Introducing the Auren: High Performance Hearing Probe

Introducing the Auren: High Performance Hearing Probe

Measuring how our ears are changing or are being damaged is tricky! While audiometry is a great tool for tracking changes to hearing, you need a cooperative patient and it does not capture everything. Instead, researchers are developing newer probes and test procedure to improve diagnoses. These new procedures try to get a better sense of what part of the ear is causing problems.

The new probes make a airtight seal with your ear canal, and you don't have to do anything but relax, and listen to the soothing sounds of... clicks... and... chirps... and... ok, maybe not relaxing but you might pretend they sounds like birds. This is where the problems start for the hardware engineers. They have to deal with problems like standing waves and noise (electrical, mechanical, and even biological). Standing waves mean that you might not be exciting the tympanic membrane with the energy you intend to, and noise issues mean you might not be able to pick up the low amplitude sounds needed for an accurate diagnosis. Developing a commercial probe that can do all these things is expensive! That means that researchers can typically afford only a handful of units which means smaller studies. And often the data and processing is a black box. That makes it hard for researchers to develop new, better tests. If only there was an open-source platform that can make high quality audio recordings and is completely open. Oh right the Tympan! Now all we need is the probe part, and that's where the Auren comes in. Let's dive a little deeper to understand what the Auren is all about.

The goal.

Here's the plan. We're going to play a sound using a MEMS (micro electro-mechanical system) speaker. Then, we'll measure the amplitude of the pressure wave going into the ear. Good so far. Then we'll also measure the amplitude of the pressure wave coming back FROM the ear. Wow. That way we can tell how much of the sound was absorbed by the ear, and how much was reflected. Depending on how much sound comes back, audiologists can diagnose a number of different middle and inner ear problems. In fairness, this only describes the Wideband Acoustic Immittance (WAI) test, but some of the same issues are relevant. Okay, so how do we do this?

No standing around, get back to work!

Let's pretend to be physicists and make the problem a lot easier. A hearing probe with an airtight seal to your ear canal is basically a perfectly cylindrical closed cavity... right? Basically a tube. And sounds in that cavity are plane waves that only change along the tube. So, the sound pressure and velocity are the same anywhere over a cross-section of the tube. Well, the first problem you run into trying to measure sounds in a closed cavity are standing waves. MEMS microphones are great, but they only measure pressure. With standing waves, there are locations, or frequencies, where forward traveling waves and reflected waves interfere with each other, destructively and constructively. In some places, the forward and return waves can completely cancel out. That leaves our microphone with no pressure change to measure. In essence, a single microphone cannot tell the forward and reverse plane waves apart.

If you want to learn more about plane waves in a tube, there are many resources on the web. Here are a few of our favorites:

Working overtime.

To get around the problem of standing waves, we can't just know one pressure at one microphone, we need to know two things. Most of the commercial hearing probes do a very careful calibration of their speakers, which means they know the amplitude of the forward wave, and by measuring the pressure they can tell the amplitude of the return wave. It's a bit more complicated than that, but it's called a Thévenin calibration. Some newer probes instead use very neat velocity sensors -- they measure the velocity and pressure at one spot in the tube. The Auren takes a different approach: it uses 4 MEMS microphones to measure the pressure at 4 different locations in the tube. Wait a minute... I thought we only needed to know 2 things? Well, it's good to have redundant information and MEMS microphones are pretty inexpensive. And more is always better! Ok, it turns out that it's helpful to have 3 mics during calibration, and we included a fourth mic near the speaker entrance so that we can do a Thévenin calibration, so we can later compare the two approaches.

Let's try it out!

The above image shows the first prototype Auren. Yeah, pretty ugly. But it was good enough to make sure the concept works. The clear tubes near the bottom of the image are calibration/validation cavities. We'll talk more about calibration in a later post where we'll walk you through the math. The plot below shows the pressure amplitude predicted by the Auren compared to a reference microphone. Ideally, the blue and black lines should be the same. Looks pretty good up to 3 kHz! Good enough, let's make a new, better version.

The first, "real" Auren.

The image above shows the "deconstructed" Auren. It's a PCB (printed circuit board) with 4 MEMS microphones, 2-USB ports, 2 MEMS speakers (not shown), and a 3-D printed enclosure. There are 2 speakers so that you can play two different pure tone frequencies needed for DPOAE exams. You also need 2 USB cables that have been "slightly" modified, and we like a flexible gooseneck wire to position the probe where we like it. The plot below looks a lot better!

 

What comes next?

We're still working on fully calibrating the probe, and we'll go over our process in a future post. Being an open-source device, we plan to eventually post the CAD files, PCB layout, bill of materials, calibration code, calibration data, the kitchen sink, ... well everything. As a preview, the above image shows a cross section of the Auren with some dimensions. After calibration we're planning to implement two tests: (1) Wideband Audiometric Immittance (WAI) and (2) swept Distortion Product OtoAcoustic Emissions (DPOAEs).

We think researchers, tinkerers, and other fine folks like you, dear reader, will be able to use the Auren and Tympan to do things like:

  • DPOAE
  • Swept DPOAE
  • WAI
  • Wideband Middle Ear Muscle Reflex (MEMR)
  • Manual Audiometry
  • Custom tests

For now, we're looking for feedback and suggestions. If you have anything to share, head on over to the Tympan forums and let us know your thoughts.

To Tympan Blog