Determining the Repeatability of Automotive Audio Testing
Introduction?
Within the automotive industry, having a fine-tuned and quality audio system is becoming a customer standard as opposed to a primary selling feature. What started as AM/FM radios morphed into multichannel surround sound that supports wireless connectivity to limitless streaming options where consumers can enjoy their favorite tracks in a ‘Hi-Fi’ environment.?
As the customers’ perspective changes, then too must the designer and engineers’ perspective. When tuning high fidelity audio systems, confidence in your measurements must be paramount. The first step to ensuring confident measurements is ensuring that the equipment being used is repeatable and reflects that of an end-user’s perspective.?
I put this alignment to the test. How repeatable is a Head and Torso Simulator (HATS) when simulating multiple measurements and tuning events on the same vehicle??
Methodology?
To accurately simulate a scenario that one might use for final validation or a tuning process in a vehicle, I used an HMS II.3 LN HEC, the labCORE with a Bluetooth module (coreBT2), and ACQUA.??
The HMS was mounted on a seat mount and placed in the driver’s seat. For each measurement group, the HMS was completely removed from the vehicle, then replaced.??
Once replaced, the overall level was checked, and adjusted, ensuring that it was 85dB[SPL] (+/- 1dB) and two measurement runs were recorded. The source file chosen for this study was the MDAQS sequence (the Multi-Dimensional Audio Quality Score), seen below…more on that later! It consists of two sine sweeps (left, then right), followed by excerpts of different genres of music. For our purposes this is both a dynamic and applicable signal when discussing Automotive Audio.?
From each measurement run, the Active Speech Level (ASL), Frequency Response, and Loudness were calculated. The testing was performed on a mid-level crossover vehicle. The source signals were played via Bluetooth and recorded via the HMS.
Results?
Standard Seat position (95th Percentile)?
The classical and standard analysis when discussing audio is frequency response so we’ll start there. The following graphs are the frequency responses for both the left and right ear for all 20 measurements recorded (10 groups of 2 runs).?
Looking at the groupings, they seem to overlay and give consistent results. There are a few takeaways looking at the raw responses.?
However, we are looking to see how consistent. The average frequency response of all 20 runs is shown below with a mask of the standard deviation.?
Yes, those graphs are small - and the variation is even smaller. Time to ENHANCE!
领英推荐
Breaking up the analysis into sub-bands, the two graphs below show the average response from 2500 Hz to 10 kHz, and 10kHZ to 20kHz respectively.?
Conclusion?
For repeatable results, the HMS is capable of consistent results within a 1dB tolerance window. The ease of using a Seat Mount helps to place the HATS in the same physical position on the seat. Care should always be taken to ensure that the seat itself does not move as changing the seat position does impact the results…which will be investigated further.???
The benefits of using a HATS in place of a stereo or mono microphone(s) allows for a true perspective of what the customer would experience in the vehicle. Following a consistent process and using calibrated equipment, repeatable results are achievable and add insight to how better tune and evaluate audio systems in the automotive industry.??
Bonus Round – MDAQS?
As we discussed previously, developing and tuning complex systems and devices towards a desired sound signature is becoming increasingly difficult. Even to the point that basic measurements (e.g. frequency response) may not represent the system in its entirety and is unable to predict human perception. To meet the demand for a quick and cost-effective solution to assess audio playback quality, HEAD acoustics developed MDAQS – the Multi-Dimensional Audio Quality Score. The algorithm calculates a Mean Opinion Score (MOS) ranging from 1 (worst value = bad) to 5 (best value = excellent).??
Using the MDAQS source signal and recording binaurally, the MDAQS algorithm calculates four MOS values for the DUT:??
Using the 20 runs we recorded and analyzed, let’s see how they stack up when running MDAQS. The average MDAQS score for all 20 runs at the 95th percentile for this vehicle is:?
?We can look at the deviation by comparing each calculated score to the average to see how repeatable results are when replacing the HATS between each measurement group. The results are shown below.?
Over the 20 runs; MOS-T, MOS-I, and MOS-O were within 0.1 MOS points from the average. MOS-D showed higher variability across the measurement groups; however, this could be secondary to the variation in higher frequencies noted in the previous analysis.?
MDAQS can be a way to help to determine audio quality and set benchmarks as a tuning event is taking place.?
Overall, there are multiple ways to approach tuning and evaluating automotive audio. Regardless of what tools are in your toolbox, the importance is having confidence in the measurements you are taking and repeatable results. Consider adding a HATS to your toolbox to help give a different perspective on customer perception, and further expanding your capabilities to incorporate advanced metrics for quickly advancing systems.?
"Measure what matters" - considering current and potential future needs of the recording. Is there a chance that, 2 weeks after audition, that tuning changes are judged as better, or worse, or the same? Making a recording of them to listen back-to-back would go a long way... Repeatability helps too! Great job Jacob Soendergaard and Matthew Lutz!
Acoustician
3 个月Nice work. Can you repeat with 10 different HMS units in & out of same vehicle?
Helping professionals solve NVH and voice quality problems for over 25 years
3 个月IMHO there is room in the world for both the famous 'Christmas Tree' and a HEAD for tuning car audio. While the 'Christmas Tree' or 6 Microphone Array has been around / is an unspoken industry standard that helps tune cars quickly, nothing will beat the reproducibility or realism of a measurement binaural HEAD. But why limit yourself to one or the other when just about everyone these days has access to way more than just 8 channels of acquisition + analysis? Want to know more about the venerable 'Christmas Tree'? Read the founding 1986 AES Journal paper (paywall...but worth it!) from Henry Blind and Dr Earl Geddes "The Localized Sound Power Method" (https://aes2.org/publications/elibrary-page/?id=5278).
Consultant - Engineering Services at HEAD Acoustics, Inc.
3 个月Happy to help! This was a fun project to take on and it provided some valuable information and insight for even us internally.