sweet spot size aka phantom center

[pozz]

@dasdoing Here's the paper I was originally thinking of: https://www.sfxmachine.com/docs/FixingThePhantomCenter.pdf

There's also this one that could potentially be implemented through DSP:

• http://elias.altervista.org/html/3_speaker_matrix.html
• http://elias.altervista.org/html/2_vs_3_stereo_high_freq.html

 

[dasdoing]

about over-toed in speakers: tried it out and it sound very strange. center is all over the place, thats why you get a "better" center of axis. but it sounds crap for me, no real stereo image anywhere. though it probably is more dramatic in my treated room, since I am used to a very defined stereo image

@dasdoing Here's the paper I was originally thinking of: https://www.sfxmachine.com/docs/FixingThePhantomCenter.pdf

There's also this one that could potentially be implemented through DSP:

• http://elias.altervista.org/html/3_speaker_matrix.html
• http://elias.altervista.org/html/2_vs_3_stereo_high_freq.html

thanks. the first link doesn't work for me atm.
about the other links: it just mixes full L+R by a coeficient to generate center? how is stereo image preserved?

 

[ernestcarl]

about over-toed in speakers: tried it out and it sound very strange. center is all over the place, thats why you get a "better" center of axis. but it sounds crap for me, no real stereo image anywhere

To be fair, only a handful of speakers would work really well with this technique, according to Matthew Poes. As I have tested this myself with my S8 monitors (including the KH120 in the nearfield). It does not sound like crap to me, and there really is a stereo image and soundstage that is maintained. It sounds just more distant and diffuse. Again, from my own experience, left and right from center seats sound more balanced to me -- but at the expense of the center seat listening position. I can still see situations where I may want to do this, but not very often indeed. And if you are the only one there in the room and sitting in the far left or right seat frequently, you could apply some DSP presets to produce MUCH better results by balancing the interaural level and time difference i.e. delays and EQ to get an essentially flat response.

 

[dasdoing]

To be fair, only a handful of speakers would work really well with this technique, according to Matthew Poes. As I have tested this myself with my S8 monitors (including the KH120 in the nearfield). It does not sound like crap to me, and there really is a stereo image and soundstage that is maintained. It sounds just more distant and diffuse. Again, from my own experience, left and right from center seats sound more balanced to me -- but at the expense of the center seat listening position. I can still see situations where I may want to do this, but not very often indeed. And if you are the only one there in the room and sitting in the far left or right seat frequently, you could apply some DSP presets to produce MUCH better results by balancing the interaural level and time difference i.e. delays and EQ to get an essentially flat response.

I angled my speakers like this: left seat of my 3 person couch I angled right tweeter straight into my right ear, vice versa for the left speaker. I din't get a clear center at sides, though it realy seams to come more form the center, it's just that center was almost as big as speaker width. but ok, it was better then normal on the sides (perhaps, didn't test long term). my biggest problem with it, and what caused my to discard the idea imidiatly was the center seat. it sounded cavelike, and center was all over the place, too. I think the problem is that in the center seat I had the right speaker blowing direct sound into my left ear (where as right ear would recieve of axis sound), and vice versa. this is where the theory fails imo. it only works if you ignore the cross talk (what left ear hears form right speaker and vice versa). If there was no crosstalk it would probably work

 

[ernestcarl]

I angled my speakers like this: left seat of my 3 person couch I angled right tweeter straight into my right ear, vice versa for the left speaker. I din't get a clear center at sides, though it realy seams to come more form the center, it's just that center was almost as big as speaker width. but ok, it was better then normal on the sides (perhaps, didn't test long term). my biggest problem with it, and what caused my to discard the idea imidiatly was the center seat. it sounded cavelike, and center was all over the place, too. I think the problem is that in the center seat I had the right speaker blowing direct sound into my left ear (where as right ear would recieve of axis sound), and vice versa. this is where the theory fails imo. it only works if you ignore the cross talk (what left ear hears form right speaker and vice versa). If there was no crosstalk it would probably work

You'd need a very, very good narrow controlled directivity waveguide.

The example Matthew gave (if I remember correctly) was the JTR Noesis 212RT (click for full review)

Actually, the ff. image from AVNirvana looks pretty amazing...

