• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

3.0 for stereo listening ?

bud947

Member
Joined
May 18, 2021
Messages
77
Likes
26
This is a general question about 3.0/3.1 setup for stereo listening as well as a specific use case with my setup.

I have a pair of polk R700 currently in a stereo setup, but because of the narrow directivity above 7/8kHz It needs significant toe-in to bring some sparkle while wife want a 30° toe out (perpendicular to the wall). I love the speakers, and honestly (except for top end) there is no better option at this price point when accounting for output, front/bottom port, distortion, wide directivity, linearity and I'll first try to "upgrade" these rather than going straight upmarket at x2 price (Mofi...)

Initially I thought about upgrading the tweeter section of the r700 with SB acoustics SB26ADC with a wideguide but this would need a significant DYI effort (redesigning the midrange/tweeter crossover, enclosure reworks...) and money.

Then looking at the measurement of the R700, the beaming is starting at 7kHz and it falls pretty quickly at 30° and higher. As my speakers are at 9 ft of each other, I thought I can put a very good 3 way-center channel (polk L400 or LSiM 706c) with a L+R mix and a -6/-9 db shelf PEQ below 7khz.

My main concern is the fact that I will now have 3 tweeters for a stereo listening. Would it work ?

Then generally speaking, is there a good way to implement listening of stereo music in a 3.0/3.1 setup ?

PS: I have a LX-805 and a minidsp flex as DSP & multichannel processor.
 
You may want to read this thread. The upshot is that with only two sets of data (left data and right data) in stereo that describe levels and phase, there is no simple way to determine what the middle speaker should deliver (without messing up relative levels and soundstage etc.)

 
Thank you for pointing me toward that very informative thread. I'll try implementing the Linear Stereo Matrices detailed and see how it works.
The idea of time/intensity trading is also interesting but needs speakers with constant but narrow directivity (which the polk are only above 7kHz).
It is also very interesting that In the end OP agrees that he needs a center channel.
 
This is a general question about 3.0/3.1 setup for stereo listening as well as a specific use case with my setup.

I have a pair of polk R700 currently in a stereo setup, but because of the narrow directivity above 7/8kHz It needs significant toe-in to bring some sparkle while wife want a 30° toe out (perpendicular to the wall). I love the speakers, and honestly (except for top end) there is no better option at this price point when accounting for output, front/bottom port, distortion, wide directivity, linearity and I'll first try to "upgrade" these rather than going straight upmarket at x2 price (Mofi...)

Initially I thought about upgrading the tweeter section of the r700 with SB acoustics SB26ADC with a wideguide but this would need a significant DYI effort (redesigning the midrange/tweeter crossover, enclosure reworks...) and money.

Then looking at the measurement of the R700, the beaming is starting at 7kHz and it falls pretty quickly at 30° and higher. As my speakers are at 9 ft of each other, I thought I can put a very good 3 way-center channel (polk L400 or LSiM 706c) with a L+R mix and a -6/-9 db shelf PEQ below 7khz.

My main concern is the fact that I will now have 3 tweeters for a stereo listening. Would it work ?

Then generally speaking, is there a good way to implement listening of stereo music in a 3.0/3.1 setup ?

PS: I have a LX-805 and a minidsp flex as DSP & multichannel processor.

I have the Polk R700 with some hypex amp/dirac too and I'm super interested about the Toe In

It's possible to share some pictures so I can compare with my actual set up ?

How much ° is your toe in ?
 
Interesting. Your wife and you obviously have different ears - unless I misunderstood and she wants a different angle for purely aesthetic reasons.

These R700 are somewhat pricey, although not by "high-end" standards. You say you love them in every regard except directivity? Hmm, difficult. It would be all to easy to suggest "try different speakers then" lol. But here's some thoughts:

The roughly comparable R600 (just a smaller version) are what I very recently read a comment about: commenter had them and clearly preferred the recently acquired Heco Aurora 700 over them. The latter have a nicely wide dispersion in the higher treble.

Now what if, and I'm just throwing ideas around, you tried listening to the Aurora 1000 model? Just like R600 vs. R700, it's the bigger model, with 2x8" woofers instead of 2x6.5", using the same tweeter. These Heco tweeters have a little waveguide (fancily called "Fluktus" in marketing speak lol) that seems to help with width. I have the 700 models, and indeed the sweet spot seems unusually wide.

