Active Sound Design entails the generation of acoustic signals which could be for vehicle sound enhancement, safety regulation, or both. With the advent and growing presence of electric vehicles on the roadway, new minimum noise requirements for EVs have been established in the name of pedestrian safety. Additionally, for the interior occupants there is a distinct auditory void left by the removal of the internal combustion engine. This missing sound, and the auditory space it leaves provides sound designers an opportunity to create a unique sonic experience for the driver and occupants, as well as craft a brand sound heard by those outside the vehicle.
Simcenter Testlab Sound Designer is an end-to-end tool for creating, tuning, and executing a rich, compelling, and lively soundscape for vehicle occupants as well as external pedestrians. This article will introduce the main capabilities of Simcenter Testlab Sound Designer, provide an overview of the underlying technologies, as well as provide an overview of the associated hardware.
This article has the following sections:
1. Introduction
1.1. Terminology/Glossary 1.2. What is Active Sound Design 1.3. Why add sound to a vehicle? 1.4. Interior vs. Exterior (AVAS) sound design
2. How to make an active sound
2.1. Order synthesis 2.2. Granular synthesis 2.3. Pitched Playback
3. How to make an active sound exciting & pleasing
3.1. Multi-layer approach
3.1.1. Base layer 3.1.2. Torque layer 3.1.3. Shepard layer
3.2. Sound parameterized to vehicle data 3.3. Live synthesis
4. How to validate an active sound in the vehicle
4.1. Vehicle Unit
4.1.1. Synthesis engine from desk to vehicle 4.1.2. Vehicle interface for live tuning 4.1.3. On-board storage
4.2. CAN Gateway 4.3. AVAS Unit
5. How to validate an AVAS design
5.1. Global AVAS regulations summary 5.2. Assess AVAS speaker mounting locations 5.3. Predict AVAS sound design performance to regulations
5.4. Example: Assessing a AVAS sound design for FMVSS141
1. Introduction
1.1. Terminology/Glossary
AVASAcoustic Vehicle Alerting System – Official name for systems put in place to meet exterior minimum noise requirements for electric & hybrid vehicles. EVElectric vehicle – Can also refer to hybrid vehicles which feature an all-electric mode of operation. ICEInternal Combustion Engine – Traditional gasoline or diesel fueled engine. FMVSSFederal Motor Vehicle Safety Standards – set of regulations issued by the National Highway Traffic Safety Administration (NHTSA) for automobiles sold in the United States of America. ECEEconomic Commission for Europe – governing body aimed at creating a uniform system of regulations relating to vehicle design in Europe. Comparable to NHTSA in the US (United States). CANCAN-bus – network for communication between various control modules in an automobile. Vehicle operational data is broadcast over CAN and describes vehicle operating state: speed, load, gear, throttle position, etc. Shepard toneAn auditory illusion created by the superposition of sine tones separated by octaves. As the pitch is changed, it gives the impression that the tone is continually ascending (or descending) though it ultimately never gets any higher or lower.
1.2. What is Active Sound Design Active Sound Design (ASD) is an umbrella term used to describe the development and implementation of sound purposefully generated for an automobile. It is “active” because it is sound created for its own sake, and not sound that is a byproduct of propulsion or other support systems, such as ancillary pumps or cooling fans. ASD consists of active noise cancellation techniques, as well as sound enhancement. This article will focus on active sound enhancement, whereby synthesized sound is added to the vehicle’s normal operational soundscape via the vehicle’s audio system. This added sound can be implemented for the interior occupants, as well as broadcast to the exterior for broadcasting brand sound or to meet minimum noise requirements for pedestrian safety.
1.3. Why add sound to a vehicle? Firstly, because it is required. Electric vehicles do not benefit from the noise generated by the traditional internal combustion engine (ICE) and can be nearly silent at low speeds. Pedestrians are often made aware of an ICE vehicle’s presence, approach, or retreat by the nature of the sound they hear. This sound is a mixture of engine and tire noise, and at low speeds is dominated by the noise of the engine. The lack of engine noise in EVs presents a safety concern for pedestrians. As a result, governing bodies around the world have begun requiring EVs to generate artificial noise to make their presence known to those around them. New external minimum noise regulations have necessitated the implementation of AVAS (acoustic vehicle alerting system) solutions to fulfill these requirements. Additionally, the sound character of a combustion engine has been part of the driving experience since the very beginning; it is a core part of a driver’s connection to the vehicle. With the electrification of the automobile, most, if not all this familiar sound character has been eliminated. Active sound enhancement allows vehicle manufacturers to replace this missing sound scape with a carefully curated sonic experience, reinvigorating the driving experience with emotion and connection for the driver and occupants.
1.4. Interior vs. Exterior (AVAS) sound design A vehicle’s sound signature has traditionally been a byproduct of the vehicle’s design: the engine design and layout, calibration, and exhaust and induction tuning, among other factors. The sound that reached the ears of occupants and passers-by communicated something about the vehicle and the brand. Active Sound Design represents the future of crafting this experience for the EV market. ASD is primarily divided into two categories: interior sound enhancement and exterior AVAS sound (Figure 1).
Figure 1: Active Sound Design encompasses both interior and exterior sound environments.
