Loudspeakers - looking for a correlation between measurements and listening impressions

You can't get sounds coming from the front with headphones
The front is always silent
Like a black hole
....
The sound never comes from the front Never
I have listened to hundreds of them
Video gamers are pushing for that
About 10% of the population have that problem so you are not alone. The solution (used by some video games) is head tracking. We sense direction using 2 sets of auditory cues.

One set is "Fixed Head". Simple binaural / headphone listening uses this. But we need precise modelling of your individual Head Related Transfer Function HRTF for it to work well ... and for some people, it is never accurate enough. Most of these people hear stuff above or in their head instead of 'straight ahead'.

The other set of cues are "Moving Head". ie the sound in each ear changes as you move your head. There are theories for small head movements and big head movements but Michael Gerzon showed that the his equations describe both. Your brain learns and adapts to the "Moving Head' cues instantly.

These theories are incorporated in a 'surround' sound system called Ambisonics which also includes height. The games use head tracking to move you around the virtual Ambisonic Sound Field
 
Sorry but you can't get sounds from the front with headphones
That's not what I said. I said that there is a soundstage with my Stax headphones (and other headphones as well).

Not sure if it has been discussed in this thread, but you might be interested in the work by Earl Geddes. He looked quite a bit in the correlation between non-linear distortion and perceived sound quality.
 
  • Thank You
Reactions: ginetto61
I think the problem is that directional hearing is highly dependent on the physical form of your ears/head, so it's different from person to person. The difficulty is not dsp calculation but every person would need their own algorythm (which makes it expensive).
But I might be wrong here, happy to be corrected!
on youtube there are binaural recordings
there is the famous one of the barber cutting the hair while turning around the recording head
you can hear it clearly on the left but when it moves to the center in front it becomes impossible to locate
at most it seems to have entered inside the head
an unnatural sensation and as such unsatisfactory
 
About 10% of the population have that problem so you are not alone. The solution (used by some video games) is head tracking. We sense direction using 2 sets of auditory cues.
Ok i am more calm now Sometimes i feel alone in my doubts
the day someone will overcome this problem headphones use will explode
Actually young people do not seem to matter Because they like sounds inside their heads 😵
One set is "Fixed Head". Simple binaural / headphone listening uses this. But we need precise modelling of your individual Head Related Transfer Function HRTF for it to work well ... and for some people, it is never accurate enough. Most of these people hear stuff above or in their head instead of 'straight ahead'.
The other set of cues are "Moving Head". ie the sound in each ear changes as you move your head. There are theories for small head movements and big head movements but Michael Gerzon showed that the his equations describe both. Your brain learns and adapts to the "Moving Head' cues instantly.
These theories are incorporated in a 'surround' sound system called Ambisonics which also includes height. The games use head tracking to move you around the virtual Ambisonic Sound Field
thank you very much for your valuable advice
with headphones the L signal never reaches the R ear unless some sort of crossfeeding is applied
i think it should be mostly a matter of delay and level
I am sure they will find an acceptable solution soon There is huge demand for these systems
Most of the humans live in flats with touchy neighbours or even worse screaming wives
There is a big need for closed HPs with great crossfeed
 
I just performed an eye opening test. With my Android phone (I don't think it works with iPhone) when I hold my phone at my left ear I can hear it at my left ear. When holding it to my right ear, I can hear the sound in my right ear. When I hold it in front of me both ears hear it and I could imagine it is in front of me. When holding it at the back of my head it is much softer but I could unmistakably recognize it coming from the back, not the sides or front. Now I wonder if there are any headphone or speaker that can master this feat. This was all done being double blind by putting my wife's black panty over my head (she was not pleased and thought I have become perverted). Goodness, why do we just not listen to the music and enjoy what we hear regardless of from where it is coming.
 
You had both ears taking part localizing the sound source, including your whole torso so full HRTF at play. Also other information like arm position and eyesight and the panty magnifying the effects were happening so your brain had quite accurate information on where the mobile was located, even if it was muted! If you had mics inside both of your ear canals recording the event and headphones calibrated to your hearing you could enjoy the same event later eyes closed, with or without the panty.

Panty aside, it's fun to test stuff like that with mobile if one wants to experiment with how audio is perceived, it can teach a thing or two although others might consider it a nut job. Did you hear the noise frequency response change as you moved it around? Thats the HRTF, at least partly, your arm also affects the sound and which way the phone is rotated as it's not ideal point source. If you put the noise playing mobile on a desk and reach it with your hand the sound already changes, as it reflects from your arm.
 
Last edited:
I just performed an eye opening test. With my Android phone (I don't think it works with iPhone) when I hold my phone at my left ear I can hear it at my left ear. When holding it to my right ear, I can hear the sound in my right ear.
yes that is the problem In an headphone the sound of the left transducer never reaches the right ear and viceversa
with stereo speakers both ears hear both channels in a different way
There has been an attempt to simulate a speaker situation with an Headphone

1729605185045.png


still the position of the transducers is not optimal I still bet on dsp It is the only hope for humanity
When I hold it in front of me both ears hear it and I could imagine it is in front of me. When holding it at the back of my head it is much softer but I could unmistakably recognize it coming from the back, not the sides or front. Now I wonder if there are any headphone or speaker that can master this feat. This was all done being double blind by putting my wife's black panty over my head (she was not pleased and thought I have become perverted). Goodness, why do we just not listen to the music and enjoy what we hear regardless of from where it is coming.
if needed, I think we could testify that what you are do8ng is for purely scientific and worthy reasons.

