How does Virtual Acoustic Tuning Develop Optimal Sound?

September 23, 2015:

 

Virtual Acoustic Tuning considers knowledge of psychoacoustics

Within the multiphysical simulation model (Tier 1- 3 MVOID®  process) the distribution of the sound can be determined right into the backmost corner by purely mathematical parameters. The optimum position for loudspeakers is found. The best possible components can be selected.

However, it isn’t sufficient to select suitable components and positioning them in the best place – whether they are standard or high-end ones. The authentic listening experience can only be achieved by adjusting the sound to the individual vehicle type and sound system on hand – the acoustic tuning process. The virtual tuning enables a computer-based adjustment that integrates the knowledge of psychoacoustics, the subjective sound perception.

 

Tasks of Virtual Tuning:

1. Minimize Resonances

Irregularities and resonances (= acoustic eigenmodes) can be minimized in the frequency response by careful virtual sound tunings. Different colored oscillations on the screen show acoustic experts the interaction of the sound system. Significant increases or reductions are visible.

At certain positions in the vehicle resonances occur: Frequencies that are too loud or too soft for the human ear. Frequency responses, for example, that show significant reductions can be an indicator that certain frequencies – an essential part of music – are missing. Musical instruments are not audible.

In addition, the sound level on all seats must be analyzed and disruptions must be eliminated, too. There may be areas within the complex architecture of the car cabin in which sound waves cancel each other out. It takes numerous measurements to analyze how each channel must be set up to reproduce a proper sound experience. Special filters – electroacoustic elements of signal processing – support the process to lift or lower frequencies.

 

2. Compensate reflection characteristics

Furthermore the virtual acoustic tuning process compensates reflections caused by resonant panel members, windows, set covers etc. A bass speaker, for example, can clearly reflect on the windshield. The various reflective properties need to be rectified for each channel in order to achieve optimal sound quality.

 

3. Adjust run-time differences to a uniform level

Another challenge is to adjust the different run-time differences of the different speakers to the listener. The speaker in the driver’s door reach the driver much sooner than the speaker, integrated in the co-driver’s door, for example. The sound does not arrive simultaneously. In addition, the volume is different. The difference of the sound pressure level must be compensated. By means of virtual acoustic tuning it can be discovered directly, for example, that speaker A arrives 5 milliseconds later than speaker B. Consequently speaker B must be delayed by 5 milliseconds.

For all seating positions, run-time differences from the various loudspeakers to the listener and the volume level are regulated to a uniform level.

 

Graphic examples:

Figure 1: Rawdata before tuning process

 

Figure 1 shows a significant drop in the frequency response. Electrical energy cannot be converted into acoustic energy. The speaker vibrates and stimulates a part that in turn causes eliminations. In this example, corrections are even not possible.

 

Figure 2: Orange oscillation: Dither oscillations of all channels. Tuning of 6 bass channels / raw data before tuning: The subwoofer (wave in purple color) is noticeably louder than the door woofers. Numerous dips in the frequency response are visible.

 

Figure 3: Measurement after tuning process. The data show that the drop cannot be compensated sufficiently.

 

As the figures 1 – 3 demonstrate, an analysis may lead to the conclusion that a planned positioning of loudspeakers is unfavorable to produce a harmonious sound.

 

Frequency response after successful tuning

Figure 4 illustrates an optimized waveform. The result is a well-balanced acoustical energy distribution across all frequencies.

Real oscillations always consist of several dither oscillations at different frequencies. In the nature perfect sinusoidal oscillations don’t exist. This can be explained by the fact that real oscillations have a finite length and are therefore restricted by an attenuation and transient effect/response. For the tuning process, this means that attenuation and transient effects as well as elements of psychoacoustics (human hearing range*) have to be considered. By contrast, a mathematically exact sinusoidal oscillation is invariant and undisturbed.

The human acoustic range covers approximately the range of 20Hz to 20,000Hz (at an older age, the ceiling may fall below 10,000 Hz).

For complex psychoacoustic processes the simpler subdivision in bass-midrange-treble is not sufficient. The following overview provides a further subdivision:


 Above list shows guidance levels. The frequency ranges cannot be assigned clearly.

After the base-tuning, minimizing resonances, compensation of reflections and adaptation of runtime differences have been made, the system does not sound optimal at this time. In other tests and measurements, the sound is further refined. A binaural reproduction technique allows us to listen to the music – the music can be heard on the screen at this time.

In the past, the large number of acoustic influences meant that acoustic experts had to adjust respectively eliminate lots of parameters manually. Today the virtual tuning process uses the extensive measurement data and can correct run-time differences automatically. This takes only a fraction of the time compared to the manual acoustic tuning of the car cabin in the past.

 

Binaural reproduction technology makes simulation audible

The transition from virtual tuning to auralization (Level 5 of the MVOID®process) is smooth. In the phase of auralization the acoustics experts finally evaluate the audio performance on the virtual model, they made further listening tests, especially in terms of spatial reproduction. When we speak about auralization it comes to the width and height of the stage – the music is to be reproduced according to a life performance. Musicians are standing a few meters away from each other. They are not at a central point. This effect is transmitted. The spatial reproduction cannot be determined only by mathematical parameters or measurements. The effect of spatial width can only be created by the auralization.

The final sound must match the highest quality requirements of the customers. Depending on the vehicle model refined nuances are desired. For a sports car, the bass, for example, can be more audible compared to the sedan. Thus, the “DNA” of the vehicle model always plays a role in the final sound-tuning. The experience of the acoustics experts decide on the desired characterization and the final sound profile that the system develops in a new vehicle.

 

Transition between virtuality and reality

Once the first prototypes exist, the results of virtual measurements and the real measurement can be compared seamlessly. The virtual tuning can be verified in this step. Numerous comparisons in the past confirmed that the virtual simulations are sufficiently accurate in order to make the right decisions already in the concept phase.

 

Benefit: Tremendously improved integration and high potential for savings

The particular benefit of this realistic simulation in the audio field is that the product is already “hearable”, before the first prototype is developed. The system architecture of the advanced development can be verified. The desired product properties can be reviewed and defined. In case of strong dips, visible by a strong decrease of the oscillations, as illustrated in Figures 1 to 3, a satisfactory modification cannot be carried out. Consequently, within the concept phase it is possible to verify how well a system can be tuned, how strong disturbing effects can be teased out. This is usually done as part of the feasibility study.

This is a bonus, which accelerates the development process many times and offers high saving potentials by the enormously improved integration.

 

From multiphysical to multidisciplinary simulation model

While we initially show the speakers up to the complete sound systems in a multidisciplinary simulation model by coupling the physical areas of electromagnetism, mechanics and acoustics (Phase 1 to 3 of the
MVOID®Process), we integrate various engineering disciplines in the process of virtual tuning. Measurement data of the amplifier and the signal processing unit are also to be considered as findings of psychoacoustics. Consequently, we are talking at this stage about a multidisciplinary simulation model.

In order to achieve most realistic simulation results and to get the maximum benefit, the whole process, starting with the analysis of the individual speaker to the analysis of the sound system up to the virtual tuning and auralization must be performed. Thereby, sound tuning is an essential step in order to meet the quality requirements of customers and consumers.