Note: If you are skipping the ear canals, you can also skip the calibration step.

Here's the one and only catch with using ear canals:

If you think about it, we basically have a microphone at the end of a tube. We all know how things sound when you listen through a tube. They sound, well... tubular.

Considering this, modeling ear canals means that we can't just listen back to the raw recording immediately. We have to process the audio in a way that removes the effects of the tube. Otherwise, the recordings will sound harsh and unnatural.

There are several ways to do this, but I've found a fairly easy way to do it using something called EQ matching. The end goal is to flatten the peaks and valleys of the frequency response for each of your dummy-head ears.

There are some EQ plug-ins and stand alone programs that have the capability to analyze the frequency spectrum of a sound file. They can then apply the inverse EQ to the audio, which results in a flat frequency spectrum. Many times this is called EQ matching. Logic Pro has an included plugin called Match EQ that does exactly that.

As you can see on the left, first I'm "learning" the frequency response of an example audio file.

Once I've learned enough, I apply the inverse EQ by clicking Match. You can see that this is simply (or not so simply) the inverse frequency response. If I were to analyze the output frequency spectrum of this "matched" EQ plug-in, it would be flat.

Audibly speaking, this creates a much more pleasing and natural sound. Keep in mind that you may need to play with the apply level of EQ Matching. I find that applying the inverse EQ at about 50% (right slider) gives a good clear and natural sound.

If you have EQ matching capability, then "calibrating" your dummy-head will be a simple, one-time process.

What you'll need to do is generate a white noise audio file and play it back over a single loudspeaker. Place the loudspeaker directly in front of your dummy-head at about 4-6 feet away. Play the white noise and record both Left and Right channels with the dummy-head. Try to get a good strong gain from the dummy-head by turning up the noise volume. Also, do your best to control reflections in the room by hanging blankets or sound baffles on the walls.

Normalize the recordings to 0 db. Apply EQ matching to each of the Left and Right channels and save those presets for future use.

Now, everytime you make a binaural recording with that same dummy-head, all you need to do is apply the same presets for EQ Matching and you are good to go.

You might want to custom tune the sound to suit your needs. You might boost the lower mids to give some presence to the sound. With my setup, I also have to cut the highs (> 15kHz) to get a more natural sound.

If you don't have automatic EQ matching you're not completely screwed... just partially.

I haven't been able to find a free EQ plug-in or program that has EQ Matching. If somebody knows about one, please inform me so I can post it here.

Here are the steps you need to follow. You can analyze your frequency spectrum with free software such as Spectrum Player or another, similar frequency analyzer. The interface will look similar to the one shown left.

Capture the white noise using your dummy as described above. Load the files into your spectrum analyzer and see what they look like. Take note of the peaks and valleys and the relative dB change between different frequencies.

Then you need to go into your audio editing software and manually apply the inverse EQ to each of your left and right channels. This may take a little longer, but it's not too bad to get a natural sounding result this way. Save your presets when you're done so that you never have to do it again!

Example Recording

This video was made using a dummy head like the one shown in this tutorial. The ears are DigDagga's SBE-1 Silicone Binaural Ears.