Here's some (limited) measurements:

As said, the bigger 1000 model uses the same tweeter, and apart from that it's a very similar 3-way 2x8" design as the R700.

I know you said you like your R700 very much except for this one thing. It's sometimes hard to find a compromise between two sets of ears and auditory brains. At least I can confirm these Heco models are excellent all around, with imaging and width being a particular strength (and bass too), so maybe it's an idea worth thinking about. Both 700 and 1000 are also about half the price of the R600 and R700, respectively. Certainly way more budget oriented, so you won't get nice real wood veneer and such. But great sound for the money, both objectively and subjectively.

TL;DR: Maybe you don't need an upgrade, but a sidegrade instead. Looking at other speakers may be worthwhile for solving your problem.
 
Last edited:
It is also very interesting that In the end OP agrees that he needs a center channel
The OP started with the belief that he needed a centre channel because, for whatever reason, he felt the phantom centre doesn't work for him.

For historical reasons, we ended up with two channels and a phantom centre image. Things might have turned out differently, but they didn't. Consequently, almost all content commercially available has zero information embedded in it to tell you how to set up a centre channel. Any attempt to do so will inevitably degrade the sound, because the centre channel is being "faked up" through interpreting data that lacks the necessary information. The accurate way to play music storing only two channels of data is with two chains - a left and a right.

But, you have a real problem to solve - and gaining acceptance from your family is more important than purism!
 
I have the Polk R700 with some hypex amp/dirac too and I'm super interested about the Toe In

It's possible to share some pictures so I can compare with my actual set up ?

How much ° is your toe in ?
The best toe in my experience is around 20° toe-in from perpendicular front wall as it keeps enough top end without too much reducing the soundstage. If you take Erins reference for "Toe in / Toe out" it would be a 10° toe-out.

What hypex amp do you have and how do you find them paired with your amp ?
 
Last edited:
Interesting. Your wife and you obviously have different ears - unless I misunderstood and she wants a different angle for purely aesthetic reasons.
For my wife it's for purely esthetic reasons and "happy wife, happy me". So I'm currently running these perpendicular to front wall.
These R700 are somewhat pricey, although not by "high-end" standards. You say you love them in every regard except directivity? Hmm, difficult. It would be all to easy to suggest "try different speakers then" lol. But here's some thoughts:

The roughly comparable R600 (just a smaller version) are what I very recently read a comment about: commenter had them and clearly preferred the recently acquired Heco Aurora 700 over them. The latter have a nicely wide dispersion in the higher treble.

Now what if, and I'm just throwing ideas around, you tried listening to the Aurora 1000 model? Just like R600 vs. R700, it's the bigger model, with 2x8" woofers instead of 2x6.5", using the same tweeter. These Heco tweeters have a little waveguide (fancily called "Fluktus" in marketing speak lol) that seems to help with width. I have the 700 models, and indeed the sweet spot seems unusually wide.

Here's some (limited) measurements:

As said, the bigger 1000 model uses the same tweeter, and apart from that it's a very similar 3-way 2x8" design as the R700.
The directivity of the R700 is very good, it's just a bit narrow at the top and the tweeter output and distortion are less stellar than the rest of the speaker.
But for anything under 8khz and/or less than 96db/1m the speaker is exceptional and will scale with your amp. I have 2x Topping L90 in mono and rendering of the system is very hard to beat unless going 2 to 3 times. Here is what I like about the R700 :
- ~4 ohm min speaker with not that much of impedance shift (3.6 to 9 ohm) meaning it's an easy load for true 4-ohm amp
- can use a lot of power (except for that tweeter that breaks above 96db)
- very good distortion figures on axis & off axis frequency response and good directivity (again except for tweeter) which makes midrange (+/- 1db) on these (slightly forward unlike KEF) very pleasing to my ears (and the very reason I don't find alternative at this price point).
- distortion free bass until 40hz (very hard to have on alternative speakers at this price point)
- they feel very substantive (33kg/77lbs) and are almost reasonance-free.

I hesitated with the HECO aurora 1000 based on Audioholics review/measurements and was mainly displeased with the bass output and cabinet build quality (but overall they were no match for R700 in the final ratings). Maybe I should try to find and listen to them.
 
For my wife it's for purely esthetic reasons and "happy wife, happy me". So I'm currently running these perpendicular to front wall.