Exterior AVAS sound design aims first to satisfy the government regulations for EVs around the world to ensure adequate pedestrian safety. The AVAS sound must adequately inform pedestrians of the vehicle’s proximity, acceleration, and direction of travel. Visually impaired pedestrians are especially reliant on their sense of hearing to determine if a vehicle is approaching, or if it is safe to cross the street. However, in addition to providing for pedestrian safety, AVAS sounds are an important part of a vehicle’s brand and sound signature. The AVAS sound is heard by everyone around the vehicle and is communicating something about the vehicle and the brand whenever it is active. AVAS sound design is therefore about more than just meeting regulatory requirements, it is a branding opportunity as well.
Interior sound enhancement is about connecting emotionally with the driver and passengers. It is an opportunity to convey the vehicle’s power, refinement, and quality through sound. In addition to providing auditory feedback to the driver’s input, it can also provide some masking noise to make annoying and undesirable noises less perceptible. Interior sound design is the acoustic backdrop for the driving experience.
There can also be interplay between interior and exterior sound designs, and the relationship between them must be understood and accounted for.
2. How to make an active sound There are several methods available to generate an active sound design. This section will introduce a few of the techniques available in Simcenter Testlab Sound Designer.
2.1. Order synthesis Perhaps the most traditional method of sound enhancement is to synthesize order content. Rotating machines such as the internal combustion engine primarily produce order-based pressure pulses, which gives them their characteristic sound. Orders are sinusoids that change frequency in relation to the rotating speed of the machine (Figure 2). See the article What is an order?for more information on orders and where they come from.
Figure 2. Spectrum map of an internal combustion engine over a speed sweep. The diagonal lines represent the orders generated by the engine.
Order synthesis consists of generating multiple sine waves at various frequencies, amplitudes, and phases, and changing the frequency of these sine waves with respect to vehicle speed (or other parameter). This technique is useful for mimicking or enhancing the order content of an internal combustion engine. A map of synthesized orders is shown in Figure 3 below.
Figure 3. Order synthesis map showing 48 half-orders (0.5, 1.0, 1.5, etc). Amplitude is shown in color.
Just as the sound and order content produced by a real engine will change when the load increases, it is important to synthesize different orders at different load levels to preserve a realistic sound using order synthesis. Simcenter Testlab Sound Designer allows designers to specify up to four different order maps, based on load: Map 0: 0-25% load, Map 1: 26-50% load, Map 2: 51-75% load, Map 3: 76-100% load (Figure 4). In this way, the designer can ensure the driver will hear a different set of order content at a given RPM when under high load (i.e. – high acceleration) versus under low load (i.e. – cruising).
Figure 4. Simcenter Testlab Sound Designer offers 4 load planes where order content can be specified. This enables the order synthesis to sound different at different accelerator pedal positions, or load.
2.2. Granular synthesis
Granular synthesis is a method of sound generation that has been used in the music and entertainment industries for many years. It begins with a starting sound waveform and breaking it up into a short section, or “grain”. This grain is then looped (played end-to-end) to create an entirely new sound (Figure 5).
Figure 5. Simplified process of granular synthesis. Beginning with a sound sample (A), a small section is selected. This section is known as a “grain” (B). The new synthesized sound is created by replaying the grain over and over (C).
Further manipulation of the grain can dramatically change the nature of the synthesized sound (Figure 6 below). By selecting a longer grain, for instance, more of the starting sound’s character is retained. Selecting a grain from a different point in the starting sound (the end versus the beginning) may provide an entirely different character to the synthesized sound. Additionally, by changing other factors such as the number of grains, applying different window functions, overlapping the grains, and altering the pitch of the grain, an almost unlimited number of variations can be synthesized, all from the same starting sound.
Figure 6: Some of the ways the grain can be modified to alter the character of the synthesis in Simcenter Testlab Sound Designer.
Randomness is another variable that can be adjusted for any of the grain parameters shown in Figure 6. By adding a small amount of randomization to the grain parameters, for instance the grain position or grain length, the synthesized sound is made continuously new, rich, and “alive” sounding. Small changes to these parameters keep the sound from being repetitive and stale. 2.3. Pitched Playback
Pitched playback is slightly different from order synthesis and granular synthesis, in that an entirely new sound is not synthesized from scratch. Pitched playback is a replication-related method: it begins with a complete sound sample that is looped continuously, and the pitch of the sample is changed as a function of the dynamic control parameter (typically vehicle speed in this case, see Figure 7). This makes pitched playback the most predictable method of sound synthesis, as it maintains the character of the pre-existing sound sample and does not require any deep knowledge or experience in sound design.
Figure 7: Pitched playback utilizes a complete sound sample, which is looped continuously (left). During driving the pitch of the sample is altered in relation to a dynamic control parameter, such as vehicle speed (right).
Several layers of sample playback can be added simultaneously. Each of the layers can use a unique sound sample, or multiple layers of the same sample can be used to create different levels and directions of pitch change. Because it utilizes a complete sound sample, Pitched Playback is often utilized for AVAS sounds. Here, the precise nature of the sound can be maintained, and simply pitched up or down to meet pitch change requirements necessary for certification in certain parts of the world.
3. How to make an active sound exciting & pleasing
3.1. Multi-layer approach If implemented poorly, an active sound design can become annoying, and tiresome to listen to. To keep the sound fresh, exciting, and pleasing for vehicle occupants, Simcenter Testlab Sound Designer features a so-called multi-layer approach (Figure 8).
