rotating components, a motor and a separate bearing of a system situated away from the motor, were considered for forming system-level multi-fault conditions. The acoustic signature of various multi-fault conditions was acquired at different rotational speeds. Thereafter, a moving window was employed on the acquired raw acoustic signals to extract statistical and sound quality features. A feature fusion approach was implemented on the extracted features, and a machine learning model was trained with optimised hyperparameters. A comparative study was performed to analyse the model's performance, which was trained with the extracted statistical, sound quality and fused features. It was found that the proposed feature fusion-based methodology was helpful for diagnosing multi-faults with acoustic signatures. The importance of psychoacoustical features on model performance was also observed. It mitigated the complexity of mounting the sensors and showed the potential usage of sound quality metrics for multi-fault diagnosis.

Computational model for predicting sound quality metrics using loudness model based on gammatone/gammachirp auditory filterbank and its applications

Sound quality metrics (SQMs), such as sharpness, roughness, and fluctuation strength, have been utilized to study the perceived pleasantness or unpleasantness of sounds. Computational models for predicting SQMs have been proposed, commonly based on Zwicker's loudness method (ISO 532B:1975). ISO 532 was revised in 2017; Moore & Glasberg's loudness method for stationary sound was adopted in ISO 532-2:2017, while Zwicker's loudness method for stationary and time-varying sound was revised in ISO 532-1:2017. There are a few important differenterences between the two methods, such as the frequency scale (Bark scale or ERB scale) and filter shape (symmetry or asymmetry). It therefore might be necessary to revise these computational models for predicting SQMs on the basis of these revisions. We previously proposed a time-domain computational model for predicting SQMs on the basis of a loudness method using the time-domain auditory filterbank. This loudness method can be regarded as a time-domain version of Moore & Glasberg's method in ISO 532-2:2017, using gammatone/gammachirp auditory filterbanks instead of the roex auditory filterbank, which does not have any time-domain representations. In this paper, we introduce calculations for three SQMs (sharpness, roughness, and fluctuation strength) using the proposed model for predicting SQMs.

Discomfort estimation for aircraft cabin noise using linear regression and modified psychoacoustic annoyance approaches.

Appropriate sound quality models for noise-induced discomfort are necessary for a better acoustic comfort design in the aircraft cabin. This study investigates the acoustic discomfort in two large passenger aeroplanes (i.e., planes A and B). We recorded the noise at 21 positions in each aircraft cabin and selected 42 stimuli ranging from 72 to 81 dB(A) during the cruising flights. Twenty-four participants rated the noise discomfort by the absolute magnitude estimation method. The discomfort values in the middle section of the aircraft cabin are 10% points higher than in the front or rear section. The discomfort magnitude was dominated by loudness and influenced by roughness and sharpness. A multiple linear (MA) discomfort model was established, accounting for the relationship between the discomfort and sound quality metrics (i.e., loudness, sharpness, and roughness). The MA model estimated noise discomfort better than the Zwicker and other (i.e., More and Di) psychoacoustic annoyance (PA) models. We modified the coefficients of independent variables in the formulations of Zwicker, Di, and More PA models, respectively, according to the present experimental results. The correlation coefficients between the estimated and measured values of the modified models were at least 20% points higher than the original ones.

Acoustic and psychoacoustic characterisation of small-scale contra-rotating propellers

This paper reports an experimental investigation of noise radiation and psychoacoustic evaluation of contra-rotating propellers with different blade counts over multiple emission angles and thrust settings. A series of aeroacoustic and psychoacoustic tests were performed with a contra-rotating propeller rig at the University of Salford, UK, where both propellers’ blade count and their axial separation z were varied at static conditions and constant thrust. The sounds were analysed using conventional acoustic metrics, such as Sound Pressure Level (SPL) and Sound Power Level for tonal, broadband, and overall noise, and using the psychoacoustic Sound Quality Metrics (SQMs) Loudness, Fluctuation Strength, Roughness, Tonality, and Impulsiveness. Configurations with axial separation-to-diameter ratio z/D between 0.2–0.4 exhibited a balance between tonal and broadband noise generation mechanisms, resulting in reduced noise levels and reduced annoyance for on-axis observers. The number of blades played a secondary role on determining noise levels: higher blade counts resulted in lower overall noise levels at larger axial separations, although at shorter distances (z/D<0.2) increasing the blade count led to higher noise levels. The SQMs Loudness, Fluctuation Strength, and Tonality were found to also reach a minimum at z/D=0.3, although increasing the number of propeller blades led to a decrease in Loudness but not in Fluctuation Strength and Tonality. The dominant noise generation mechanisms observed here present a directivity null perpendicular to the propellers’ axis, and no clear trends were observed for acoustic and psychoacoustic metrics at this emission angle. A listening experiment was set up to validate the results of the acoustic and psychoacoustic analysis, and its results suggest that the reported noise annoyance is driven by Loudness, Fluctuation Strength and Tonality metrics (p<0.01) at on-axis radiation, while at 90° emission angle, Loudness (p<0.01) and Tonality (p<0.05) are the main contributors to noise annoyance.

