Binaural Recording, Pt 2

After trying the commercial microphones in Part 1 of this series on Binaural Recording, I thought I should try something more home-made – although this also involved a large-ish initial purchase.

What I bought was this handsome life-size mannequin head, intended for work in a hat shop:

The idea, of course, was to install microphones in the ears of the dummy head.  It was made of a fairly hard, but not too brittle, plastic (PVC, I believe it said in the eBay listing), which seemed to be a couple of millimetres thick.

There were several things I particularly liked about this style of head: first of all, the realistic appearance – the whole point of binaural recording is realism, so the closer the recording device resembled the human head, the better.  Unlike some mannequin heads, however, this one wasn’t painted to look like a real person – that would be too spooky! . . .

In particular, the ear was quite well-fashioned:

A big part of the way we hear things is because of the size and shape of our ears, so the accuracy of the ears of the dummy head would have an effect on the quality of the recordings.  For similar reasons, some dummy heads for recording include shoulders, as sound will bounce off these into the ears.

Finally, the underside of the base had a socket which would make it possible for the head to be mounted on a pole or stand, so as to be set at an appropriate sitting or standing height,  whichever was required for a particular recording situation.

*

My task in this case was essentially to drill suitable holes in the mannequin’s ears, and insert a pair of electret capsules.  I began by soldering a pair of capsules to the cut ends of the twin phono lead I had left after removing 10cm of one end for the previous project.

The capsules were like these:

The lead with the three connections to the side of the capsule is the Ground lead, the other is the signal.  I connected the two capsules this way, with some shrink tubing to make the joints stronger and stop them short-circuiting.

Turning to the mannequin head, the base was only attached by a dab of glue on one side, so came off easily with a little twisting and a cut with a craft knife.

It was a fairly quick procedure to drill 3 holes in the head: one in each ear, slightly smaller than the size of the electret capsules, and another, larger one at the back for the leads to exit from:

I pushed the lead in through the hole in the back, ran a big blob of hot glue round the front edge of the electret capsules and stuck them just behind the ear holes.  I chose hot glue as it’s easy to remove in case the capsules need replacing at some point in the future; it didn’t matter if a bit of the glue came over the front edge of the capsule as the actual hole on the front which the sound goes in through is very small, just a millimetre or so, right in the middle of the capsule.

Looking inside through the base, you can see how the electret capsules are stuck inside the ear, and the cables are held in place with more hot glue:

This took only a matter of minutes, and the final result looked like this:

As you can see, the microphones are held discreetly in the ear holes, and the twin phono lead, which connects to the preamp, exits from the large hole in the back of the neck, where it’s held in place by further hot glue.

*

The cost of this project was £10.50 – about half as much as the first one.  The mannequin head was £9.50; the electret microphone capsules about 25p; and the half phono lead was 75p.

The only other thing to consider is whether the head should be filled – and, if so, what with – to more accurately reflect the fact that human ears are separated by more than air.  The human brain is about three-quarters water and has the consistency of jelly or tofu; it’s quite heavy, but soft and squishy, and you can’t really pick it up until it’s been preserved in some way, which most brains we see pictured have been.

So what the best thing would be to fill the head is difficult to decide, given that jelly or tofu would soon go off.  In one article that I read the dummy head maker installed the microphones then filled the head with liquid silicone, which gradually set solid.  That seemed to be a good plan, although there’d be no way of getting to the microphones again if there were a problem with the capsules or the wiring.  My thinking is it would be sufficient to use something sound deadening, like wool or felt, to ensure that the microphones would only be picking up sound from outside.

[Edit: This is what I did, filling the empty head with a pyjama jacket, which I had been bought but had never worn].

*

In the third part of this series, I’ll complete the final project, do some recording and compare the results.

In the meantime, here’s a short extract from the first recording I made with the head:

Binaural Recording, Pt 1

I’d done some field recording with conventional microphones, and after recording with contact microphones and hydrophones, as described in this post, I decided it would be interesting to try binaural recording.

Binaural recording is an attempt to record in as likelike a way as possible.  Since both the size and shape of our ears and the fact that they are placed on opposite sides of our head are important factors in establishing the quality of the sound we naturally hear, binaural recording attempts to replicate this by, most usually, placing microphones within the ears of a dummy head.

‘Lifelike’ aspects which could be captured by binaural or dummy head recording include time differences in the arrival of sounds at one ear or the other, and types of frequency-dependent level differences and distortions which vary with the direction of the sound source.  These would allow a listener using headphones to gain extra information about the precise location and distance of sounds they were listening to; information which would not as readily be apparent if the recordings were made with conventional microphones or played back via loudspeakers.

This article [note: it’s a pdf] refers to three ways in which binaural recordings preserve ‘cues’ as to sounds’ direction and location.  These are, in order of importance: the shape of the ear; the time difference between sounds arriving at one ear, then the other; and, least significantly, it says, the presence of the bulk of the head between the two ears.

(This isn’t a universally-held view: a number of binaural recording devices feature two microphones, side by side, and separated by a sound-absorbing panel.  A notable example of this is the Jecklin disc, which has a diameter of 35cm, is covered with sound-absorbing foam or fleece, and placed between two omnidirectional microphones).

I experimented with three ways of creating binaural recording devices, comparing the results in terms of quality, practicality and cost.  I intended to use the pre-amp I had originally designed for use with electret capsules, and which I had built into a handy case when developing an inductor pickup, so I didn’t include this in the cost.  The pre-amp – which in any case was very cheap and simple – looked like this:

and the case like this:

The preamp is stereo, so inside there are two of the circuits above.  The inputs are phono sockets; the output, a 5-pin XLR, is compatible with the connecting lead of my recording device, a Marantz PMD-660.  The 3.5mm mono socket was for a 9v power source; the velcro on the top of the case was to mount the holder for a PP3 battery.