Figure 8: Multi-layer approach featured in Simcenter Testlab Sound Designer. Using multiple layers provides realism and flexibility to the sound design.
Just as the traditional sound from an internal combustion engine varies with different driving situations, so too must the synthesized sound. The various layers in Simcenter Testlab Sound Designer work together to provide a rich, dynamic soundscape that feels alive and authentic to the driver. As the inputs from the driver change, the sound design can change with it. More information on each of the layers can be found in the next sections.
3.1.1. Base Layer The base layer is the foundation of the sound design; it provides the overall character of the sound and how it evolves during driving. Typically, the base layer is tied to vehicle speed: as the vehicle speed increases, the base layer changes accordingly, often rising in pitch like a combustion engine does. The base layer can be synthesized using any of the available methods (Order Synthesis, Granular Synthesis, Pitched Playback, etc.) The primary role of the base layer is to give the driver some auditory feedback to their inputs: shifts in tone with changes in speed, perhaps a reduction in sound level when cruising at a constant speed, turning, braking, etc. It serves as the “sonic canvas” for the vehicle and communicates information about the brand and driving experience to the vehicle occupants.
3.1.2. Torque Layer While the base layer is primarily concerned with setting the overall tone or feeling of the sound design, the torque layer is used to “spice up” the sound when the driver calls for something more dynamic. When the driver accelerates quickly, perhaps merging onto a freeway, or if the vehicle begins to climb a steep grade, the load on the powertrain goes up significantly. Classically, in an internal combustion engine this increased load is accompanied by a noticeable increase in sound level and change in sound character. The transmission often downshifts, the RPM of the engine shoots up, and the engine produces a more powerful sound. While an EV manufacturer may not want to directly replicate the same sound character of an ICE under load, if the sound does not appreciably change when the load goes up, or when the accelerator pedal is pressed, the driver will likely feel as if something is missing. This is where the torque layer comes in. The torque layer provides an additional sound layered in over the base layer sound. It can be designed to increase the sound level, add roughness, or perhaps additional order content. It is used to convey the emotion of power and connect the driver to the vehicle in an emotional way. The torque layer is controlled and tuned by a dedicated page in the Simcenter Testlab software, so it is completely under the control of the sound designer. The torque layer sound can be synthesized much like the base layer: using Granular Synthesis, Order Synthesis, etc. However, instead of using vehicle speed as the dynamic control parameter, the sound designer will link the torque layer to “Load”. For an EV, this is often a weighted combination of accelerator pedal percentage and torque. The torque layer can utilize the same starting sound as the base layer, or a new sample can be added for use by the torque layer.
3.1.3. Shepard layer The Shepard layer is so named after an auditory illusion known as the “Shepard tone,” first described by American cognitive scientist Rober Newland Shepard. The auditory illusion consists of sine waves superimposed on each other, separated by an octave. As the bass pitch of the tone begins moving upward, it is subtly replaced by another, an octave lower. This process gives the illusion of a continually increasing tone, though the pitch never truly leaves the octave in which it began. This superposition technique is utilized by the Shepard layer in Simcenter Testlab Sound Designer to ensure that as the vehicle speeds up (and the frequency content of the Base layer is pitched upward) there always remains some low frequency content in the soundscape. The low frequency keeps the sound balanced and full, while also giving the impression of rising pitch.
3.2. Parameterized to vehicle data The multi-layer system for sound synthesis is a good start, but for the synthesis to sound truly authentic, as if it were being generated by the vehicle itself in a natural way, it needs to be connected to the vehicle itself. Simcenter Testlab Sound Designer accomplishes this by parameterizing the sound design itself in terms of vehicle performance data (Figure 9).
Figure 9. Simcenter Testlab Sound Designer allows for data from the vehicle to guide and shape the way the sound design evolves during driving. Vehicle parameters such as speed, torque, and RPM are mapped to sound design parameters via the software interface.
By mapping sound design parameters to vehicle performance data, the active sound can evolve and change along with the driver in a direct and natural manner. No two driving trips may be the same, and so neither should the sound generated by the active sound design. As the vehicle data changes, so does the sound synthesis. In this way, the driver hears a natural sound that relates well to the driver’s commands, resulting in a dynamic soundscape that feels as if it is “alive.”
3.3. Live synthesis In addition to live vehicle data informing how the sound design evolves during driving, the synthesis itself must be done in real time to feel authentic and pleasing. Simcenter Testlab Sound Designer synthesizes all active sound live on demand – there are no static looped samples or look-up tables. In the audio processing unit, there is only the short sample sound (in the case of granular synthesis and pitched playback) and a library of control parameters linked to vehicle data to shape the synthesis. As a result, the synthesized sound is always new, dynamic, and rich.
In the beginning stages of an active sound design, the primary focus is to experiment with different sounds, textures, and moods for the project. The goal is not fine tuning, but rather to create multiple avenues for design exploration and play. As such, most of the work is performed in the sound studio, or perhaps at the ASD engineer’s desk – no vehicle is required, and perhaps doesn’t even exist yet (Figure 10).
Figure 10: Testlab Sound Designer is a fully functional sound design tool: It can perform the live synthesis, create audio output, generate its own CAN-bus data, and provide live editing of the sound design.
