Actuator-based sound reproduction is increasingly adopted in automotive audio systems due to its advantages in packaging efficiency, reduced mass, and seamless integration into vehicle structures. This work presents the implementation and tuning of a hybrid automotive sound system combining distributed mode loudspeakers (DMLs) with conventional dynamic drivers in a constrained open-cockpit vehicle platform developed by Morgan Motor Company. Two full-range actuators were integrated into a 2 mm aluminum dashboard panel to reproduce mid and high frequency content, supplemented by conventional woofers, a subwoofer, and low-frequency actuators. A comprehensive measurement methodology was employed, including spatially averaged frequency response measurements using an eight-microphone array and time-domain analysis via centralized impulse response capture. Measurements were conducted under multiple operating conditions, including stationary and driving scenarios with varying roof configurations. Digital signal processing, implemented using tools developed by Sennheiser, was applied to achieve frequency response linearization, time alignment, crossover integration, and final subjective tuning across the hybrid system. Additional tuning was guided by structured critical listening evaluations across representative use cases. Results indicate that actuator-based reproduction can provide improved spatial impression and apparent source width compared to conventional full-range door-mounted loudspeaker configurations, particularly in acoustically challenging environments characterized by high ambient noise. The findings demonstrate the viability of hybrid actuator-conventional systems and highlight the importance of panel behaviour, placement constraints, and robust tuning methodologies in achieving consistent performance.
Automotive OEMs invest significantly in branded audio systems to create differentiated, premium in-vehicle experiences. However, loudness tuning—critical for maintaining consistent perceived sound quality across listening levels—is often underdeveloped or inconsistently implemented. As a result, systems optimized at a single reference level can sound thin, unbalanced, or inconsistent under real-world usage conditions, negatively impacting perceived quality and brand identity. This paper presents a practical and production-ready methodology for loudness tuning based on ISO equal-loudness contours. Starting from a calibrated reference listening level, level-dependent compensation curves are derived to address changes in human auditory sensitivity at reduced playback levels. These compensations are implemented using a small set of efficient filters controlled via a volume-dependent lookup table, enabling scalable integration into automotive audio platforms. To further validate the approach, a benchmark dataset of over 200 production vehicles was analyzed, combining objective measurements with listener preference rankings. High-performing systems were used to derive effective loudness compensation characteristics, which shows a good alignment with ISO-based predictions while revealing systematic deviations due to in-cabin acoustics and playback conditions. The results demonstrate improved spectral consistency and preservation of brand sound signature, providing a high-impact, low-cost opportunity for OEMs to enhance perceived audio quality across all listening conditions.
Immersive object-based spatial audio is now firmly established in the music industry as the standard for production, distribution, and playback. The number of automobiles integrating such content to provide premium entertainment experiences is steadily increasing, driving the development of new audio rendering techniques. While loudspeakers integrated into automotive headrests have been around for more than 50 years, they have not yet achieved status as a standard feature in new cars. However, they represent a powerful tool for reproducing immersive audio by enabling the creation of personal sound zones with reduced passenger distraction while effectively complementing existing cabin speakers. We conduct subjective assessments using paired comparison experiments to measure preference and multiple spatial audio attributes. We plan to model the resulting probability outcomes using probabilistic choice models, such as Bradley-Terry-Luce rank ordering. We expect the findings of this work to provide a technical roadmap for the integration of headrest speakers in next-generation automotive spatial audio systems.
Standardized measurement of in-car audio systems remains an ongoing challenge, particularly regarding microphone configuration and placement, test signals, and correlation with perception. Current recommendations from the AES Technical Committee on Automotive Audio (TC-AA) advocate spatial averaging using multi-microphone arrays to improve repeatability. While this method reduces the influence of reflections and standing waves to improve frequency magnitude measurement consistency, it limits frequency, phase and amplitude resolution, which help identify the root causes of the distortions and evaluate their perceptual impact. An alternative approach is to apply frequency-normalized distortion analysis, where reflections or standing waves in the frequency response are also echoed in the distortion results, and eliminated by direct comparison of the two measurements. In this paper, spatial averaged distortion measurements using the TC-AA recommended 6 microphone array are compared with normalized distortion calculations, and a single microphone capture using the normalized distortion method. BSR measurements are also challenging. The TC-AA recommends a crest factor algorithm, available in most audio measurement systems. This adequately detects transient distortions, but is susceptible to background noise, therefore requires a tightly controlled environment for making measurements. Two other methods, enhanced perceptual Rub & Buzz and enhanced Loose Particles offer improved repeatability in the presence of background noise, and the results are easier to correlate to audibility. The three methods are compared, both in a quiet environment, and with background noise.
