Skip to content

A good enough microphone example

Microphones! Who doesn't want a good one? Although digital voice processing has gotten more advanced over the last decades, most of you know the unfortunate reality of echoes that aren't completely cancelled, unsatisfactory or greedy noise cancelling and so on. At their core, these problems come from digital processing being unable to cope with very compromised recording setups. Providing decent recordings feels like politeness or hospitality to my call-mates. Commercial solutions are either expensive, or of unknown quality. Time to build one!

Well, let's hold our horses. What's the precise goal here? I wanted a microphone that could pick-up a decent recording. I'm not recording podcasts or concerts so I do not need a huge dynamic range. The microphone does not need to be portable or extremely rugged -- I'm mostly recording from one fixed location in my home. However the whole point is to have a plug and play experience with internet telephony, so I would like something that integrates seamlessly with my computer. Computers have microphone ports, I should be able to plug things in and have them just work (tm).

The microphone itself

The shell of my microphone comes from an old speaker, bought at an amateur radio flee market. Once the (broken) speaker membrane is ripped out, it's an empty shell with a grille, a hole for a wire, and a bracket for easy mounting with an adjustable angle. It bought it because it had plenty of space for messing around, before having a precise idea of what to do with it.

Once the project was underway, it was time to decide how to usefully stuff this shell. The two main designs of pickup are the following:

  • a dynamic microphone acts like a reverse speaker. As you speak into it, a membrane vibrates next to a magnet, which a coil on the membrane picks up as a current.

  • a static, or here electret microphone uses a tiny capsule. Simplifying, this capsule holds a crystal that, when crushed by sound waves, creates a voltage. This tiny voltage is amplified in the capsule itself by a transistor.

You may know electret microphones from their application in, well, everything from phones to headsets. Dynamic microphones are generally confined to studio equipment. This is because they're larger and need a preamplifier to generate a decent signal level. On the other hand, they offer (for voice signals, at least) a better signal-to-noise ratio and less distortion. However, by-and-large computers expect static microphones, and have the circuits to ensure their proper operation. Building a compatible dynamic preamp seemed too complex and brittle, for too little reward. So, I went with an electret design. 1

How to mount the capsule inside the shell? At first glance, I'd need to design a standoff, maybe 3D print something... Or use cardboard, like I did :). Easy, cheap mounting with some damping.

Just because we use the boring pickup design doesn't mean we can't have a bit of fun! I used two electret capsules. Sound waves coming from the sides of the microphone hit the two capsules with a delay. By summing the capsules' outputs, interference means some frequencies and angles don't get picked up. Long story short, at 10kHz the microphone theoretically doesn't respond at all to sound coming from the sides. (A so-called "figure 8" response). At other frequencies, the response isn't so neat, but overall noise coming from the sides is attenuated compared to sound coming straight-on.

The spacing of 1.7cm dictates the frequency where the response is a clean figure 8. I chose 1.7cm (and, therefore, 10kHz), because I expect most outside noise to be approximately white. Indeed, the noise sources when I'm at the computer are its fan and its mechanical hard drives. I believe the choice of 10kHz minimizes the overall picked up white noise across the audio band. But I haven't run the numbers.

Standing up

So far, I had a plastic shell that acts as a microphone. It already provided better quality than my laptop's integrated microphone, courtesy of not being in a mechanically and electrically noisy chassis. I could lay it on the desk, and speak to it. But, each halving of the distance between my mouth and the microphone would raise the signal-to-noise ratio by 6dB. Great gains can thus be had by having a proper mount!

Studio microphone mounts are complicated. They are made of multiple interlocking brooms and clamp to a table side. I quickly decided not to recreate that design, because it is too complex. It trades-off simplicity for ease of adjusting. My setup is going to be mostly stationary, so I decided I only needed one upright beam standing on a base. The only possible adjustment is the height of the microphone, in case I change desks for some reason.

Simplified doesn't mean featureless, and the base has a channel carved into it. This simplifies cable management when used with the brushed aluminum cover. The three screws hold the cover shut but are also the stand's feet. Three points define a plane, so it can't rock by itself.