Simcenter Testlab Sound Designer software contains all the functionality necessary to create a new sound, and the designer can begin immediately mapping the sound design to vehicle parameters. The software performs the live synthesis, generates the CAN-bus data stream, provides methods to manipulate the vehicle performance data (e.g. virtually “drive” the vehicle) and generates audio output. All that is needed is a set of speakers or headphones to begin creating an active sound design. The software also provides live editing capabilities: the designer can change the parameterization while listening to the sound and receive immediate feedback. This fully-featured capability makes Simcenter Testlab Sound Designer impactful from the very beginning of a new vehicle program or ASD campaign.
4. How to validate an active sound in the vehicle
Designing active sounds in the studio or office is a good first step. However, once the sound designs have begun to take shape and the features of the design have been roughed out, they must then be validated in a real vehicle (Figure 11).
Figure 11: An active sound design must first be developed at the desktop (software only) level once it leaves the design studio. The next step is to validate the sound design in a real vehicle.
In-vehicle sound design development requires hardware that can connect the Simcenter Testlab Sound Designer synthesis and tuning software, the vehicle performance CAN-bus data, and the vehicle audio system. This hardware is discussed in the following sections.
4.1. Vehicle Unit [HW-SD-VUNIT2] The Vehicle Unit is a central part of the Simcenter Testlab Sound Designer solution, as it performs multiple functions throughout the entire sound design life cycle – from design exploration to final validation. It is a small, orange box that features an Ethernet port (laptop communication), a DB-9 port (CAN-bus), and a micro-USB (service port) on the front face, four audio jacks and a DC power input on the rear face (Figure 12).
Figure 12: Simcenter Sound Designer Vehicle Unit. The unit serves as the main connection between the software and the vehicle during development/tuning.
The Sound Designer Vehicle Unit allows for in-vehicle sound synthesis, eases the sound-tuning process, and allows for long-term in-vehicle and simulator evaluation. It provides the link between sound synthesis, dynamic vehicle CAN-bus data, live synthesis editing, and audio output to speakers or the audio head unit of the vehicle.
4.1.1. Synthesis engine from desk to vehicle
During the initial development of an active sound design, the synthesis is being performed by Simcenter Testlab Sound Designer software on the host PC. The Vehicle Unit duplicates the TLSD software’s ability to synthesize sound, but independent from a PC. The Vehicle Unit sound generation is 100% compatible with the synthesis engine on the PC, while also taking hardware limitations such as sample length into account. This compatibility ensures that the sound synthesis developed at the desk will remain the same when performed in the vehicle.
The TLSD Vehicle Unit brings in the CAN-bus data from the vehicle (via the CAN Gateway, Figure 13), performs the live synthesis, and outputs the audio to the vehicle speakers. It also provides an ethernet port to connect to a computer to perform live editing to the sound design (see section 4.1.2).
Figure 13: In-vehicle sound design setup featuring the Vehicle Unit. The Vehicle Unit brings together vehicle parameters (CAN), sound synthesis, and audio output into one module.
Eventually, in production vehicles the sound synthesis will likely be performed by the audio head unit, infotainment system, or other dedicated audio hardware. The TLSD Vehicle Unit aids in this transition at the early development phases by taking hardware-related boundaries such as memory size and size of sound samples into account. This allows for easy and safe transfer of tuned sound signatures to production vehicles.
Once a sound design can be experienced during real driving, the next stage of development and tuning can begin. The combination of live driving and actual performance data from the vehicle connects the driver and the sound design in a powerful way. Additionally, hearing the active sound as it is reproduced by the vehicle speakers (as opposed to headphones or studio monitors) provides additional realism and feedback. Live driving also provides realistic ambient sounds such as wind and road noise, noise from ancillary vehicle systems, etc., that will need to be balanced with the active portion of the sound design.
4.1.2. Vehicle interface for live tuning
The TLSD Vehicle Unit also allows for real-time editing of the sound design via connection to the ethernet port (Figure 14). Live editing of the sound design allows for rapid tuning of the sound design, as well as saving multiple iterations for later evaluation.
Figure 14: In-vehicle hardware setup, including live editing capability using Testlab Sound Designer on the laptop computer.
Real-time editing of the sound design offers zero-latency feedback for tuning efforts: one evaluator can drive the vehicle and offer design feedback, while another designer rides in the passenger seat and implements the suggestions in real time. This immediate feedback allows for rapid tuning of the sound design, without needing to wait for edits to be made off-line and reloaded into the synthesis engine.
4.1.3. On-board storage
Another valuable feature of the Vehicle Unit is on-board storage. This storage enables the Vehicle Unit to be loaded with multiple sound designs, which can be loaded directly from internal memory without the need for a computer or the Simcenter Testlab Sound Designer software. The Vehicle Unit offers 4 memory slots, and the user can switch back and forth between slots during evaluation to quickly compare the different designs, or perhaps to represent different driving modes (Eco, Performance, Standard, etc.) For long term buy-off type evaluations, the Vehicle Unit can be concealed in a storage compartment or under the seat to avoid distraction and provide a realistic production experience. Sound design selection and recall can be made via an available smart phone application, giving the driver full control of the selected design, and a seamless interface to the sound design selection.