The directivity of the R700 is very good, it's just a bit narrow at the top and the tweeter output and distortion are less stellar than the rest of the speaker.
But for anything under 8khz and/or less than 96db/1m the speaker is exceptional and will scale with your amp. I have 2x Topping L90 in mono and rendering of the system is very hard to beat unless going 2 to 3 times. Here is what I like about the R700 :
- ~4 ohm min speaker with not that much of impedance shift (3.6 to 9 ohm) meaning it's an easy load for true 4-ohm amp
- can use a lot of power (except for that tweeter that breaks above 96db)
- very good distortion figures on axis & off axis frequency response and good directivity (again except for tweeter) which makes midrange (+/- 1db) on these (slightly forward unlike KEF) very pleasing to my ears (and the very reason I don't find alternative at this price point).
- distortion free bass until 40hz (very hard to have on alternative speakers at this price point)
- they feel very substantive (33kg/77lbs) and are almost reasonance-free.

I hesitated with the HECO aurora 1000 based on Audioholics review/measurements and was mainly displeased with the bass output and cabinet build quality (but overall they were no match for R700 in the final ratings). Maybe I should try to find and listen to them.
Well it has to be said, the Aurora models are definitely one category below the R700. Both in price and especially construction. Still a "budget" series that just so happens to have turned out really well for the price. Definitely things you will notice if you're a bit "spoiled" by speakers that cost double as much: foil instead of veneer. Plastic parts (feet, terminal plates). Still very well made if I can extrapolate from my 700s, just from cheaper materials.

The bass is really good, deep and clean, but I doubt they can match the R700, which you say are exceptionally good in that regard.

I used them as an example for suggesting "what if you tried other speakers" because the tweeters are really good, especially in terms of directivity - which is your only concern about the R700 as I took it. Overall very nice sounding, and the whole speaker is very balanced and can do high volumes effortlessly too. I guess you'd be served best with R700s with a Heco tweeter. (⁠☞゚⁠ヮ゚⁠)⁠☞

I guess this is a typical situation of "you can't have everything". Still I'd say it's definitely worth to go listening to alternatives, including the Aurora 1000.
 
Last edited:
I used them as an example for suggesting "what if you tried other speakers" because the tweeters are really good, especially in terms of directivity - which is your only concern about the R700 as I took it. Overall very nice sounding, and the whole speaker is very balanced and can do high volumes effortlessly too. I guess you'd be served best with R700s with a Heco tweeter. (⁠☞゚⁠ヮ゚⁠)⁠☞
R700s with a Heco tweeter => That's a damn good idea!
Similar sized cabinet with flat front baffle... that should works. If someone happen to have detailed measurements of the heco tweeter (maybe I won't even need to rework the crossover)
 
Oh my. It was a funny comment! But if you really want to look into the Frankenspeaker idea:

Screenshot_20241209_152959_YouTube.jpg


20241209_153324.jpg


Taken from this video:


700 and 1000 models use the same tweeter. Should be cheap too. However by 99% chance it will not be a drop-in replacement. Do you really want to mess with your otherwise fine speakers that way? :D

Edit: part number is: HT 30 KF-GE 470 FED, outer diameter 104mm. Turns out, not exactly cheap as a replacement part. I found them for around 80 EUR here - each.
 
Last edited:
Why not bolt some little jig to the floor? When you want to "critically listen" you can slide the speakers up against the jig. Afterwards, you can spin them out to wherever your wife wants. You could get fancy and do it with some sort of "lazy susan" contraption, even if it's just 1 piece of wood sliding over another, with the "jig" as part of the lower panel.

jig.png


the centre channel is being "faked up" through interpreting data that lacks the necessary information
I'm not so sure, because there is common sound in both channels that can be extracted (as in mid/side processing). Then you basically have "3 audio tracks".
  • 2 channel = LeftMix + RightMix
    ⇅⇅⇅ e = ∑∞ⁿ⁼⁰ ¹ₙ ⇅⇅⇅
  • "Mid/Side" = LeftOnly + Center + RightOnly
So extracting these 3 tracks opens up some options to then re-mix them in to 3 discrete output channels (the basic options being "how much mid in the sides", and "how much sides in the mid"). I'm not sure how 3 speakers will work with regards to "acoustic interference", but speaker positioning would play a role, and there's a variety of DSP options that could improve the situation too (for each track in each mix you could consider IIR options like delays, gains, hi/low/band passing, etc. And don't forget FIR!)