Sound sensing through structural vibrations has emerged as a robust alternative to conventional air conducted microphones for automotive exterior applications, where environmental exposure can degrade microphone performance. This study evaluates the feasibility and performance of sensing airborne speech via vehicle body surface vibrations using surface mounted accelerometers. Three miniature accelerometers with different signal to noise ratios were mounted at seven locations on a vehicle body to sense speech generated by an artificial mouth positioned outside the vehicle. Speech quality was objectively assessed using ETSI TS 103 281 perceptual metrics under various driving conditions. The results demonstrate that acceptable to good quality of speech can be reconstructed from body surface vibrations under relatively low noise conditions such as parked or idling states. Accelerometer SNR is identified as a key performance factor, with higher SNR sensors consistently yielding superior speech quality across all mounting locations and test conditions. Sensor mounting location also plays a significant role, particularly under elevated driving noise, with relatively flexible and noise isolated body panels providing better performance. In addition, increasing the speech level improves performance, consistent with the benefits associated with higher SNR accelerometers. Finally, the results indicate that introducing a high pass characteristic into the accelerometer frequency response does not provide a consistent performance benefit.
Microphones used in automotive hands-free and Automatic Speech Recognition (ASR) systems are typically required to meet wideband or fullband specifications defined by standards such as ITU-T P.1110 and P.1120. In practice, however, compliance with these standards is often challenged by vehicle cabin integration constraints and automotive-grade durability requirements. Moreover, there is limited empirical evidence clarifying how specific microphone characteristics influence perceptual audio quality and ASR performance. This paper presents an experimental investigation into the effects of microphone frequency response and bandwidth variations on system-level performance in automotive environments. Noise signals recorded under real-world driving conditions are used to evaluate perceptual speech quality using ETSI TS 103 281 metrics, including S-MOS, N-MOS, and G-MOS, as well as ASR accuracy quantified by Word Error Rate (WER). In addition, the interaction effects between microphone characteristics and signal processing algorithm is also studied by processing the signals through Noise Reduction (NR) modules. The results aim to identify the most critical frequency response attributes for properly determining microphone specifications in automotive hands-free and ASR applications.
This paper presents a vehicle-scale multichannel automotive audio demonstrator based on piezoelectrically actuated flat-panel loudspeakers (FPLs) integrated into existing cabin surfaces. The proposed architecture combines piezoelectric driven radiators embedded in headrests, front and rear doors, and an OLED center console with a conventional electrodynamic subwoofer for low-frequency extension, enabling full-band reproduction in a hybrid configuration. Operating frequency bands are assigned from electroacoustic characterization of each radiator and implemented through dedicated crossover design and DSP tuning. The system is analyzed through numerical simulation and evaluated experimentally in a prototype vehicle using AES automotive assessment procedures, objective electroacoustic indicators, and perceptual evaluation through Multi-Dimensional Audio Quality Score (MDAQS) metric. Results show that surface geometry, actuator placement, and time alignment improve in-cabin sound field distribution and directivity. After tuning, spectral balance improves and impulsive distortion artifacts are no longer evident, with corresponding enhancements in perceived timbre and overall quality. However, maximum output level remains constrained by the subtractive equalization required to control distortion and spectral balance. Overall, the study demonstrates the feasibility of hybrid piezoelectric--dynamic automotive systems, highlighting their packaging advantages and potential for reduced electrical power consumption while identifying the remaining limitations in low frequency capability, distortion performance, and achievable Sound Pressure Level (SPL).
