Dummy Head Recording - a flaw I notice...

[danbronson]

I see some dummy heads being used where the mics are positioned where a person's eardrums would be, actually inside the head a bit. The idea is to fake the ear canal and give it's effect. The Neumann pictured below even models the entire ear as well.

This should only be appropriate if you're positioning the speaker directly over the eardrum though, which never happens. Maybe this would help if the listener is using ear-buds (but I'm not aware of any earbuds with a speaker that far down the ear canal), but certainly for over-the-ear headphones, this will only make things less realistic since what's coming out of those headphones is going to go down your ear canal anyway.

So it seems to me, the most realisic way to record using a dummy head is to position the mics not where the eardrum will be, but where the speaker will be, just outside the head maybe even up to an inch, facing straight out, which is the opposite direction the speaker will face. Or just barely inside a fake ear, where the earbud will be. Remodeling the entire ear and ear canal will actually make things less realistic. Unless I'm missing something...?

[John Willett]

Originally Posted by dylansmale

I noticed that the ears look like hard plastic just yesterday and thought "wouldn't that **** up the sound before it actually even touches the microphone?"

because, well...our ears aren't made of plastic...

The ears on the Neumann are soft and pliable with the same consistency of real ears - they are not hard at all.

I have also heard that putting a wig on a dummy head helps with sound location when played back - not tried it though.

[Blast9]

A mullet wig gives it the authentic German sound

[Eganmedia]

Originally Posted by Blast9

A mullet wig gives it the authentic German sound

I thought it gave it an authentic Southern Rock sound...

Guess it's where your mullet comes from...

[Fabi]

Originally Posted by danbronson

Remodeling the entire ear and ear canal will actually make things less realistic. Unless I'm missing something...?

I asked myself the same question, but then just listen to these recordings (on headphones of course) and it sounds sooooo good.

I heard (but I realy don´t know) you can make prints of your own ears and use them on the "fritz"

Now how cool is that?

Put the dummy in the sweet spot and stream audio to your headphones at the beach, giving orders to the assistant.
This way you can mix everywhere and sunbathe at the same time
And your ears are in the studio

Or you make ears of your favorite Mastering Engineer, then asking him constantely if that Mix would work for him

[theblue1]

Originally Posted by Fabi

I asked myself the same question, but then just listen to these recordings (on headphones of course) and it sounds sooooo good.

I heard (but I realy don´t know) you can make prints of your own ears and use them on the "fritz"

Now how cool is that?

Put the dummy in the sweet spot and stream audio to your headphones at the beach, giving orders to the assistant.
This way you can mix everywhere and sunbathe at the same time
And your ears are in the studio

Or you make ears of your favorite Mastering Engineer, then asking him constantely if that Mix would work for him

Um... I guess it would work -- as long as, when you are actually physically on the job, you never move your head, even slightly.

That, as I see it, is the one structural/theoretical problem with binaural recording.

Just as human eyes are in constant motion to "refresh" the optic image (if they were not, we, like dogs and cats and most 'lower' animals, we would only see things when they were moving in our field of vision [or, of course, when we were moving]) most of us subconsciously (or sometimes otherwise) are constantly moving our heads as we carefully listen to sounds.

Since we know from testing that very small movements within many auditory environments can make large changes in what we hear (cfr Ethan Winer's paper on test mic placement and large frequency response shifts from small placement changs), it stands to reason that our impression of sound is formed over time and presumably, for many of us, many small head movements that give us different audio 'perspectives.'*


But a dummy head, God love it, is just gonna sit there like it was carved out of... uh... well... whatever they make those things out of, now.

So it's great in many ways, when done right, but still not the same as being there.


*I'm thinking it's a little like the problem with convolution reverb. A convo IR is like a 'snapshot' that doesn't reflect the constant changes in a typical audio environment. As we listen to the convo reverb we can become aware of a certain quality to the reverb typically produced that is, again, less than entirely satisfying. (Surprise, surprise. You want a good room sound, play in a good room. I don't have one myself but... )

[danbronson]

Hmm...

In that case I've decided to build a hat thingy which I can use to mount my little omnis next to my ears. I'll walk around the room while a band is playing. Or maybe record a conversation with someone else. It might be cool to record myself playing some acoustic guitar and singing. Could be very interesting...

[MarkRB]

Some of that binaural stuff can sound amazing but what people tend to forget is that the sound once recorded has to get out of the speakers and into our ears again, thereby doubling the effects of the "ear canal".

[FollowTheMusic]

So-called "diffuse-field" equalized headphones compensate for the ear canal resonance to give an (ideally) flat response from the transducer to the eardrum. (This is done by averaging over all directions, and using the inverse as the EQ. This removes non-direction-dependent components). This avoids the "double-ear-canal" issue you're describing. Many headphones are diffuse-field equalized, because it often sounds better even with non-binaural recordings.

Alternatively, some binaural recordings are in fact made at the entrance to the ear canal, rather than at the eardrum. Signal processing and HRTF libraries make it possible to convert between the two methods. Of course none of this is perfect, and doesn't address dynamic (head-motion) cues. That gets you into head-tracking...

[andychamp]

Dummy-head recordings are intended to be heard with headphones.
They will not only give you left-to-right stereo placement, but also front-to-rear and up-down localization.
Dummy-head radio plays were a big thing in the late 70s/early 80s and I've never heard anything as realistic since, I mean: f**k 5.1!

[theblue1]

Originally Posted by danbronson

Hmm...

In that case I've decided to build a hat thingy which I can use to mount my little omnis next to my ears. I'll walk around the room while a band is playing. Or maybe record a conversation with someone else. It might be cool to record myself playing some acoustic guitar and singing. Could be very interesting...

Do not forget to mimic the nodding head motions of someone drifting off while listening to a tired chestnut of a classical favorite for the umpteenth time...


And on your hat thingie... you'll need a mod for tracking southern rock with it -- two PBR cans dangling from it with curly tube straws...