Regarding the mics placement i would try two omni mics very close one to the other at a distance similar to the head width
Usually they are distant one to the other
Inner ear mics will not work for soundstage imho
 
Last edited:
Very steep xovers don't sound good .. especially EVIL digital FIRs (the most common type). You want a smooth transition so there are no kinks in the directivity.

I won't go into why FIRs are truly EVIL cos there are many here who believe they are the bees knees 😊

Er. MiniDSP and others offer IIRs for filter traditionalists.

More to the point, FIRs can be designed to have almost any slope depending on the filter length and other parameters.
Here's a shallow slope lowpass FIR, -3 dB at 1000 Hz, -12 dB at 2000 Hz, and about -50 dB at 4000 Hz, followed by a medium slope highpass FIR with -6 dB at 2000 Hz, -26 dB at 1000 Hz, -50 dB at 500 Hz, and about -80 dB at 250 Hz. The last is a classical example of a steep lowpass FIR which falls off the edge of the world after the cutoff frequency.

The first is not far from a 2nd order analog LPF for the first octave and chance, the second is close to a 4th order Linkwitz-Riley HPF, and the last is what most people consider a typical FIR.

So why use an FIR for a similar response to a standard filter? Two reasons:

First off, FIRs are linear phase so the group delay is a constant, which to my ears matters especially for percussion instruments. Flat group delay in the low mids around 250 Hz matters when reproducing sharp transients such as snare drums. Snares are supposed to be rude, making you blink when a drummer smacks them hard. FIR crossovers preserve that quality, whereas analog and IIR crossovers do not, making the drum a bit too polite for belief. Moreover I hear linear phase around 100 Hz as helping bass guitar and kick drums from running into each other. I haven't done a careful listen for other frequencies, but seeing as how the ear uses interaural time difference (ITD) below about 1350 to 1500 Hz for localisation [1] I have little ideological problem with using FIRs at 2000 Hz since the overlap could extend down to the ITD sensitive range.

Second, FIRs can be tuned for useful attenuation outside the passband as shown by the steep HPF example, again -6dB at 1000 Hz, then -19 dB at 700 Hz, -36 dB at 500 Hz, followed by -90 dB at 250 Hz. The rolloff is not far from a Linkwitz-Riley 4th order for about the first half octave (implying almost a full octave of overlap when used with the complementary LPF for the other driver) but it increases dramatically beyond that point. This means the tweeter doesn't have to deal with much signal outside its passband as neither does the midrange, which could help tame drivers with resonances beyond the crossover point.


Conclusion: like any other tool FIRs can be useful or harmful depending on how they're used. As an aside, phase response is worth considering but only until 2000 Hz or so.


[1] https://bionumbers.hms.harvard.edu/bionumber.aspx?s=n&v=4&id=112329
 

Attachments

  • ShallowFIR.png
    ShallowFIR.png
    510.8 KB · Views: 26
  • MediumSlopeFIR.png
    MediumSlopeFIR.png
    497.4 KB · Views: 30
  • SteepFIR.png
    SteepFIR.png
    496.8 KB · Views: 24

Loudspeakers - looking for a correlation between measurements and listening impressions​

In my Book "High Performance Loudspeakers - 4th Edition 1991" by Martin Colloms (Monitor Audio), there are tables and descriptions matching measurement and tonal observations vs. audiophile descriptions...

Let me take some pictures of those pages and be back ! 😉

T

@ginetto61 : Here are the pages in question, concerning loudspeaker assessment :

1729618150935.png


1729618199790.png


1729618288774.png


T
 
of course not
As I see it, it's possible to have a driver working with no audible distortions, including a sound of the cone material.

i guess paper cone
Interesting because paper damps the cone resonances to some degree. It's a reasonable choice when you have to cross high, near breakup. However if you do it properly it shouldn't matter so much.
 
  • Like
Reactions: ginetto61
Hi I have seen a video worth some discussion
Imho
There's a cone driver with + and - terminals connected to a scope
When the cone is hit an impulse appears on the scope
Like an electrical response to a mechanical action
Often I see someone hitting the cone and listen for the sound
I like this test with the scope much better
It's possible to see also the damping
Now I want a scope badly
 
@ginetto61 : Here are the pages in question, concerning loudspeaker assessment :

View attachment 1370999

View attachment 1371000

View attachment 1371001

T
Hi sorry for the belated reply
Actually I crave for something quite less subjective like lab measurements
Instruments can hear signals that I have no hope to hear
I like so much the John Dunlavy approach
When I see labs I see a lot of money spent on instruments
Are there for fun?
It's better to rely on the hearing system of a mature man maybe with dirty ears?
 
  • Like
Reactions: tubelectron