A White Paper for in-car measurements version 1.0 was officially published in 2023 by the Automotive Audio section of the AES. It comprises basic characteristics of car audio systems. One of the intentions was to make audio systems as well as different cars comparable. This paper is divided into two parts: In the first part, measurements of several cars with different sizes, ages, and audio performance are presented based on version 1. The feasibility of the suggested procedure is evaluated, and improvements are suggested. The first version addresses frequency responses at different levels and at maximum sound pressure level, as well as defect symptoms such as rattling. In a second part, additional measurements are suggested that address compression and non-linear distortion. The suitability of different distortion measurements is discussed and practically demonstrated. Time variant effects, which are typical for DSP-based algorithms, such as limiters, protection systems, and adaptive control, are also investigated, and methods to characterize those are proposed. Stationary behavior is not ensured in such systems; thus, methods to cope with transient effects are required.
Automotive audio systems rely on complex signal-processing algorithms. Technologies such as noise reduction, acoustic echo cancellation, beamforming, and in-car communication processing must be meticulously tuned across vehicle platforms and their unique microphone and loudspeaker layouts. This optimization process requires constant evaluation of where and how specific algorithm parameters alter the output. Pinpointing these deviations remains a major challenge, as no existing tool is designed for such granular comparison. We introduce a comparison methodology that adapts the blink comparator — an instrument invented in astronomy over a century ago and used to discover Pluto — to the analysis of spectrograms of automotive audio signals. This technique rapidly alternates between aligned spectrograms of different processing variants in a flicker-free display, making differences in time, frequency, and intensity immediately visible, even when scalar metrics such as Echo Return Loss Enhancement or wideband Signal to Noise ratio improvement indicate similar performance. To further expose subtle changes that might otherwise remain hidden, the method incorporates spectral difference masking, which highlights local spectral deviations between processing variants. We also extend the classical visual blink comparator with synchronized audio playback switching, enabling the engineer to see and hear the difference simultaneously. We demonstrate its effectiveness through automotive case studies involving noise reduction tuning, echo cancellation artifacts, and beamformer comparisons.
We propose a machine-learning method that predicts trained-listener panel ratings of vehicle audio systems from in-situ microphone-array measurements. Using a dataset collected over more than 10 years from 1177 vehicle audio systems, the model is trained to predict listener-score distributions, not only mean scores, for subjective sound-quality attributes on a 1–10 scale. In this paper we focus on the timbral/spectral attribute and evaluate performance on a held-out set of 237 systems. Results show that prediction accuracy improves consistently with dataset size and that distribution-aware modelling captures not only expected score level but also listener disagreement and uncertainty. The study highlights both the feasibility of virtual listener panels and the importance of large-scale, reliable benchmarking data.
A method is described for measuring and comparing the perceived sound quality of vehicle audio systems in situ. The process has been refined through use at company facilities for over two decades and has been applied across a variety of internal use cases within the organization. Panels of experienced evaluators, aligned processes, and dedicated facilities are used at multiple sites around the globe to support consistent data collection. The collected data are benchmarked against a periodically refreshed reference population of recent vehicle evaluations, allowing each vehicle to be positioned relative to the contemporary competitive landscape. Hundreds of current vehicle models are available for comparison at any given time. The resulting information is used within the organization to support sound system target setting and validation, competitor analysis, sales pursuits, and product development and prototyping. More recently, the method has also been explored as a basis for sound-quality predictor modeling. This paper describes the overall method, the associated analysis workflow, and a set of anonymized case studies intended to illustrate practical comparability and discrimination in routine use.
A core challenge of acoustically optimizing an automotive sound system is realizing a balance between ideal and viable loudspeaker placement. Placement is largely influenced by structural, aesthetic and safety design considerations leading to non-ideal acoustic response characteristics within a given car cabin. These constrains often force drive units to be oriented off-axis from the listening region. To fully benefit from a transducer’s performance, the on-axis response is desirable at the listening location. This work presents the design and evaluation of design of broadband acoustic meta-material lenses to perform beam steering on an incoming wavefront. The parameters of the meta-material lens are learned via a differentiable acoustic simulation, evaluating the far-field directivity of the lens by solving the inhomogeneous Helmholtz equation via Fourier spectral methods coupled with a far field approximation of the boundary integral method. To realise a physical lens from the theoretical lens, a differential evolutionary algorithm is used to optimize the geometry of a parametric 2D model for each element of the metamaterial lens’s cross-section, matching the target effective density and bulk modulus via Finite Element Method (FEM) analysis. The performance of the whole lens is then evaluated via FEM analysis. After this, resin 3D printing is used to construct the metamaterial lens which is used to verify the results in real world.
