Why don't headphones work like this?

Archived from groups: rec.audio.pro (More info?)

In the latest issue of TapeOp, there was an interesting article about
stereo miking using a fake head. It was interesting, but would seem to
only work if you listened to the recording via headphones.

The article got me thinking about the difference between listening to
something on headphones, and listening to something using speakers.

Clearly, most music is mixed to be listened via speakers. And clearly,
when you listen via speakers, you can hear some of each channel with
each ear. This isn't true for headphones (at least closed ones).

Why, then, don't we mix some of the right channel into some of the left
channel (and vice versa) before outputting the signal into the
headphone jack? Or is this done already?

I realise that there's more going on than just level (phase etc) but
wouldn't this at least make the recording more realistic in terms of
pan?

Archived from groups: rec.audio.pro (More info?)

huwgareth@my-deja.com wrote:
> In the latest issue of TapeOp, there was an interesting article about
> stereo miking using a fake head. It was interesting, but would seem to
> only work if you listened to the recording via headphones.
>
> The article got me thinking about the difference between listening to
> something on headphones, and listening to something using speakers.
>
> Clearly, most music is mixed to be listened via speakers. And clearly,
> when you listen via speakers, you can hear some of each channel with
> each ear. This isn't true for headphones (at least closed ones).
>
> Why, then, don't we mix some of the right channel into some of the left
> channel (and vice versa) before outputting the signal into the
> headphone jack? Or is this done already?
>
> I realise that there's more going on than just level (phase etc) but
> wouldn't this at least make the recording more realistic in terms of
> pan?

Something along the lines of what you are talking about is being done
by the Headroom corp. http://www.headphone.com

However there are engineers, such as Michael Bishop at Telarc, using a
dummy head microphone setup as a main stereo pair. Then spot mics are
panned to match the aparent position of the acoustic sources rendered
by the binaural setup. Seems to work pretty well.

Karl Winkler
Lectrosonics, Inc.
http://www.lectrosonics.com

Archived from groups: rec.audio.pro (More info?)

> You have just re-invented the shuffler. Headwise makes a nice one.

And when applied to speakers, it's called "spatial stereo"TM or
something.

Archived from groups: rec.audio.pro (More info?)

On 6/3/05 9:56 PM, in article b55e0$42a10a4c$186009c6$9641@KNOLOGY.NET,
"Dave" <davemor@knology.net> wrote:

>
>>> In the latest issue of TapeOp, there was an interesting article about
>>> stereo miking using a fake head. It was interesting, but would seem to
>>> only work if you listened to the recording via headphones.
>>>
>>> The article got me thinking about the difference between listening to
>>> something on headphones, and listening to something using speakers.
>>>
>>> Clearly, most music is mixed to be listened via speakers. And clearly,
>>> when you listen via speakers, you can hear some of each channel with
>>> each ear. This isn't true for headphones (at least closed ones).
>>>
>>> Why, then, don't we mix some of the right channel into some of the left
>>> channel (and vice versa) before outputting the signal into the
>>> headphone jack? Or is this done already?
>>>
>>> I realise that there's more going on than just level (phase etc) but
>>> wouldn't this at least make the recording more realistic in terms of
>>> pan?
>>
>> This is a long-understood difference between BINAURAL recording/playback
>> (earphones only) and how that can aply and be adapted to speakers. Most
>> binaural recordings actually are pretty good on speakers. It's the essence
>> of both the ORTF, Crown SASS, and Jecklin Disc approaches as well.
>>
>> POLK used to (still does?) make a series of speakers that used
>> cross-channel
>> reverse-phase midrange drivers to cancel out the cross-channel acoustic
>> information at the ear and make speakers sound more like headphones.
>
> Was this the same system that Carver used to use with their Holographic
> Generator?

Nope... POLK's was entirely mechanical, each speaker system simply had a
second midrange driver that was connected to the OPPOSITE channel and
displaced enough to match the ear-to-ear distance. The result was a
sweet-spot cancellation at each ear of significant directional information
from the channel opposite that ear.

Archived from groups: rec.audio.pro (More info?)

huwgareth@my-deja.com wrote:

> In the latest issue of TapeOp, there was an interesting
article about
> stereo miking using a fake head. It was interesting, but
would seem to
> only work if you listened to the recording via headphones.

