How do you hear headphone 'soundstage'

[wiggum]

Funny thing though that notching the 10kHz with a filter does nothing for the sound stage.
Also many headphones actually have peaks there.
The 10kHz notch is caused by the 'ear' and in measurements by the fake ear (pinna + earcanal) and depends on the the fixture + pinna used.

This all makes the evidence of the 10kHz notch (Rtings) extremely 'thin'. At Rtings they thought it (along with the difference between a pinna + pinnaless (but with ear canal) was something they could 'rate' easily which was the goal. It was kind of determined by looking at a few headphone that were said to have good spatial effects.
Those 2 or 3 headphones all had similar 'effects' caused by the pinna and that became the target. The 10kHz dip is part of that pinna.
When you look at the (older) plots you will find the 'target' to be very close to the HD800 response.

I mean ... wouldn't it be great if all headphones with poor sound stage suddenly would get a good sound stage if we would just notch 10kHz.
Sadly that doesn't work.

Did you ABX with a 10 kHz notch? Easy to create 2 samples files with ffmpeg?

 

[solderdude]

Extensively... experimented with different frequencies and even Q factors.

That said... even binaural doesn't do it for me. This is not strange either given the poll.
It is a perception thing + pinna effects.
What 'works' for you (notching 10kHz) may well fool your brain into creating depth but it won't for others so is no universal indicator for sound stage.

I can hear the 10kHz getting 'softer' but does not change the sound stage for me.

 

[wiggum]

Extensively... experimented with different frequencies and even Q factors.

That said... even binaural doesn't do it for me. This is not strange either given the poll.
It is a perception thing + pinna effects.
What 'works' for you (notching 10kHz) may well fool your brain into creating depth but it won't for others so is no universal indicator for sound stage.

I can hear the 10kHz getting 'softer' but does not change the sound stage for me.

Try these samples. I am also not convinced that much as I find using the bs2b or sofa filter has more impact. I think ITD(interaural time difference) is more significant than this notch.

 

[jacobacci]

The Smyth Realizer A16 sheds some light on how 3D spatialization is created.
We have two eardrums. That is essentially a two channel detector. The brain only has two channels to work with. The individual HTRF (shape of the head, shape of the earlobes) diffracts the sound coming from different directions in a personally individual way. The brain has adapted to extract a 3D spatial sound "image" from these two channels, so can detect the sabretooth tiger behind you.
The Smyth Realiser maps the 3D sound (as captured in a Dolby atmos stream) through two transformation functions:

• The HTRF which represents the way 3D sound is diffracted by the face and earlobes. Usually in recording sound signatures of speakers in a listening room the HRTF of a particular person is used (mics placed in the ears of the person). This is close enough to work for the majority of listeners.
• The personal HTRF, which represents how sound produced by a headphone is diffracted by the earlobes and the ear canal. This personal HTRF is derived by calibration with in ear microphones.

These two transfer functions are used to convert a Dolby atmos (or Auro 3D or DST:X) stream to a binaural (headphone) signal that at the eardrums is as close as possible to what the individual listener would hear in the real room.
Hence a plain stereo signal reproduced over headphones is always a compromise that may work for some people (where the HTRF underlying the stereo recording just happens to be close) and not work for others. Definitely a 10kHz notch will not improve things.
Something that does work: Apple Music and Tidal actually output a binaural signal from an iPad / iPhone. Of course they use generic HTRF to create this binaural signal, but it actually works quite well with a reasonably good headphone. Simply connect a DAC to the iPhone via USB-C and you're off to the races. Enjoy.

 

[threni]

I thought this was interesting:

Explains the futility of recommending headphones with a good soundstage. Is there a pattern behind headphones which get described as such? Perhaps they have to envelop the whole ear, for example.

I wonder if it is/will be possible to create an app which takes photos of your outer ears (ear? are they symmetrical?) in 3d (via video I guess) and compute the processing required to produce the transformation function needed to create the effect of a wider soundstage. Are people's ears like fingerprints so everyone "needs" their own function? Or are there a limited number of functions such that you could just try a few popular ear shapes? Perhaps everyone who hears good soundstage in headphones share some physical characteristic of their outer ears?

 

[fcserei]

Apple can create personalized HRTF with scanning your head with an iPhone. It works quite well, in my experience results are comparable to my personal HRTF I've created with mesh2hrtf.
The problem with Apple Spatial on IDevices is that they are not customizable, it's the usual Apple one solution fits all approach. You can not EQ or select BRIR for the spatialization.
But if you use Logic Pro, there are much more possibilities. You can get your BRIRs with Space Designer. It has a built in "Spatial Audio Monitoring" Plugin and also a Dolby renderer for headphones. On a Mac in Logic Pro you can use your personal HRTF with Apple headphones with head tracking or any headphone without head tracking to monitor any source up to 7.1.2 with Apple Spatial (selectable movie and music mode with) or with Dolby rendering.
Also there is there is a speaker option, where you can use 2 close spaced speakers, and the plugin with crosstalk cancellation virtualize a complete surround system, like the BACCH Stratos.

 

[wiggum]

I thought this was interesting:

Explains the futility of recommending headphones with a good soundstage. Is there a pattern behind headphones which get described as such? Perhaps they have to envelop the whole ear, for example.

I wonder if it is/will be possible to create an app which takes photos of your outer ears (ear? are they symmetrical?) in 3d (via video I guess) and compute the processing required to produce the transformation function needed to create the effect of a wider soundstage. Are people's ears like fingerprints so everyone "needs" their own function? Or are there a limited number of functions such that you could just try a few popular ear shapes? Perhaps everyone who hears good soundstage in headphones share some physical characteristic of their outer ears?

Probably not though the sound stage is perceived better if the sound interacts with the pinna, so a over-ear headphone would sound better than an IEM.

I have attached 2 samples. One is plain while the other one is processed through a BRIR. Listen to both using a IEM and over-ear headphone.

 

[threni]

Probably not though the sound stage is perceived better if the sound interacts with the pinna, so a over-ear headphone would sound better than an IEM.

I have attached 2 samples. One is plain while the other one is processed through a BRIR. Listen to both using a IEM and over-ear headphone.

I listened on my HD660S. It sounds a bit reverby, making the main instrument a little less clear. So the music sounds more like it's in a (bigger) room, but that room is still inside my head.

 

[wiggum]

I listened on my HD660S. It sounds a bit reverby, making the main instrument a little less clear. So the music sounds more like it's in a (bigger) room, but that room is still inside my head.

Looks like there is still some other X factor at play. For me, with over-ear headphones, it feels out of my head, but with IEMs the effect is not so great.