 

[dasdoing]

You'd need a very, very good narrow controlled directivity waveguide.

The example Matthew gave (if I remember correctly) was the JTR Noesis 212RT (click for full review)

Actually, the ff. image from AVNirvana looks pretty amazing...

View attachment 69036

in the middle you still get more direct sound in the oposit ear.

 

[ernestcarl]

in the middle you still get more direct sound in the oposit ear.

Again, it shouldn't matter at all... and if I remember correctly, someone else did mention maintaining room reflections (avoid absorbers) and maybe diffusors could be beneficial. This isn't gonna work optimally just with any speaker and room arrangment.

 

[ernestcarl]

Compare the above graph to the RM28ac (uses very similar horn shape as the S8 xo 2.4kHz up) which costs ~$4,500:

Very, very far from ideal to use for the extreme toe-in experement that I did... LOL. I wonder what the off-axis response is for your speakers, but I suspect far from ideal as well.

 

[pozz]

about the other links: it just mixes full L+R by a coeficient to generate center? how is stereo image preserved?

There will be certain colourations and so on but this is how they work. The time/amplitude variations are preserved. You'll see MS processing is one of the methods.

Other more complicated methods include delays and frequency-specific filterbanks. The first link I gave was to a pdf of a paper that discusses these as well as some of the combfiltering that's introduced by regular stereo playback.

 

[dasdoing]

Compare the above graph to the RM28ac (uses very similar horn shape as the S8 xo 2.4kHz up) which costs ~$4,500:

View attachment 69038

Very, very far from ideal to use for the extreme toe-in experement that I did... LOL. I wonder what the off-axis response is for your speakers, but I suspect far from ideal as well.

I have no clue, actualy. but since they are studio monitors probably on the narrow side

 

[dasdoing]

I was thinking about a solution while reading about mid-side processing (which works diferently anyways).
centre speaker shouldn't only play what comes dead center. what is panned 50% to the left would be played equaly loud by left and center speaker; the center speaker does the part the right speaker would do in normal stereo.

CS=(L+R)-root of (L-R)*(L-R)
LS=L-R
RS=R-L

if we have a signal panned to center L=50; R=50
CS= 100
LS= 0
RS=0

if we have a signal panned L=75; R=25
CS=50
LS=50
RS=0 (negative)

if we have a signal panned L=100; R=0
CS=0
LS=100
RS=0 (negative)

 

[dasdoing]

@markus gave me a link to this intresting document in another forum: http://decoy.iki.fi/dsound/ambisonic/motherlode/source/7036.pdf
have not read it yet

 

[dasdoing]

https://www.meridian-audio.info/public/trifield[2563].pdf

 

[dasdoing]

You'd need a very, very good narrow controlled directivity waveguide.

The example Matthew gave (if I remember correctly) was the JTR Noesis 212RT (click for full review)

Actually, the ff. image from AVNirvana looks pretty amazing...

View attachment 69036

since I am flirting with a line source PA speaker (with controlled directivity) I am playing around with the PA software and remembered this.
this is more or less my room. mic 2 is center, 4 is 60cm to right, 5 is 60cm to left. aimed at center:

so the ideal would be to bring 4 and 5 as close together as possible. since the room is symetric it would mean that at position 4 both speakers would have same volume. here is the kind of surprising best angle I found

over 1Khz they are not too far off.
now the software doesn't calculate any reflections though. but I might try this out.

 

[ernestcarl]

since I am flirting with a line source PA speaker (with controlled directivity) I am playing around with the PA software and remembered this.
this is more or less my room. mic 2 is center, 4 is 60cm to right, 5 is 60cm to left. aimed at center:

View attachment 169684

so the ideal would be to bring 4 and 5 as close together as possible. since the room is symetric it would mean that at position 4 both speakers would have same volume. here is the kind of surprising best angle I found

View attachment 169686

over 1Khz they are not too far off.
now the software doesn't calculate any reflections though. but I might try this out.

That is basically what I do -- I'm using some cheap plastic rotating platforms under the monitors and labeling the angles with tape as I like being able to switch angles quickly whenever I feel like it). There is one issue in that the sound at the center listening position becomes a little less "focused". Other monitors like the large JTRs may perform much better here.