Development of a Sound Quality Evaluation Model Based on an Optimal Analytic Wavelet Transform and an Artificial Neural Network

The purpose of this study was to develop a sound quality model for real time active sound quality control systems. The model is based on an optimal analytic wavelet transform (OAWT) used along with a back propagation neural network (BPNN) in which the initial weights and thresholds are determined by particle swarm optimisation (PSO). In the model the input signal is decomposed into 24 critical bands to extract a feature matrix, based on energy, mean, and standard deviation indices of the sub signal scalogram obtained by OAWT. The feature matrix is fed into the neural network input to determine the psychoacoustic parameters used for sound quality evaluation. The results of the study show that the present model is in good agreement with psychoacoustic models of sound quality metrics and enables evaluation of the quality of sound at a lower computational cost than the existing models.

Sustainable Soundscape Monitoring of Modified Psycho-Acoustic Annoyance Model with Edge Computing for 5G IoT Systems

Next-generation IoT systems will allow sustainable performance in long-term monitoring systems. This sustainability concept applies to soundscape description, as it allows monitoring in urban environments. In this work, the implementation of psycho-acoustic annoyance models in a 5G-enabled IoT system is proposed, applying two edge-computing approaches. A modified Zwicker’s model is adopted in this research, introducing a term that takes into account the tonal component of the captured sound. These implementations have been validated in a measurement campaign where several IoT devices have been deployed to evaluate different sound environments of a university campus. Then, the analysis of the sound-quality metrics is conducted in a different location, showing that if tonality is present in a noisy environment, it results in greater subjective annoyance. Moreover, the Just-Noticeable Difference of these results is derived from Zwicker’s psycho-acoustic annoyance to establish a limitation for this metric.

A Psychoacoustic Test for Urban Air Mobility Vehicle Sound Quality

This paper describes a psychoacoustic test in the Exterior Effects Room (EER) at the NASA Langley Research Center. The test investigated the degree to which sound quality metrics (sharpness, tonality, etc.) are predictive of annoyance to notional sounds of Urban Air Mobility (UAM) vehicles (e.g., air taxis). A suite of 136 unique (4.6 second duration) UAM rotor noise stimuli was generated. These stimuli were based on aeroacoustic predictions of a NASA reference UAM quadrotor aircraft under two flight conditions. The synthesizer changed rotor noise parameters such as the blade passage frequency, the relative level of broadband self-noise, and the relative level of tonal motor noise. With loudness constant, the synthesis parameters impacted sound quality in a way that created a spread of predictors both in synthesizer parameters and in sound quality metrics. Forty subjects listened to the suite of UAM noise stimuli in the EER and judged each sound individually on a standard scale of annoyance. Additionally, a subset of the UAM noise stimuli were compared to a reference sound that varied in loudness. From these responses, the relative effect of changes in loudness or changes in other sound quality metrics on annoyance was evaluated. This paper covers background and motivation for the test, details of how the sound stimuli were generated, and details of the test design and execution. Test results investigate how sound quality may affect perceived annoyance to UAM vehicle noise, indicating the importance of sharpness, tonality, impulsiveness, and roughness on annoyance to UAM noise.