From the article linked above;
MS_Channels_infographic_sonible.jpg
 
Last edited:
I'm not so sure, because there is common sound in both channels that can be extracted (as in mid/side processing). Then you basically have "3 audio tracks
I thought the same when I was a kid and started experimenting with multiple speakers and stereo. It must be straightforward, I thought, surely! But it's impossible to do properly, and an information-theory way of approaching the problem explains why.

This is an excellent paper Blumlein 88 introduced in the other thread I linked to:
 
I thought the same when I was a kid and started experimenting with multiple speakers and stereo. It must be straightforward, I thought, surely! But it's impossible to do properly, and an information-theory way of approaching the problem explains why.

This is an excellent paper Blumlein 88 introduced in the other thread I linked to:
I skimmed it. It's very much focussed on optimum sound reproduction for a single seat, and with analogue circuits/processing in mind. It barely mentions trading off "best single seat performance" for "broad area performance", nor things like custom delays via DSP (and when I say custom, I mean you could do things like delay only the highs on the center. You can chop it up heaps of ways if you have the hardware for it).
what about these options.png

Here's a paragraph from the article;
The preceding work has been applied to stereo layouts
where all loudspeakers are equally distant from a listener
at the ideal stereo seat. These decoders can be used
with other stereo loudspeaker layouts, such as the one
shown in Fig. 1(g), but the results will be less than
ideal, although they will generally still be quite ac-
ceptable to noncritical listeners
. Generally such layouts
will still tend to work quite well for very noncentral
listeners
, but give degraded results at or near the tra-
ditional stereo seat.

One more quick comment - "preserving the spatial cues of a recording" may not be everyone's goal/requirement, because some music doesn't even have a "recording". For example, I mostly listen to heavy metal, and I believe a bunch of studio stuff is recorded "solo", and then convolved with various IR's to get the desired tone of guitar/bass/drums, and stuff like that. Do they even record the sound coming out of a guitar amp any more, or just send the guitar straight in to an interface? Who knows?!

Here's a fun little video on the topic;
 
A few questions i thought of that could help us advise on improving your listening experience:
1. Room dimentions/layout?
2. What is your listening distance to the phantom center?
3. How close are your mains to the side walls?
4. Are the sidewalls first reflection points a similar layout?(or asymmetrical, like a window on one side and a sofa on the other)
5. Do the sidewalls have any sort of acoustic treatment?
6. Are you running a subwoofer? if yes which one? and are your mains highpassed?
7. if No to #6, have you considered bookshelf speakers paired with a sub or subs?
 
I skimmed it. It's very much focussed on optimum sound reproduction for a single seat, and with analogue circuits/processing in mind. It barely mentions trading off "best single seat performance" for "broad area performance", nor things like custom delays via DSP (and when I say custom, I mean you could do things like delay only the highs on the center. You can chop it up heaps of ways if you have the hardware for it).
what about these options.png

I think it's inevitable that 2-ch stereo, or multichannel derivations from 2-ch stereo, be optimized at a single seat/ single position.
Due to the simple nature of the triangulation physics.

I've found that to hold true, experimenting with 3-ch LCR implementations of Gerzon's matrices..
That said, I've also found that the 3-ch LRC rig widens the acceptable single seat area, vs 2-ch stereo.

In the rig below, L & R have centers 14ft apart. Listening distance is 14-15ft So the classic equilateral triangle.....(with back wall coincidentally 15 ft behind)

When playing normal stereo using just L&R, center image is rock solid dead center (to the point friends often make me prove center speaker is not playing).
But stays centered for only 2-3inches of lateral head movement

When playing 3-ch LCR, the center image is just as solid/dead center as in stereo.
It stays dead centered for about +/- 1.5ft of lateral movement. ......(about the same as the 3ft width of the 90H x 60V horns.
1733848540662.jpeg


As far as truly optimizing spaciousness & soundstage, I've found the timing of the arrivals is critical...they must arrive at the same time.
Gerzon made that clear, and for me it's DSP to the rescue with precise digital delays.
Can also say DSP really helps implement his and other's LCR matrices...waaay better than analog, I have to think.

The thing with LCR spaciousness and soundstage of my rig though, is that it is very recording specific, as to which I prefer, stereo vs LCR.
Really matters how the recording was mastered. Whereas, image centering as described above, is not so recording dependent.

