Authors: Kilwa Ärölä and Markus Vaalgamaa
Wireless earbuds have become very popular, and we see people with wireless headphones in public places, vehicles and at homes daily. Users have learned that, regardless of all innovations, it’s very difficult to find one design that fits every individual ear. Companies that develop and manufacture earbuds or in-ear headphones struggle with this issue.
We at SoundlyFit have created a cutting-edge technology to solve the problem of comfortable design. SoundlyFit’s solution is a comfort simulation using physically realistic FEM models of the ear and earbud. We can help the design team to test and optimize earbud ergonomics faster and more precisely than ever before. We would like to elaborate on some technological motivations and details in this article.
Slow and inaccurate user tests
The requirements for an ergonomic in-ear headphone or earbud design are a comfortable and secure fit, and visual appeal. For best acoustics, sound quality, and exterior noise suppression, the earbud tip must create an airtight seal between the outside world and the ear channel. Like so often in product development, in this case too the design requirements are contradicting. A stiffer ear tip retains its position in the ear well, but it can create an uncomfortably high contact pressure and deformation in the sensitive skin of the ear. On the other hand, a very soft ear tip can feel almost unnoticeable in the ear and form a tight fit with the ear channel, but the contact pressure and friction are too low to keep the earbud in place.
Traditionally, the design of earbuds has relied on iterations through trial and error. Prototypes of the earbud design have been made, and their ergonomics have been evaluated with smaller or larger scale human ergonomics tests. In the ergonomics test, selected test subjects try out the design in their ears and rate the designs for fit, comfort, and other variables. Based on the outcome, the best design candidates are selected for further development. This design process however has shortcomings:
Creating physical prototypes is time-consuming,
An adequate number of test subjects must be recruited for the ergonomics tests, and
Results may be biased results due to possible problems in the test design.
We have conducted numerous ergonomics tests and gained several insights on the subject, including the fact that planning and execution of full user testing takes between one to two months. The fastest time we have seen is about three weeks, but this process required very good preplanning and fast delivery of updated silicon tips and earbud mock-ups. In practice, these time frames allow a maximum of two tests within a typical two–three-month design cycle of a new earbud, which is not sufficient for making significant improvements in ergonomics.
However, the biggest issue in user tests is not the test time, but the inaccuracy of human perception around ear. The test participants are unable to pinpoint the exact cause that makes the earbud uncomfortable, and therefore selecting the appropriate corrective action is very difficult for the design team. So, as an outcome, the ergonomics test often indicates that there are problems in the current design, but quantifying what should be improved concretely is very difficult. This fact and the lack of an adequate number of “trial and error” iterations can make the human testing ineffective and often even useless.
SoundlyFit’s Simulations – A Global First
SoundlyFit has created a unique solution to resolve these ergonomics testing problems using computer simulations. By integrating realistic physics-based simulations in the design process the designs can be developed and validated faster. Clear feedback of possible problem areas can be provided to the designers. Ultimately, this leads to reduced design costs and faster time to market.
SoundlyFit’s simulation process consists of two parts. The first part is a structural mechanics simulation of the contact between the earbud and the ear based on the Finite Element Method (FEM). The second part converts the simulation results into the comfort and ergonomics perceived by people when wearing the earbuds.
The FEM is a commonly used simulation method in design processes of smartphones, cars, bridges, skyscrapers, airplanes, and space shuttles. We adapted the same technology for ears and earbuds to simulate and calculate the deformations of the earbud, ear tip, and ear, and the contact pressures and the strains in the ear. To the best of our knowledge, we are the first company in the world to introduce a precise mechanical FEM model of the human ear. We at SoundlyFit are super happy and proud of this achievement.
Our FEM simulation consist of three phases. The first phase is creating accurate physical models of the ear and earbud using FEM. We’ve done precise physical force-versus-displacement measurements to stretch a variety of human ears and from a variety of places, and then built a mechanical FEM model of these ears in our computer. This ear model has the same parts as a human ear has: skin, cartilage, bone and joints. We have optimized the material parameters, locations, and thicknesses of these parts iteratively tens of times until we got the same results in the simulations as in the actual measurements in human ears. We have confirmed that our FEM model works not only for an “average” ear but for all kinds of human ears, even for those that are small, large, soft or stiff. We have also made mechanical models for earbuds. In the second phase, we have automated the placing of the earbuds into the ears, mimicking how humans do it. This sounds like a simple thing to do, but it took considerable effort to realistically fit slightly oversized earbuds into the ears. We finally mastered this, and we have confirmed that in more than 90% of the simulations we achieve the same position in the simulations as what humans did with real earbuds. In the final phase, the outcome of the simulations shows all the contact surfaces between the earbud and the ear—we use measures such as contact pressure, stress, and displacement from the simulations.
Once the simulation is completed, the results from the mechanical simulations are further postprocessed using a comfort model exclusively developed by SoundlyFit. This brings human perception into the picture by calculating the rating a human test subject would give the device being tested. We have developed our comfort model using an immense amount of knowledge collected from perceptual quality models used in audio and visual quality science. This allowed us to achieve very good results with a robust and generic comfort model. We matched our comfort simulation models to actual human ergonomic test results, and we have achieved an extremely good match between the data of a group of people from human ergonomic tests and computer simulations of the same group of people.
A suitable test group of people can now be selected from a population of virtual ears based on, for example, age, ethnic background, or gender. Once material properties have been applied to the 3-D CAD models of the new earbud and tip design, the virtual ergonomics test can be conducted very smoothly/effortlessly using SoundlyFit’s simulation process.
Technically, the large deformations, multiple contacts between the parts, and nonlinear materials all represent major challenges in FEM simulations of a mechanical behavior. Expressed simply, the ear is too complex a mechanical system to just build and fix with any FEM solver. We have run simulations with the state-of-the-art Dassault Systèmes SIMULIA brand Abaqus/Explicit solver. On top of that, we have built a system that automatically selects and creates ear-earbud-tip combinations for simulations. We have optimized the simulations in various ways such that now by using 30 computer cores, one simulation case can be run in about 45 minutes using a workstation grade computer. Effectively, this means that with one such computer the simulations of the ergonomics of ten people (20 ears) can be completed/run overnight and, naturally, just scaling up the server count enables extending the studies to tens or even hundreds of people within a few days. Therefore, since the execution time of the simulations used to be orders of magnitude longer when our initial working models were developed, we can now consider SoundlyFit’s physics-based realistic simulation to be computationally fast.
SoundlyFit’s simulation is a leap towards precision and squeezing down test time—we can pinpoint exactly what the problem is in the current design. The simulations show the specific point or areas on the earbud design that creates the most discomfort. We can choose target user groups—a blend of genders, geographical regions, small, medium and large ears—and identify the specific design issues for each of them. A virtual ergonomics test for 30 people and 60 ears can be conducted in about two days with one simulation workstation. Consequently, together with the designer, we can test and improve the design in a cycle of a few days, which is more than ten times faster and ten times easier compared to the human tests. If desired, we can even specifically design earbuds for a specific user group, be that female, male, children, people with small or large ears, or people in specific countries.
Imagine knowing exactly how comfortable your product is. Imagine knowing the comfort of your competitor’s products. Imagine improving your products to surpass the best of your competitors’ products and doing it fast.
With SoundlyFit’s Realistic Comfort Simulations all this is possible!
Comments