The automotive industry is rapidly migrating from 12 V to 48 V electrical architectures to support growing power demands from compute and electrified subsystems. Audio amplifiers appear well positioned to benefit through reduced current, lower conductor losses, increased instantaneous power capability, and an additional 12 dB of voltage headroom for improved transient reproduction in non-boosted amplifier systems. However, directly scaling conventional Class D amplifier architectures and ICs to 48 V introduces significant and often underestimated challenges. Usage weighted analysis of real audio content shows that over 95% of playback occurs below 5 W output power, where conduction losses are minimal and voltage dependent switching losses dominate efficiency and thermal behavior. The resulting impacts on efficiency, EMI, noise floor, thermal design, and system architecture motivate a re examination of long standing assumptions for next generation 48 V automotive audio systems.
Engine order enhancement is central in automotive sound design, where selective harmonics are synthesized to shape perceptual qualities such as sportiness, refinedness, or power. This paper investigates a neural network-based approach to combustion engine sound modeling that extends conventional engine order analysis and enhancement by deriving synthesis parameters from audio data with machine learning and incorporating stochastic components into the synthesis framework. The system parameterizes engine sounds as a compact representation capturing per-order and broadband timbral variation across the full RPM-torque operating range, while remaining manually tunable and compatible with established automotive audio frameworks. The approach leverages gradient-based optimization and analysis-by-synthesis through an end-to-end differentiable implementation. The resulting synthesis parameter set is directly transferable to conventional DSP implementations for deployment on embedded targets. Spectral metrics and listening tests confirm high reconstruction fidelity, and integration into an established automotive audio framework EVx Suite demonstrates technical feasibility on deployment-ready embedded systems.
Automotive audio systems require precise spectral tuning to achieve consistent sound quality across vehicle models, trim levels, and seating positions. Cabin acoustics introduce strong resonances and seat-dependent responses, while manufacturing tolerances, loudspeaker variability, and interior materials create additional differences between vehicles. As modern infotainment platforms incorporate multi-speaker playback, active sound design, speech communication, and personalized audio processing, manual equalization becomes impractical for production environments. This paper presents an automatic parametric equalization approach for production calibration of automotive audio systems. The method estimates the minimal number and determines the parameters of a cascade of biquad filters such that the resulting frequency response matches a predefined target curve while satisfying constraints typical for embedded DSP implementations. The solution is formulated as a constrained non-linear optimization problem tailored for stable and efficient biquad coefficient estimation. The proposed approach enables automated tuning workflows in which measured acoustic responses are directly converted into equalization parameters without manual intervention. Applications include factory calibration, premium audio tuning, and speech communication optimization, resulting in improved repeatability, reduced tuning time, and consistent spectral performance across vehicles.
Tactile transducers, or "shakers", are increasingly being included within automotive seats, both as a safety feature for driver alerts, and as part of the in-vehicle sound system. High sound pressure level auditory experiences such as live sound events are often accompanied by tactile sensations, and so the inclusion of tactile excitation alongside the audio rendered by the vehicle's loudspeakers can enhance the listener experience. In audio-tactile systems, the micro-dynamic properties of the driving signals can be manipulated in order to enhance either transient or steady-state elements. This can be carried out both as part of the tuning of the system, or in order to cater to different user preferences. This study reports a subjective test where 50 subjects rated audio-tactile experiences with differing balance of transient and steady-state elements, in order to determine whether there is a relationship between user preference for transient-steady-state balance and musical genre. The results suggest that there are no discernible trends for user preference versus genre.
In automotive audio systems, musical beats and drum events can control synchronized in-cabin experiences such as ambient lighting and music-driven visual effects. Compared to beat tracking, automatic drum transcription (ADT) offers richer control signals by detecting and classifying drum onsets of multiple drum classes (e.g., kick, snare, and hi-hat), enabling more precise and musically meaningful synchronization. Deploying ADT in vehicles, however, requires low latency, computational efficiency, and robust performance for various input signals. This paper investigates improvements to low-latency ADT suitable for automotive deployment, using the Separate-Tracks-Annotate-Resynthesize Drums (STAR Drums) dataset and a block-based processing strategy that achieves an average detection delay of around 60 ms. We explore three strategies: (1) lightweight architecture modifications inspired by recent advances in image classification, combined with a temporal convolutional network (TCN); (2) re-rendering STAR Drums to increase drum timbre diversity and augmenting the re-synthesized drum stems; and (3) refinement training with pseudo labels obtained from source-separated mixtures. Our results show that data augmentation and increased drum timbre diversity yield modest performance gains, whereas pseudo-label refinement provides the largest effect, with up to 18 % relative improvement in global F-measure. In the real-time eight-class setting, our best model achieves a global F-measure of 0.76 on MDB Drums, competitive with state-of-the-art offline systems, demonstrating that accurate and efficient ADT is feasible for automotive deployment.