The laptop screen is showing a custom q-sys remote that has presets for each channel mono, any two speakers stereo, and 3 types of LCR matrices.
Along with delays for each top and sub, and a sliding high-pass for the center when in any of the LCR modes.
(and a number of other general speaker controls)

Like you say, you can chop it up heaps of ways ! :D
Anyway, my findings/experiences...
 
700 and 1000 models use the same tweeter. Should be cheap too. However by 99% chance it will not be a drop-in replacement. Do you really want to mess with your otherwise fine speakers that way? :D

Edit: part number is: HT 30 KF-GE 470 FED, outer diameter 104mm. Turns out, not exactly cheap as a replacement part. I found them for around 80 EUR here - each.
Thanks for the part number, will have a look at it during the holiday. What makes you think it will not be a drop in replacement ? Its seems that the sensitivity and impedance of the tweeter are very similar...
 
Edit: I get it, the parameters of the tweeter (inc. bandwith) need to be very close for this to work. Otherwise the crossover will need to be redone (at least the high pass & crossover point); Does anyone know of the measurements and parameters of polk reserve tweeter ?
 
I'm not so sure, because there is common sound in both channels that can be extracted (as in mid/side processing). Then you basically have "3 audio tracks".
  • 2 channel = LeftMix + RightMix
    ⇅⇅⇅ e = ∑∞ⁿ⁼⁰ ¹ₙ ⇅⇅⇅
  • "Mid/Side" = LeftOnly + Center + RightOnly
So extracting these 3 tracks opens up some options to then re-mix them in to 3 discrete output channels (the basic options being "how much mid in the sides", and "how much sides in the mid"). I'm not sure how 3 speakers will work with regards to "acoustic interference", but speaker positioning would play a role, and there's a variety of DSP options that could improve the situation too (for each track in each mix you could consider IIR options like delays, gains, hi/low/band passing, etc. And don't forget FIR!)

From the article linked above;
View attachment 412703

I read the article and it was very insighful.
Do you have listening feedback on Mid/Side vs Gerzon matrix for 3 channel upmix ? The end goal is to have a wide area performance, not for critical listening but for sessions with a couple of friends (let say 3 to 4 seats) as I understand that for critical listening it would be very challenging to get right (if possible at all).

I skimmed it. It's very much focussed on optimum sound reproduction for a single seat, and with analogue circuits/processing in mind. It barely mentions trading off "best single seat performance" for "broad area performance", nor things like custom delays via DSP (and when I say custom, I mean you could do things like delay only the highs on the center. You can chop it up heaps of ways if you have the hardware for it).

Interested in your views on how to implement this via DSP (inc. FIR which the FLEX has)
A few questions i thought of that could help us advise on improving your listening experience:
1. Room dimentions/layout?
2. What is your listening distance to the phantom center?
3. How close are your mains to the side walls?
4. Are the sidewalls first reflection points a similar layout?(or asymmetrical, like a window on one side and a sofa on the other)
5. Do the sidewalls have any sort of acoustic treatment?
6. Are you running a subwoofer? if yes which one? and are your mains highpassed?
7. if No to #6, have you considered bookshelf speakers paired with a sub or subs?
1. Room is a 7x5m rectangle (european appartment) with roughly 4.5x5m as listening area.
2. Listening distance to the phantom center is 3-4m depending on the seating position
3. Main are very close to front wall (I managed to have ~1 foot), very close on one side wall and open on the other
4. Asymmetrical (a wall and glassdoor at 1 foot on one side and open to almost 4m on the other)
5. No acoustic treatment (rented appartment)
6. No subwoofer needed. R700 can reach 26hz (dirac measures) in room.
7. No (reasons below)

I previously had a system with 2 presonus eris e8 xt + 2 subwoofers (presonus as well) and crossover was a pain to setup and the system was somehow output limited and I hated the raised noise floor when using DSP.
To be honest my listening experience is wonderful as polk r700 really shines with the topping LA90 (and probably with any state of the art amplifier) and I have played and measured enough position to have the best experience. Now, I'm trying to accommodate for more comfort and use cases (wife, friends...) without losing to much.
 
How's this look? Can i make any adjustments to increase accuracy? Where is the glass door?

Layout8.JPG
 
Back
Top Bottom