The piece of wood holding the microphone cannot rotate, although it held by only one screw. That's because a groove is carved into it, which matches with the broom's profile. You can see there are multiple grooves here. A recent furniture-making adventure introduced me to the world of wood-carving so I was primed to experiment with it a bit more. Making the cross-holes in the broom was a first for me. To do them, I prepared a jig which holds the broom in place while also guiding the drill bit perpendicularly. This method is very well-known, especially in machining circles, but I had never done it.

microphone in its mount I was very happy about this setup, until it fell on its side! The very old plastic bracket promptly broke. Liberal application of superglue was only a temporary fix -- old plastic is very brittle. Thinking about it, I blamed two things:

  1. The stand could fall, if tilted enough. That's because its center of gravity was fairly high.

  2. While the microphone itself would be hard to break, its bracket is definitely a weak point.

The naive way to tackle point 2 would be to 3D print another bracket identical to the old one. But I wouldn't have a guarantee of it being much stronger, and CAD is boooring. However, I was curious about metal bending. I wondered whether scoring a bracket template, then hammering it along a table edge would work. To give myself a chance, I picked aluminum, the bendiest metal I had available. And in the end, it did work! It definitely wasn't perfect, but for a first try it was forgiving. It did help that the bracket itself doesn't have many critical dimensions. That fixed point 2.

aluminum bracket, mounted

For point 1, I remembered I had bought (at that same flea market) a pop filter. It's meant to be mounted on those professional microphone stands. But my stand also has a broom, meaning I can mount the filter on it! That lowered the center of gravity dramatically, while providing some shielding to the microphone it case the stand still falls. And, of course, it filters pops.

microphone mount with pop filter I can't talk at length about a microphone project without some audio samples, so here you go! These two recordings were made separately either with my current setup, or my laptop's integrated microphone. I won't claim I achieve anything near studio quality, but that wasn't the point.

There is one major downside to the external microphone. It tends to pick up electrical noise from my phone somehow. I think my phone is particularly noisy, and the shielding on the new microphone is worse than on the integrated one. I only need to remember to turn off my phone before calling, which does not bother me. One day, I may investigate using proper shielded microphone cable.

A minor downside comes from the directionality. Since the microphone is slightly directional, I need to speak into it. If I don't, the received signal is much lower-level. This risks not triggering the voice detection, and in any case my voice becomes distorted by the very non-flat frequency response. It's a skill to pick up for sure, but not a deal breaker (and I suspect, similar skills are required in studio situations.)

Projects as intersections

I'm happy about this project. I now have a more than good-enough recording setup that isn't too fussy and I got to learn about multiple crafts. Reflecting, I figured that for me fulfilling projects usually stem from an overlap of three elements:

  1. Projects need to have a purpose, and fulfill a need. Needs which could be considered frivolous (eye candy, toys...) are valid, making something only for academic/intellectual purposes is straddling the line. I hate having projects gather dust as soon as they're finished.

  2. Projects need to provide opportunities to experiment with new techniques, or challenge me on old ones. If a project ends up being implementation only with the design feeling trivial to me, it can become boring. An exception is made when doing the project is, by itself, fun or relaxing. This includes things like hand sewing.

  3. Projects should be (mostly) doable now. Taking bits and pieces I already own and assembling them together is fun. Shopping for precisely the right components and tools is not. Most of the time, there are multiple ways to achieve a project's goal, so this isn't as limiting as it seems and provides an interesting design challenge. This does mean I will stock up on things I do not need right now, if I think I will be able to use them for... something later on. Drawing the line between "useful component that I'll be able to use later" and "(e-)waste" is a bit tricky, I believe most hackers learn that skill over time.

My criterion evolve over time and are oftentimes implicit. However I think it's good to be aware of them, since they help filter project ideas that would in the long-run not be satisfying to work on or are doomed to be abandoned part-way. I'm curious about yours, so feel free to reach out!

  1. High-grade consumer microphones are almost always dynamic I believe. That's why they're powered by USB -- they need to power their preamplifier. And also why they're expensive. ↩