4.2. CAN Gateway [HW-SD-CANGATE2]
As shown in Figures 13 and 14 above, the CAN Gateway (Figure 15) is used in conjunction with the test vehicle (or simulator), Testlab Sound Designer software, and the TLSD Vehicle Unit. The gateway routes CAN input and output messages for the Simcenter Testlab Sound Designer software: it allows TLSD software to read the CAN messages from the vehicle or simulator, as well as writes (outputs) CAN messages to the TLSD software, Vehicle Unit, or the AVAS Unit. When used with the AVAS Unit (Section 4.3), the gateway is required to edit the AVAS signature stored onboard the unit.
Figure 15: The TLSD CAN Gateway allows for CAN messages to be safely transferred between the TLSD software, the vehicle, and the TLSD Vehicle Unit. It is also used for communication with the AVAS unit.
The gateway manages the communication streams between the vehicle and the rest of the TLSD chain. For messages from the vehicle, the unit is listen-only and prevents any CAN-bus messages from being broadcast on to the vehicle (or simulator) CAN-bus network, as this can cause dangerous vehicle information conflicts. The gateway instead duplicates the vehicle messages and sends them to the Vehicle Unit. For earlier development stages where the vehicle is not present, the CAN Gateway can read the CAN parameters from the TLSD software and send them to the Vehicle Unit for sound synthesis.
4.3. AVAS Unit [HW-SD-AVAS]
The AVAS Unit [HW-SD-AVAS] is a compact, lightweight loudspeaker used for developing and optimizing AVAS sound production for electric and hybrid vehicles. The stand-alone unit is capable of sound synthesis without the need for a PC and is modelled after production AVAS equipment found in the marketplace today (Figure 16).
Figure 16: The development AVAS speaker requires only DC power and a CAN-bus signal to produce AVAS sounds.
Typical mass production solutions for AVAS devices show limitations regarding sound reproduction capabilities – small, lightweight, and inexpensive loudspeakers must be used. Sounds designed in Simcenter Testlab Sound Designer using headphone or high-quality speakers will not take these restrictions into account and will therefore not deliver the optimal AVAS sound. Using the AVAS Unit for development incorporates these limitations on sound reproduction, providing the ability to optimize the AVAS sound design with the end in mind (Figure 17).
Figure 17: AVAS Unit receives CAN-bus signals from the vehicle, and synthesis is performed on the unit. AVAS sound signatures can be edited live using Testlab Sound Designer software; this flexibility makes adjusting the sound design fast and easy.
The final mounting position and direction of the AVAS loudspeaker will also have a significant impact on the AVAS sound design as observed at the pedestrian locations. By utilizing the AVAS Unit early in the vehicle development process, several options for speaker mounting location can be assessed for noise performance in addition to other constraints: vehicle packaging, dust and water ingestion, crash and safety, etc. The sound design can be optimized based on the various mounting locations to ensure it meets regulatory requirements.
5. How to validate an AVAS design
5.1. Global AVAS regulations summary
Since the mid-2019, legislation all over the world had made AVAS mandatory for vehicles with an electric mode of operation, imposing a minimum noise level. Two major regulations have been developed to define electric vehicle warning sounds: United Nations Economic Commission for Europe (UN ECE) Regulation 138, and in the United States, the National Highway Traffic Safety Administration’s Federal Motor Vehicle Safety Standards (FMVSS) 141 (see Figure 18).
Figure 18: AVAS regulations around the world. There are two primary regulation standards: ECE R138 & NHTSA FVMSS 141.
Apart from the detectability of the presence, direction and location of quiet vehicles, AVAS sound must allow pedestrians to identify critical operating scenarios such as constant driving speed, acceleration, or deceleration.
Most countries adopt the UN ECE Regulation 138 in their local legislation, sometimes with slight modifications. For example, in China, GB/T 37153 has become mandatory. It is fully aligned with UN ECE Regulation 138, except that all minimum noise levels are increased by 2 dB. In the U.S., FMVSS 141 is quite like ECE R138, with a few deviations. ECE R138 and FMVSS 141 regulations are summarized in the table below (Table 1).
Table 1. Summary of EU ECE R138 and NHTSA FMVSS 141 minimum noise standards for electric vehicles.
Highlights and Differences:
ECE R138:
Tests can be done outdoors, or in a large hemi-anechoic chassis dyno.
Moving speeds can also be simulated via CAN-bus input to the AVAS speaker(s).
For 10 kph & 20 kph tests, 2 octave bands between 160-5000 Hz must meet minimum noise levels. Additionally, one of these bands must be below 1600 Hz in frequency.
Reverse test must be at least 47 dB(A), either moving at 6 kph or stationary.
No other stationary requirement.
Pitch shift requirement (0.8% / kph) to indicate speed increase.
FMVSS141:
All tests must be performed outdoors.
Includes a stationary directivity test using a third microphone on the CC’ line (see Figure AVAS4)
Adds a third forward operating speed at 31 kph.
Slightly modifies operating speeds to 11 kph, 21 kph, 31 kph, each +/- 1 kph.
Reverse test must be stationary.
For all tests, US legislation allows for two different compliancy paths:
4 non-adjacent bands in a span of at least 9 bands (315 – 5000Hz).