*

The first, and simplest, method involved purchasing the microphones!  I found that Roland made a very useful-looking pair of microphones resembling ear-buds, the idea being that you would wear them while recording – no need for a dummy head when your own head could do the job!   This set (CS10-EM) sounded as if it would be particularly effective, as the microphone earpieces also contained earphones, enabling recordings to be monitored while being made.  The downside, however, was the cost: over £70 on Amazon – a reasonable price, I suppose, for a potentially very useful pair of microphones, but not in the price-range for projects in this blog.

However, I noticed that Roland also made a similar pair of microphones for use with GoPro cameras, the WPM-10 WearPro:

Although this set didn’t have actual earphones for monitoring, the cost was considerably less: just under £20, including postage; so I invested in a set.  When it arrived, it contained a choice of different-sized earpieces, so it was possible to select the best fit for your ears; and, as can be seen from these drawings, it also included a pair of foam covers that might have an effect – albeit a small one – on pick up of wind noise.  Having fitted them, it didn’t look to me as if they would stay on for very long, though, so I didn’t plan on using them.

Because of their intended purpose, the plug on these microphones is not an audio plug, it’s a mini USB.

Except that it isn’t a simple mini USB plug at all.  A standard sized USB plug has 4 pins; a mini- or micro-USB has 5; this one is the size and shape of a mini-USB, but has 10.  I suppose it could be called a proprietary connector, as a number of manufacturers use them, but not always in the same way.  GoPro uses them like this:

The connections in the row along the bottom are the standard mini- or micro-USB set: +5v, Data-, Data+, ID and Ground.  ‘ID’ is the one omitted from the standard-sized USB connection: using standard-sized USB connections, the ‘in’ socket on a host device (e.g. a laptop) should be the narrow, rectangular one, Type A; the ‘out’ socket on a peripheral device (e.g. a printer) should be the square one, Type B.  With mini- and micro-sized sockets, there is no distinction in shape between these ‘in’ and ‘out’ sockets, so the job of distinguishing them is done by the ‘ID’ pin.  If the socket is performing the job as a peripheral – e.g. a camera, the ID pin is not connected; if as a host, e.g. a laptop, it is connected to Ground.

In the case of the GoPro, the device responds differently to different resistances between the ID pin and Ground.  As shown in the diagram, a resistance of 100k between ID and ground causes the device to function as a video and audio source, and it can be plugged into an external video/audio receiver; a resistance of 330k, and it will receive signals from a microphone.  In both cases, the presence of a resistance between ID and Ground allows the upper 5 pins to come into operation; their specific use is shown in the diagram above.

(I have seen it suggested that a resistance of 33k allows both these functions at the same time, but that was not confirmed by experiment in the article where I read it).

The reason for the resistor shown with a dotted line is that the conventional place to connect the 33k/100k/330k resistor would be the case of the plug, but one experimenter who posted a video on YouTube had difficulty with this, and used the Ground pin in the centre of the upper level instead and confirmed that this worked fine.

*

However, I don’t have a GoPro, and this is not the way I intended to use the microphones – in fact, it was the exact opposite.  What I needed to do was remove the mini-USB plug altogether and attach instead two phono plugs, so that the microphones could be used with the preamp shown above, for recording on my Marantz machine.

I was banking on the fact that these microphones would be a pair of electret capsules, and would receive sufficient power from the preamp to operate correctly.

So, as I had done before, instead of buying a pair of separate in-line phono plugs, I bought a 2m twin phono lead – it was only about £1.50 and would, when chopped in half, make two single-ended twin leads.

I took one of the leads and the WearPro microphones:

and cut off the mini-USB plug from the microphone lead.  As expected, this left two signal leads, Red for right, Yellow for left, and two unenclosed ground connections.

I didn’t need a whole metre of extra cable, and my next project would require as much of the 2m as could be spared; so I cut off one pair of phono leads with about 10cm of cable.  In this case the two signal leads were red and white, and the two grounds were black.

All I had to do was connect these 4 leads together, red to red, yellow to white and black to copper, and test the microphones with my recorder.

The test was fine, the microphone in my left ear was recording to the left channel of the recorder, the microphone in my right ear was recording to the right channel; so I sealed the wiring with shrink tubing and duct tape.

I used duct tape because the quality of the electrical insulating tape I’ve been coming axross recently has been terrible: neither flexible enough or sticky enough.  After I took the lower picture I added an overall layer of duct tape to bind the two wires securely together.

*

The cost of this project was £19.75 – the microphones were £19 and half the phono lead was 75p. Afterwards I felt slightly guilty at being lazy and buying the microphones.  It would have been possible to buy a pair of headphones or earphones and adapt them by gluing electret capsules on the outsides and transferring the wiring from the speakers to the microphones.  The electret capsules would have cost no more than about 20p each, and the price of a pair of not-very-good headphones or earphones would have been minimal, so it could have been done for a quarter or a third of the price.

However, this was a neat solution, took only a few minutes to finish, and the resulting set up looks good.  The quality comparison of my different binaural systems – and sound files – will come later, after I’ve finished all three projects, but this one is going to have the advantage when it comes to practicality, as it’s very simple and discreet, certainly the best solution for situations in which I don’t want to be advertising the fact that I’m recording.

This was my first recording with these microphones. I started in the hallway, walked out of the house into the shed, out of the shed, brushing past some bushes, picked up and started filling the watering can from the water butt, then turned and walked back towards the house. No noise reduction has been applied to the recording, which shows you can get a decent clear sound from the WPM-10s.

As soon as I had completed the project, I took the remaining part of the phono lead and moved on to the next one, which is described in Part 2 of this series, here.