As Bob Cain pointed out, this is all about HRTFs. In the
case of fake head recordings, the head applies a HRTF to the
acoustic signal that reaches the mics in the dummy's ears.
If you listen through speakers, the HRTF of your head gets
added to the existing HRTF that was applied to the recording
by the dummy head, and the result is often a less than
natural sound. If you listen through headphones, the HRTF
of your head is bypassed, and the result is often a more
realistic sound.

> The article got me thinking about the difference between
listening to
> something on headphones, and listening to something using
speakers.

Right, when you mic conventionally, and playback through
speakers, then your HRTF gets added, for more natural sound.
However, there is some vagueness that slips in via the
reconstruction of the sound field that the mics pick up, by
the loudspeakers.

> Clearly, most music is mixed to be listened via speakers.
And clearly,
> when you listen via speakers, you can hear some of each
channel with
> each ear.

HRTFs include the leakage of sound around your head to the
ear on the other side.

>This isn't true for headphones (at least closed ones).

Its pretty much true for most headphones, because even the
open ones don't project much of an external field. If they
did, I'd hear what you are listening to through headphones
in the room more clearly. As things stand, most open-air
type headphones project only a very weak sound into the
room.

> Why, then, don't we mix some of the right channel into
some of the
> left channel (and vice versa) before outputting the signal
into the
> headphone jack? Or is this done already?

Over the years, a number of products have been developed
based on this idea. Some of the newer ones are
fully-cognizant of the more recent information that is known
about HRTFs.

> I realise that there's more going on than just level
(phase etc) but
> wouldn't this at least make the recording more realistic
in terms of pan?

There's a counterpoint to this. The human ear/brain system
can be very adaptable, and it is possible to become
acclimatized to some degree to straight-up headphone
listening, which is admittely different from loudspeaker
listening. Its mostly about time and being flexible and
adaptable. It never seems to happen for some people, but it
happens for many others.

Archived from groups: rec.audio.pro (More info?)

Come to think of it, if you want to get AudioMulch (free),
Voxengo's Pristine Space Light ($59.95), and Virtual Audio
Cable ($30.00) I've got some HRTF's recorded from Dolby
Headphones and modified to be a little flatter that I could
send you that would convert stereo to binaural.


Bob

Bob Cain wrote:
>
>
> huwgareth@my-deja.com wrote:
>
>> Why, then, don't we mix some of the right channel into some of the left
>> channel (and vice versa) before outputting the signal into the
>> headphone jack? Or is this done already?
>
>
> This works even better if what is called a HRTF (Head Related Transfer
> Function) is employed. That defines in full detail what happens to
> sound going to each ear from each speaker by measuring such a setup with
> litle mics at the entrance to the ear canals. It uses a process called
> convolution to map stereo to binaural using a HRTF calculated from such
> measurements. Dolby Headphone is one implementation of this that you
> could google for. It's quite impressive.
>
> Some think that another person's HRTF can't be used to create a good
> binaural experience but that is not my experience at all having done
> such measurements, processing and recording myself. For example if
> someone is listening to a demo binaural recording I've done with a dummy
> head mic in which there is a door slamming behind it, invariably the
> listener will spin his head around to see what happened and where the
> hell a door came from. Quite fun, actually. :-)
>
> This points out something kinda interesting. Often front/back
> discrimination is hard to achieve even with a dummy head _unless_ it is
> a startling, impulsive thing like a door slamming from behind. Then it
> seems to work 100% of the time. Go figure.
>
>
> Bob

--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

Archived from groups: rec.audio.pro (More info?)

What is the purpose of engineering pan displacement on a mix when you are
trying to cross fade etc..to get rid of it.
Might as well just have a mono mix and be done with it.
<huwgareth@my-deja.com> wrote in message
news:1117827911.942370.3030@g47g2000cwa.googlegroups.com...
> In the latest issue of TapeOp, there was an interesting article about
> stereo miking using a fake head. It was interesting, but would seem to
> only work if you listened to the recording via headphones.
>
> The article got me thinking about the difference between listening to
> something on headphones, and listening to something using speakers.
>
> Clearly, most music is mixed to be listened via speakers. And clearly,
> when you listen via speakers, you can hear some of each channel with
> each ear. This isn't true for headphones (at least closed ones).
>
> Why, then, don't we mix some of the right channel into some of the left
> channel (and vice versa) before outputting the signal into the
> headphone jack? Or is this done already?
>
> I realise that there's more going on than just level (phase etc) but
> wouldn't this at least make the recording more realistic in terms of
> pan?
>