2 non-adjacent bands, one below 1000Hz and one above 800Hz, + minimum band sum.
In both 4-band and 2-band option, each of the bands selected must meet individual band minimum SPL.
2-band method adds minimum band sum requirement (e.g. band sum > 44 dB)
No requirement on pitch shift with increasing speed.
Adds increasing Volume requirement with increasing vehicle speed (3 dB+ per 10 kph increase).
Measurements for both stationary and moving tests for ECE R138 & FMVSS 141 are made at the microphone locations shown in Figure 19. Microphones are placed 1.2m above ground level aligned with the front plane of the vehicle (Line PP’, Figure 19). Microphones are placed 2m from the centerline of the vehicle (line CC’, Figure 19). For moving tests, the vehicle drives along line CC’ at a constant speed, and noise levels are evaluated as the vehicle enters the test zone until the front plane of the vehicle crosses line PP’. For stationary tests the front plane of the vehicle is placed at line PP’ (for Reverse tests, the rear plane of the vehicle is placed at the line PP’). FMVSS 141 adds a center microphone along line CC’ for directivity tests called out in the standard.
Figure 19: Microphone measurement locations (green circles) for stationary tests in both ECE R138 and FMVSS 141. FMVSS 141 adds the center microphone (orange circle). Microphone locations are the same for Reverse tests, with the rear face of the vehicle placed on the PP’ line. For moving tests, the center microphone is removed, and the vehicle is driven along the center line (CC’) at constant speed.
5.2. Assess AVAS speaker mounting locations
The sound that reaches the regulatory microphones during minimum noise compliance testing will be a combination of a) the sound design itself (frequency content), b) the level of playback of the sound design (amplitude), and c) the acoustic transfer function between the speaker location and the microphone locations. Different mounting locations will each produce a unique acoustic transfer function, dependent on the specific location, direction, and environment the speaker is in (see Figure 20). This transfer function shows the relationship between the speaker and microphone locations in terms of frequency, pressure, and voltage.
Figure 20: The mounting location of the AVAS speaker will affect the sound recorded by the regulatory test microphone. Three different possible packaging locations (A, B, C) for the speaker will each have a unique acoustic transfer function. The same sound played at each of these locations will all sound different at the microphone.
After leaving the speaker, the more objects and barriers the sound must travel around, through, or bounce off before reaching the microphone will affect how the sound is perceived and measured by the microphone. For example, a speaker placed behind the bumper of the vehicle will sound different than the same sound played through a speaker mounted in front of the bumper. An example of a real acoustic transfer function is shown below in Figure 21.
Figure 21: An acoustic transfer function between an AVAS speaker and a microphone. This function illustrates the ability of the speaker to create pressure (per unit volt) at the microphone position at various frequencies.
The units are measured in “Pa/V” or pressure per unit voltage. This function is created by playing broadband noise through the speaker and measuring the response with a microphone placed at the regulatory positions shown in Figure 19. In addition to measuring the pressure response at the microphone, the voltage at the input to the AVAS speaker is also recorded. This enables the transfer function between these two signals (pressure and voltage) to be measured and plotted versus frequency. Peaks in the transfer function show the frequencies at which the speaker is efficient at producing sound at the microphone, while valleys show the frequencies where the speaker is not very good at producing sound at the microphone position. These peaks and valleys play an important role in determining whether the AVAS sound design will pass or fail regulatory testing and must be accounted for by the sound designer. This is discussed further in the next section.
5.3. Predict AVAS sound design performance to regulations
Knowing a product will receive certification prior to testing is crucial to avoiding wasted time, effort, resources and money. The sooner the engineer can learn the design will fail and requires modifications to pass, the better. The “AVAS.CHECK” module of Simcenter Testlab Sound Designer allows engineers to assess their sound designs against global minimum noise regulations and predict whether it will pass or fail at any stage of the vehicle development timeline. The AVAS.CHECK module takes as input recordings of the sound design at the various operating conditions set forth in the standards (stationary, reverse, 10 kph, 20 kph, etc), and the acoustic transfer functions measured between the speaker mounting location and the microphones (see Figure 22). With these two pieces of information, AVAS.CHECK can predict the acoustic signal at the regulation microphone location and compare the results to the requirements. The result is a matrix of regulations and regions, showing pass (green) or fail (red) for each condition (Figure 22).
Figure 22: The AVAS.CHECK module inside Testlab Sound Designer can predict whether a sound design will meet the various regulation requirements set forth in the global standards. If the design fails one or more conditions in the regulation, either the sound design or the mounting strategy must be amended.
Once the sound design is loaded and a set of transfer functions are provided, AVAS.CHECK can produce the evaluation to the standards. The various outputs of the AVAS.CHECK module are shown below in Figure 23.
Figure 23: The AVAS.CHECK module shows results in 4 areas: The main summary matrix (upper left), overall sound pressure level versus time (upper right), frequency & level information (lower right), and when a cell in the summary matrix is selected, the details of the status will appear in the detail area below the matrix (lower left).
Results Summary Matrix The status report matrix shows the predicted performance of the sound design (combined with the acoustic transfer function) to the global standards. Across the top row are the various global regions: EU (European Union), US-A (4-band method), US-B (2-band method), JP (Japan, matches EU) and CN (China, EU + 2dB). The test types across all regions are listed in the first column on the left, and once the prediction is made, the cells of the table turn green (“ok” = pass) or red (“NOK” = fail). See Figure 24.