Signal processing using artificial intelligence (AI) has gained increasing interest because it outperforms existing solutions in many fields. A significant challenge for deep neural networks lies in meeting strict requirements regarding latency, computational load and memory, which is vital in automotive audio. This paper presents CUpGAN (Conditional Upmix GAN), a computationally efficient method for extracting upmix signals with low latency, leveraging signal separation for two upmixing concepts using a conditional generative adversarial network (CGAN). One upmix approach utilize spatial positions of direct sources within the stereo image, allowing for the distribu- tion of sources around the listener. The second approach separates direct and diffuse signals to create an ambience signal for rear surround loudspeakers. By employing phase-aware loss functions, integrating residual connections in the generator, and training with coherent input and target signals, we achieve high sound quality in the generated signals. This methodology also facilitates the computation of a cost-efficient complementary signal for both upmixing concepts through the difference between input and generated signals. The proposed technique reduces memory as 96% of the parameters can be shared between both applications, allowing seamless switching between upmixing approaches without the need for parameter loading; instead, parameters are computed by a small control network. The GAN generator is trained on synthetically generated data, enabling control over separation characteristics that surpass traditional methods. We present an evaluation using listening tests and computational metrics, demonstrating the advantages of our approach compared to classical signal processing methods.
As intelligent cockpits develop, music source separation (MSS) is increasingly used in automotive audio to address complex sound mixtures failing to meet users’ diverse needs. In karaoke, it separates vocals and accompaniment for humming and enables independent male/female vocal volume adjustment for duets. For in-vehicle audio up-mixing, extracted stems reconfigure stereo mixes into cockpit-optimized multichannel layouts. For real-time rendering, it is required to enhance specific tracks to adapt to cabin noise and user preferences. However, audio sources other than vocal, bass and drums lack research for real-time automotive applications. This paper proposes targeted optimizations for data augmentation and model structure: for target tracks (guitar, piano, male/female lead/backing vocals), a parallel single-track model is used for piano/guitar separation, and a two-stage model for male/female voice separation (first separating general vocals, then splitting into lead and backing vocals). A "Random Mixing" and "Aligned Mixing" combined method adapts to harmonic overlap in real songs. In terms of loss function, besides time-domain L1-loss and Multi-scale STFT loss, a GAN-based training procedure with individualized discriminators for each instrument stem improves audio quality and separation accuracy. Training uses a dataset from MedleyDB, MoiseDB and 3,000 private songs. To enhance the real-time causal model’s time-dimension receptive field, a modified SCNet with dilated convolutions and source-based band split is adopted. The model achieves 64 ms latency with 4.36M parameters, and its SDR values (6 dB piano, 5.6 dB guitar, 8.9 dB male vocals, 7.6 dB female vocals) outperform SOTA models like DTTNet and SCNet.
With the increasing adoption of electric vehicles (EVs), the number of vehicles equipped with active driving sound (ASD) systems has also grown. Although many EV driving sounds emulate the acoustic characteristics of internal combustion engine (ICE) vehicles, EVs lack engine induced vibrations, resulting in a mismatch between auditory cues and tactile seat sensations. Providing vibrations synchronized with virtual engine sound can mitigate this discrepancy and enhance driver immersion. This study analyzes the seat vibration characteristics of an ICE vehicle and proposes a vibration generation algorithm that integrates vehicle state information with already existing ASD sound. The algorithm was implemented on a DSP platform, and vibration actuators were installed in an EV seat to deliver tactile feedback in conjunction with ASD sound.