Tonality Perception Lessons Learned

The future is expected to bring Advanced Air Mobility (AAM) vehicles, including small unmanned aerial systems (sUAS), urban air mobility (UAM) vehicles and regional air mobility (RAM) vehicles. These manned and unmanned vehicles are propelled by rotors. Rotors tend to generate tonal sound as their blades interact periodically with airflow features. Since people are more sensitive to tonality, including tones, than broad band sound, AAM generated tonality is expected to be an important consideration for design. In this paper several tonality metrics are examined for their ability to explain perceived annoyance of AAM flyover noise as measured by NASA's Rotorcraft Sound Quality Metric 1 (RoQM-1) test. The various investigated metrics use one-third octave band, narrow band, and autocorrelation analysis. It is observed that tonality influences but does not control perceived flyover noise annoyance due to other sound qualities like roughness, consistent with previous work. The metrics are also examined for their ability to explain perceived tonality of sounds generated by IT equipment. The metrics based on autocorrelation are observed to best explain perceived tonality while also being among the best for explaining flyover annoyance. The plausibility of the auditory nervous system as a physiological auto-correlator is discussed.

Multitarget prediction and optimization of pure electric vehicle tire/road airborne noise sound quality based on a knowledge- and data-driven method

With the promotion of pure electric vehicles (PEVs), the overall interior noise level has been gradually reduced. Tire/road noise is increasingly becoming noticeable in PEVs and represents a primary concern for passengers. Vehicle acoustic packages are crucial for suppressing tire/road noise, and numerous studies have focused on improving the acoustic package performance and sound quality of tire/road noise. However, the prediction and optimization of tire/road acoustic comfort have two deficiencies: (1) The characteristics and transfer paths of tire/road noise are complex. Using knowledge-driven methods (such as simulation models) and data-driven methods (such as neural networks) for analysis has difficulty in accuracy parameter acquisition and unexplainable models. (2) The sound quality of tire/road noise is multidimensional. In the development of multitarget prediction models, when multiple targets contradict each other, prediction bias may result from wide differences among targets in the gradient value of the loss function in training. Therefore, in this paper, a knowledge- and data-driven method is proposed, which introduces the knowledge graph technique to develop a vehicle tire/road noise knowledge graph architecture and uses an improved residual network to drive reasoning in the domain knowledge graph. In addition, a multitarget prediction method based on the adaptive balanced learning mechanism for a residual network is proposed, which uses the dynamic weighted average method to adaptively adjust the loss weight of each target according to the convergence speed and learning difficulty of each target. The proposed dual-drive method is applied to predict and optimize the performance of vehicle acoustic package and to further improve the multiple sound quality metrics of tire/road noise. In the experimental validation, the proposed method outperforms the traditional prediction and optimization methods in effectiveness and robustness.

Evaluation of tone annoyance metrics on reverberation room measurements of HVAC equipment

Reverberation rooms are a primary tool for HVAC equipment manufacturers to calculate equipment sound power levels, which lead directly to equipment sound ratings. While sound quality has been included in standards, namely, the withdrawn AHRI 1140, the adoption and value of those standards has been minimal leaving specifiers to construct their own tone quality metrics. This paper will evaluate the application of common sound quality metrics, specifically tonal metrics, to reverberation room measurements. Room response is a critical factor in measuring tone prominence, especially at lower frequencies where adjacent frequency tones can vary by 6dB or more. Metrics evaluated include TTNR, Prominence, Tonality and a new metric developed in ASHRAE 1707-TRP.

Sound quality metric indicators of rotorcraft noise annoyance using multilevel analysis.

It is hypothesized that sound quality metrics, particularly loudness, sharpness, tonality, impulsiveness, fluctuation strength, and roughness, could all be possible indicators of the reported annoyance to helicopter noise. To test this hypothesis, a psychoacoustic test was conducted in which subjects rated their annoyance levels to synthesized helicopter sounds. After controlling for loudness, a previous analysis using linear regression identified sharpness and tonality as important factors in predicting annoyance, followed by fluctuation strength. The current work focuses on multilevel regression techniques in which the regression slopes and intercepts are assumed to take on normal distributions across subjects. The importance of each metric is evaluated, and the variation of regression parameters among subjects is evaluated using simple models. Then more complete models are investigated, which include the combination of selected metrics and subject-specific effects. While the conclusions from linear regression analysis are affirmed by multilevel analysis, other important effects emerge. In particular, subject-specific intercepts are shown to be more important than subject-specific slopes. In addition, subject-specific slopes for fluctuation strength and sharpness are more important than for tonality. Using a multilevel framework, the relative importance of sound quality metrics is reexamined, and the potential for modeling human annoyance to helicopter noise based on sound quality metrics is explored.