What makes speakers "disappear " and can it be measured?

[tecnogadget]

you can also increase the side-wall reflections by e.g. heavy toe-in to increase the enveloping sound.

A few folks even try “Toe-Out” for this purpose. It would be nice more definitive testing, research on Toe-In-Out, straight, etc.

 

[Thomas_A]

Yes indeed, toe-in will undoubtedly have an effect on phantom image position (as explained on the pdf I uploaded). The way I see it, or at least I’ve empirically perceived, about the ratio of reflected vs direct sound, is that it affects the soundstage, how wide, how much depth or how much “3D like” the presentation feels.

My subjective appreciation after testing:
Direct sound: more like “in studio sound”, very focused or pinpoint, and not so big sound stage. Reflected: wider, like a “wall of sound”, bigger sound, livelier.

The thing about phantom image shifting if you are close enough to one of the sepakers is that since you are closer...sound time arrival is earlier than the other speaker, and our brain cares about that, so there isn’t any work around that, it’s the physics of listening to Stereo.

Well the work-around to some extent is that toe-in gives less power of the closest speaker and more of the distant one which is on-axis. It is no absolute fix but it helps. In my system the mono dialogue is still fixed to the center even if I am placed in front of one speaker. It works fine at listening position in the sofa, but moving closer to the speaker - then it tilts. So I guess there is also a contribution of reflected sound that helps in this matter.

The interesting question (for me at least) is if there is any way to measure phantom centre position?

 

[Thomas_A]

Yes indeed, toe-in will undoubtedly have an effect on phantom image position (as explained on the pdf I uploaded). The way I see it, or at least I’ve empirically perceived, about the ratio of reflected vs direct sound, is that it affects the soundstage, how wide, how much depth or how much “3D like” the presentation feels.

My subjective appreciation after testing:
Direct sound: more like “in studio sound”, very focused or pinpoint, and not so big sound stage. Reflected: wider, like a “wall of sound”, bigger sound, livelier.

The thing about phantom image shifting if you are close enough to one of the sepakers is that since you are closer...sound time arrival is earlier than the other speaker, and our brain cares about that, so there isn’t any work around that, it’s the physics of listening to Stereo.

One issue with having more direct sound is that you will encounter larger errors due comb filtering, creating an uneven frequency response curve. So I wonder if this reveals the loudspeakers? If resonances or variations in the frequency response is an hall-mark of "identifying the speakers", you need to partly compensate for these variations in the frequency response curve and use reflections to fill in the dips.

 

[kukocz]

What do you think about super-toe-in setup? I mean the focal point is created before the listener.
Ken Ishiwata used to use it for Marantz presentation on audio shows...

 

[Tangband]

All of the above are fine by me. However, delays of 20 ms, or 6.8 meters, and avoiding shorter ones to fill up the gaps and peaks will not be possible in most rooms.

I didnt say that one should avoid reflections between 5-20 ms, wich as you say is impossible.
But a bigger room has the advantage of later reflections, at 20-30 ms its gonna be beneficial.

 

[Thomas_A]

What do you think about super-toe-in setup? I mean the focal point is created before the listener.
Ken Ishiwata used to use it for Marantz presentation on audio shows...

View attachment 144445

I have always used it.

 

[tuga]

What do you think about super-toe-in setup? I mean the focal point is created before the listener.
Ken Ishiwata used to use it for Marantz presentation on audio shows...

View attachment 144445

As far as I know such a setup is used to widen the "sweet spot" (when you have more than one listener).

 

[tuga]

I noticed that some speakers just disappear and it feels like the sound appears out of thin air, and others don't feel like that at all, and the sound is just coming out of 2 speakers.
What makes speakers "disappear" and can it be measured?

Absence of transduction distortion and cabinet-generated issues, omnidirectional directivity, untreated boundaries (walls, floor, ceiling).

I can be difficult to pinpoint the location of an instrument or voice in a stone wall church unless you're sitting very close to the sound source.

 

[Thomas_A]

I didnt say that one should avoid reflections between 5-20 ms, wich as you say is impossible.
But a bigger room has the advantage of later reflections, at 20-30 ms its gonna be beneficial.

Yes, but again. Very difficult in most instances. One have to also consider the level of the first reflections. If it is low enough compared tot he direct sound it will not disturb.

 

[Thomas_A]

As far as I know such a setup is used to widen the "sweet spot" (when you have more than one listener).

It will also change the ratio between the power of reflections regarding speaker <-> nearest wall vs. opposite wall, given that the speaker has a dispersion pattern with lower energy with higher angles.

 

[Jim Matthews]

How to make speakers disappear:
1. Burglary
2. Vanishing cream.
3. Black magic.
4. Blindfold.