Figure 24: AVAS.CHECK prediction result table. Colors indicate whether the requirements for a region and test condition are met and pass (green) or fail (red).
Cells at the intersection between a test type (row) and region (column) show the result of the prediction. If the sound design is predicted to not meet local regulatory requirements, the cell is colored red and labelled with the text “NOK,” short for “not ok”. If the sound design is predicted to meet the requirements for that test, the cell is colored green and labelled “ok”. Cells where particular test types are not required for that region are left empty.
To highlight how AVAS.CHECK can help a sound designer identify issues of non-compliance and remedy them, consider the following example. The results of an AVAS.CHECK prediction are shown in Figure 25 below.
Figure 25: Results of an AVAS.CHECK prediction for an example sound design and speaker configuration. The result matrix shows several tests where the current design does not meet the EU regulation ECE R138 (highlighted with a yellow box).
Considering the regulation for the EU (ECE R138), it is clear from the red boxes in the result matrix that the evaluated AVAS sound design is predicted to fail several tests: For the 10 kph speed, the design fails the minimum sound pressure level, the 1/3rd octave band requirement, and the pitch shift requirement. For the 20 kph speed condition, the design is predicted to fail the minimum level and the 1/3rd octave requirements. This is good information for the sound designer, but more particulars about why these tests failed is needed. By clicking on the test condition in the first column of the result matrix (for example “10 kph Level min,”) additional information is provided in the displays to the right of the matrix (see Figure 26).
Figure 26: Selecting an operating condition in the first column of the result matrix populates the displays to the right of the matrix. Overall level versus time is shown in the upper display (different colors for various forward speed conditions), and the lower display shows the 1/3rd octave details for the selected speed condition.
In the upper overall level versus time display, multiple speed conditions are shown in different colors. (The 10 kph speed is shown in orange and highlighted by the arrow in Figure 26). The Max and Min levels for each condition are shown as straight red and green lines, respectively. These requirement lines change based on the operating condition selected in the result matrix. It is clear from the upper display in Figure 26 that the 10 kph overall level (orange line) does not stay above the green Minimum line, thus the requirement has not been met. The lower display in Figure 26 shows the predicted 1/3rd octave spectrum for the selected operating condition as well as the required levels for each band according to the test specification. In this example, all the octave bands are below the required level. The red numbers along the X-axis display the number of decibels the prediction is below the required level for each band. For designs where the octave band levels are met, this number would turn green and specify the number of dB above the requirement.
Matrix Detail Panel In addition to the information shown in the overall level versus time and 1/3rd octave band displays, clicking on a cell in the status report matrix will populate the detail panel below the matrix with additional details about the test prediction. See Figure 27.
Figure 27: Clicking on a red NOK cell will populate the status window with details of why the condition failed, and what changes need to be made to pass the requirement. For example: Pitch shift needs to be increased by 0.1% to meet the EU requirement at 10 kph (blue arrow), and the max level in the 20 kph test needs to be reduced by 3.1 dB to pass (green arrow).
The matrix details panel provides specific information about what the regulation requires, what was predicted, and the difference between these two values. Figure 27 above shows two conditions that did not meet the ECE R138 regulation: Pitch shift for the 10 kph speed, and maximum level for the 20 kph speed. Clicking on the NOK cells shows detailed information in the details panel below the matrix. For the 10 kph condition, the pitch shift needs to be increased by at least 0.1% from the current design (Figure 27, blue box). For the 20 kph test (Figure 27, green box), the current AVAS design is too loud at this condition. The regulation maximum is 75.0 dB(A), and the predicted level is 78.1 dB(A). The sound design must be changed to reduce the overall level for the 20 kph operating condition by at least 3.1 dB.
1/3rd octave band requirements For the 1/3rd octave band requirements, the detail pane uses “flags” to report the status of the various band requirements for the different regions. These flags are numbered 1-4, and shown left-to-right in the result matrix cells:
1. The number of octave bands meeting the minimum level 2. The frequency range of the bands 3. Bands must be non-adjacent (US only) 4. Band sum must meet minimum level (US, 2-band only)
The specific regional requirements for 1/3rd octave bands are repeated below:
ECE R138:
For 10 kph & 20 kph tests, 2 octave bands between 160-5000 Hz must meet minimum noise levels (Flag 1, number of bands). Additionally, one of these bands must be below 1600 Hz in frequency (Flag 2, frequency range of bands).
FMVSS141:
For all tests, US legislation allows for two different compliancy paths:
4 non-adjacent bands in a span of at least 9 bands (315 – 5000Hz). (Flags 1 & 3)
2 non-adjacent bands, one below 1000Hz and one above 800Hz, + minimum band sum. (Flags 1-4)
In both 4-band and 2-band option, each of the bands selected must meet individual band minimum SPL.