Engine orders are a major source of tonal noise in vehicle cabins. In production engine order cancellation (EOC), good performance is not only high attenuation, but also stable behavior in real conditions. After successful cancellation, the target order can become masked by broadband noise (low SNR). In addition, uncorrelated in-band intrusions may occur in the same frequency region. In these cases, observation-based on/off rules that rely on the current visibility of the tone can cause false deactivation and mode chattering. This paper presents a multi-channel EOC method for an in-cabin audio system. Orders are modeled in real time using a quadrature sin–cos basis synchronized with rotational-speed information. The main contribution is an order-level state logic that uses a reconstructed estimate of the current control contribution through secondary-path models. This enables a practical distinction between “absent/irrelevant” and “suppressed but masked,” and supports stable operation by scheduling the adaptation step size and the accumulation length in order periods. The method also accounts for fast RPM changes by shortening the accumulation window while keeping a moderate adaptation rate for tracking. A second contribution is a dynamic order-management layer that periodically evaluates residual-based order salience (accounting for current cancellation) and selects a bounded top-K set for control. The layer includes optional fade-in/out transitions for switching and a lightweight safeguard that de-rates poorly controllable orders to reduce the risk of noise amplification. The method is evaluated on a low-reverberation bench and in a vehicle under steady-speed and run-up conditions, including background noise and audio playback.
Active Road Noise Control (RNC) has become an important complement to passive treatments for mitigating low frequency tire–pavement noise in automotive cabins. While numerous global and local RNC systems have been reported in the literature and implemented in production vehicles, their performance is often evaluated using inconsistent indicators and measurement methodologies, making direct comparison difficult. This paper presents a critical review of performance evaluation metrics and measurement methods for automotive RNC systems. Key performance indicators, including sound pressure level reduction, effective frequency range, spatial effectiveness, adaptability, and robustness, are discussed with an emphasis on noise reduction measurements used in both industrial practice and academic research. Existing standards and test procedures are reviewed alongside commonly used experimental methods, such as single point microphone measurements, artificial head measurements, and small and large microphone array techniques. Reported noise reduction performance of representative global and local RNC systems is summarized to illustrate the influence of system architecture and measurement methodology on published results. The review highlights current limitations in RNC performance assessment, including test scenario complexity, measurement variability, and the lack of unified international standards. These findings provide guidance for selecting appropriate evaluation methods and support future development and standardization of automotive RNC system evaluation and measurement.
This research focuses on two of the main challenges for headrest loudspeakers in cars – their directivity and their low-frequency performance. In part one, we discuss a method to characterize the directivity in meaningful ways despite the complex acoustic target environment (acoustic near field conditions, presence of head and torso, car interior). Headrest speakers of different classes (open back, closed back, panel) are investigated with a nearfield scanning technique, and best practices for measurements are derived. In part two, we study how nonlinear adaptive control of transducers can improve the bass response and the overall quality of sound reproduction of headrest speakers, especially in typical applications relying on linear, time-invariant characteristics. The theoretical advantages of this approach have been discussed in earlier papers. Here, we will provide measurements to quantify the effect for different driver concepts (see part 1), discuss challenges and implications for concerned active sound algorithms such as individual sound zone control, hands-free communication, and active noise control. The results show that nonlinear adaptive loudspeaker control can considerably expand the usable frequency range while retaining robustness, reducing distortion, and providing a stable response even under varying environmental conditions.
AVAS design and optimization is a critical part of electric and hybrid vehicle development for regulatory purposes. Trial-and-error testing is often done to ensure compliance. However, this comes late in the vehicle design phase, take time, effort, and specialized test equipment and facilities, and is subject to testing variance which may overestimate or underestimate AVAS design viability. Increasingly, vehicle manufacturers and suppliers look to support AVAS design using simulation techniques. These can indicate during early design phases the sound levels at the AVAS certification measurement points from a given transducer and the sensitivity to different locations and angles of orientation. In the same simulation, the sound pressure on the vehicle glasses and body panels can be predicted, which can be combined with other simulation methods to predict how much noise from the AVAS transducer is transmitted to the interior and perceived by the vehicle occupants. In combination with objective optimization of transducer positioning for AVAS with acoustic transfer function simulation, subjective evaluation may also be carried out early in the design phase by combining candidate or measured AVAS transducer signals and inputs with the simulated acoustic transfer functions, which act as a filter from the output at the transducer to yield the virtual expected result at the AVAS measurement positions. This allows psychoacoustic evaluation of the expected sound at the AVAS measurement positions.