5. Divorce settlement
6. Grandkids off-leash

 

[Jim Matthews]

I have always used it.

View attachment 144446

Very cool.

Wideband driver "up top"?

Nicely done.

 

[Thomas_A]

Very cool.

Wideband driver "up top"?

Nicely done.

Thanks, the monitor plays from 80 Hz and up. Sub stands reach down to about 21 Hz. Toe-in is such that each speaker is on-axis for the left and right listening positions in the sofa. The dispersion characteristics is "quite wide" up to 7-8 kHz.

 

[pozz]

The interesting question (for me at least) is if there is any way to measure phantom centre position?

You have to use a dummy head and torso to make binaural measurements, and plot interaural time and level differences and crosscorrelation (ITD, ILD and IACC) per frequency with a variety of signals (sweeps, noise, clicks). That allows inferences about apparent source width (image size) and position. Virtual auditory display is a another term used in the lit to refer to the problem.

I'm not in the field so I can't answer precisely, but a lot of work is being done on virtualization/binauralization based on acoustic or anechoic data. For example deriving spatial impressions from ambisonic recordings. Headphone based head tracking systems, like by ones by Audeze and Smyth Research, also seem to be pretty successful. Hopefully the techniques aren't fully locked into esoteric commercial or academic settings. It would be excellent if there was a readymade way to input Klippel data and plot approximate spatial results for stereo/multichannel speaker setups.

 

[Jim Matthews]

Thanks, the monitor plays from 80 Hz and up. Sub stands reach down to about 21 Hz. Toe-in is such that each speaker is on-axis for the left and right listening positions in the sofa. The dispersion characteristics is "quite wide" up to 7-8 kHz.

Your design?
The fit and finish are sharp.

 

[Thomas_A]

Your design?
The fit and finish are sharp.

Thanks! However they are nothing special in terms of finish. I did what I could with the skills and tools I had.
The shape of the speakers, no they are based on a Mirsch speaker system around 1990 but with other drivers and filter. I posted some other pictures in the DIY forum quite some time ago.

 

[Thomas_A]

You have to use a dummy head and torso to make binaural measurements, and plot interaural time and level differences and crosscorrelation (ITD, ILD and IACC) per frequency with a variety of signals (sweeps, noise, clicks). That allows inferences about apparent source width (image size) and position. Virtual auditory display is a another term used in the lit to refer to the problem.

I'm not in the field so I can't answer precisely, but a lot of work is being done on virtualization/binauralization based on acoustic or anechoic data. For example deriving spatial impressions from ambisonic recordings. Headphone based head tracking systems, like by ones by Audeze and Smyth Research, also seem to be pretty successful. Hopefully the techniques aren't fully locked into esoteric commercial or academic settings. It would be excellent if there was a readymade way to input Klippel data and plot approximate spatial results for stereo/multichannel speaker setups.

Thanks, these are the keywords I need. I suspected that dummy head recordings are needed. I'll see if I have some time to study "Virtual auditory display."

 

[Tangband]

Thanks, these are the keywords I need. I suspected that dummy head recordings are needed. I'll see if I have some time to study "Virtual auditory display."

One has to remember that optimal recordings for playback on loudspeakers are done in an entirely different way than recordings that are optimal for headphone listening.
Its a big difference how you place your microphones, and the distance between them. If we are talking two-channel recordings in real stereo.

 

[Thomas_A]

One has to remember that optimal recordings for playback on loudspeakers are done in an entirely different way than recordings that are optimal for headphone listening.
Its a big difference how you place your microphones, and the distance between them. If we are talking two-channel recordings in real stereo.

Yes, perhaps. But this question relates to find position and size of virtual phantom sources in stereo, where recording of the test signals needs to be done with dummy heads.

 

[Duke]

I have always used it [super toe-in setup].

View attachment 144446

VERY NICE! I REALLY like the way you incorporate that bevel into BOTH the mains and the subs!!

Here is my (commercial) variation on the theme, photo by Eric Franklin Shook of Part-Time Audiophile. The stands include a subwoofer and an adjustable up-and-back firing coaxial driver, whose output contributes to the reverberant field (yeah I know that term in imprecise when applied to the size rooms we use, but imo it conveys the idea well):



You can see that from where the photo was taken, well off to the side of the central sweet spot, you're on-axis of the far speaker but well off-axis of the near speaker, so you still get good soundstaging due to time-intensity trading. The waveguide in the main speakers is a 90-degree constant-directivity device, crossed over to the midwoofer where the latter's pattern has narrowed to 90 degrees.

Imo the additional reverberant energy from the rear-firing coaxial (the lower cone in the photo below) contributes to the speakers "disappearing" as the apparent sound source. When the coaxial is turned off, to my ears it seems like I'm hearing "more of the playback room" and "less of the recording venue":