2-band method adds minimum band sum requirement (e.g. band sum > 44 dB)
When a 1/3rd octave cell in the result matrix is clicked, the status of the 4 flags is shown (Figure 28). Simcenter Testlab Sound Designer uses the following nomenclature for the flags: 1 : Criteria met 0 : Criteria not met -1 : Criteria does not apply
Figure 28: Clicking on a 1/3rd octave cell in the result matrix shows the status flags for the selected operating condition & region. Clicking on the “NOK” cell (step 1 above) reveals the status flags (step 2). The flags “0 0 -1 -1” indicate that the AVAS design did not meet the “1. number of bands” requirement, nor the “2. frequency range” requirement. “3. Non-adjacent bands” and “4. Band sum” requirements do not apply to this region.
The result shown in Figure 28 is “0 0 -1 -1”. These flags correspond to the requirements put forth in the ECE & FMVSS standards and are summarized in the detail pane: 1. Number of bands, 2. Frequency range of the bands, 3. Non-adjacent bands, and 4. Band sum. The result “0 0 -1 -1” indicates that the AVAS design is predicted to not meet the first two criteria (number of bands meeting the minimum level, and the frequency range of these bands), and the last two criteria (non-adjacency of bands and the band sum) do not apply to this region. This design will need to be changed in order to satisfy the requirements for this region.
5.4. Example: Assessing an AVAS sound design for FMVSS141
To further examine how the AVAS.CHECK module can be used by a sound designer, consider the following example where a sound design is assessed with respect to a portion of the FMVSS141 standard for the United States. To further expand on the 1/3rd octave result flags and how they work, consider the AVAS.CHECK prediction shown below in Figure 29. Focusing on the 20 kph operating condition for the US region, it is clear from the red “NOK” shown in the result matrix that both the US-A (4-band) and US-B (2-band) fail the 1/3rd octave requirements for this operating condition (Figure 29).
Figure 29: The AVAS sound design is predicted to fail both the 4-band and 2-band method options for FMVSS141. The US only requires one of these methods to pass, however, this design will need to be modified to satisfy the requirement.
Clicking on the red “NOK” cells in the result matrix provides additional insight as to why the design fails. Clicking on the “NOK” cell for the US-A (4-band), shows more info in the detail pane (Figure 30). The status flags for the 4-band result are shown to be “0 -1 1 -1”: criteria #1 fails (number of bands), criteria #3 passes (non-adjacent bands), and criteria #2 & #4 are not applicable for this region, denoted by the “-1” flag value. The first criteria for the 4-band regulation requires at least 4 octave bands meet or exceed the minimum dBA levels laid out in the FMVSS141 standard (shown as light red bars in Figure 30).
Figure 30: Detail for US-A (4 band) prediction. The 4-band method requires at least 4 octave bands to meet the minimum level. Only 1 band (3150 Hz) is predicted to have a sufficient sound pressure level. All other bands are too quiet to meet the 4-band regulation.
For the assessed AVAS design, only one octave band reaches the required level, the 3150 Hz band. All other octave bands are too quiet, and the design will need to be changed to meet the 4-band FMVSS141 regulation.
What about the 2-band method? Detail for the 2-band method is shown below in Figure 31. The 2-band method only requires two octave bands to meet or exceed the minimum dBA level but adds requirements for the frequency and spread (distance apart) of these two bands, as well as sets a minimum band-sum level for the two bands. For the 2-band method all 4 status flags are applicable (Figure 31).
Figure 31:Detail for US-B (2 band) prediction. The 2-band method requires at least 2 octave bands to meet the minimum level, as well as specifying the frequency range and adjacency of the bands. Additionally, the 2-band method stipulates a band-sum SPL that must be met by the two selected bands. In the above instance, the octave band requirements are all satisfied (indicated by the “1”s in the first three flag positions), but the band-sum is not sufficient to meet the regulation (indicated by the “0” in the 4th position).
To meet the FMVSS141 regulation, the sound designer needs to increase the levels in at least two of the octave bands to satisfy the 2-band method band-sum requirement. Rather than make changes to the AVAS speaker location, the designer chooses to increase the gain of the sound design for the 20 kph operating condition. Once the design has been altered, the various operating conditions are recorded in the TLSD software, and the design can be re-evaluated in the AVAS.CHECK module. The updated results for the US-A (4-band) and US-B (2-band) predictions are shown in Figures 32 & 33, respectively.
Figure 32: Updated predictions for FMVSS141 4-band method. All criteria are predicted to be fulfilled: 4 bands meeting minimum SPL, bands cannot be adjacent and must span at least 9 octave bands between 315 Hz and 5000 Hz bands. With all but 3 octave bands meeting minimum levels, a number of combinations satisfy the 4-band requirement, one such option is shown.
Figure 33: Updated predictions for FMVSS141 2-band method. All octave bands meet the minimum SPL level. The loudest band in the 315-800 Hz range is selected, along with the loudest band in the 1000-3150 Hz range (green arrows). These two bands cannot be adjacent and must span at least 3 octave bands. In addition, the 2-band sum must exceed 57 dB(A) for the 20 kph condition (57.857 dB(A), green box lower left). All criteria are satisfied!
Active sound design presents an exciting opportunity for sound designers to bring brand character and passion to drivers of electric vehicles through interior sound enhancement. External AVAS sound design provides pedestrians auditory feedback about an electric vehicle’s position and movement, providing critical safety information. Simcenter Testlab Sound Designer software and hardware provide a powerful set of tools that can be utilized from the very inception of a vehicle all the way through regular production for crafting rich, lively, active sound designs for interior occupants and pedestrians alike.
