Can "standard" Stereo be easily converted into Binaural?

[pablolie]

As I sit here listening to stereo through headphones, and enjoying myself... I was wondering... with all the DSP horsepower we have waiting for something to do these days...

Why can't my DSP-enabled headphone amplifiers simply convert stereo into binaural (or several simultaneous options thereof) on the fly? Why are binaural recordings singular things?

 

[staticV3]

Why can't my DSP-enabled headphone amplifiers simply convert stereo into binaural

They sure can. See e.g. the DX5II.

 

[pablolie]

They sure can. See e.g. the DX5II.
View attachment 512028

Wow. Does it work well?

 

[staticV3]

Wow. Does it work well?

I don't own the DX5II, so have not heard it.

Over on the DX5II review thread, people are praising the effect but also lamenting the strong drop in volume with it enabled. Guess it's good that the DX5II has so much Amp gain.

 

[myHRTF]

In principle, you are right: done properly, headphones can sound indistinguishable from real speakers in a room using digital signal processing (DSP). I'm solely listening to music this way and my speakers are used for watching TV with the family only. But unfortunately, it is not simple at all as outlined below, and this also explains why binaural (aka dummy head) recordings don’t work for the vast majority of users and recordings.

Individual perception of sound
Among other cues, the human auditory system uses tonal sound characteristics (timbre) to locate sound sources. The timbre is created by the interaction of sound waves with your torso, head and auricles, mathematically described as head-related transfer function (HRTF). Consequently, timbre is highly individual as is the size and shape of our bodies and auricles. Differences between individual HRTF can be higher than 20 dB (!) easily.
If the timbre doesn’t match the individual expectation, sound sources are perceived as being at the wrong location and too near or even inside of the head, an experience that is typically connected to generic HRTF solutions.

Headphone frequency response
Individual differences in HRTF are also responsible for the fact that different users perceive the sound from the very same headphones differently. In addition, frequency responses vary greatly between different headphone models. Both need to be taken into account, too.

Head rotation
When listening to headphones, the sound doesn’t change with head rotation, which is highly unnatural. For a small group of users it is not enough to match the individual timbre. They require head-tracking to match the expectation when rotating the head.

 

[respice finem]

In principle, you are right: done properly, headphones can sound indistinguishable from real speakers in a room using digital signal processing (DSP). I'm solely listening to music this way and my speakers are used for watching TV with the family only. But unfortunately, it is not simple at all as outlined below, and this also explains why binaural (aka dummy head) recordings don’t work for the vast majority of users and recordings.

Individual perception of sound
Among other cues, the human auditory system uses tonal sound characteristics (timbre) to locate sound sources. The timbre is created by the interaction of sound waves with your torso, head and auricles, mathematically described as head-related transfer function (HRTF). Consequently, timbre is highly individual as is the size and shape of our bodies and auricles. Differences between individual HRTF can be higher than 20 dB (!) easily.
If the timbre doesn’t match the individual expectation, sound sources are perceived as being at the wrong location and too near or even inside of the head, an experience that is typically connected to generic HRTF solutions.

Headphone frequency response
Individual differences in HRTF are also responsible for the fact that different users perceive the sound from the very same headphones differently. In addition, frequency responses vary greatly between different headphone models. Both need to be taken into account, too.

Head rotation
When listening to headphones, the sound doesn’t change with head rotation, which is highly unnatural. For a small group of users it is not enough to match the individual timbre. They require head-tracking to match the expectation when rotating the head.

Yes it is very individual, and some people will only accept headphones generally as a rather inferior substitute, while others will prefer them. I'm listening quite often with headphones, and my experience is, not every recording really needs crossfeed (but still, most do). Being "conditioned" early for headphone listening, I regard the sound not changing with head position as a feature, allowing me to listen in any position. It makes more obvious that i'm "listening to a machine", but I know it anyway, so what?

 

[myHRTF]

not every recording really needs crossfeed

Exactly, the contribution of crossfeed to successful speaker virtualization through headphones is highly overrated.

I regard the sound not changing with head position as a feature

True, in some applications like mixing and mastering it's an advantage rather than a disadvantage, provided that the user belongs to the group of people who don't need head-tracking in the first place as described above.

 

[MaxwellsEq]

As I sit here listening to stereo through headphones, and enjoying myself... I was wondering... with all the DSP horsepower we have waiting for something to do these days...

Why can't my DSP-enabled headphone amplifiers simply convert stereo into binaural (or several simultaneous options thereof) on the fly? Why are binaural recordings singular things?

As others have said, it's possible to use DSP to improve the way stereo is presented when listening on headphones. However it won't be possible to turn a commercial two channel recording into a dummy-head binaural output. The information required to do this is not captured in the two channels.