$30K Budget - On the quest for my "end game" speaker

[benanders]

Just to be clear. When I'm referring to the multiple subwoofer approach I'm talking about randomely placing three or more subwoofers working at the same frequency area in the room with the goal of achieving an even response.
I'm not talking about simply using several subwoofers in order to get lower distortion and reach necessary SPL. The latter is something I find crucial for a high-end setup unless one is using one subwoofer with super high SPL (like a hugh horn subwoofer).

In regards to capacity I'm not really sure multiple subwoofer approach where one or two are placed in the opposite direction works that well. My experience is that this approach leads to a lack of slam and tactile experience compared to having all in the front pointing in the same direction. Perhaps because they end up partially working against each other.

I tried the opposing-subwoofers approach (anecdotally and subjectively).
I did so on recommendation of a DIY’er of a L / R channel opposing-subwoofer approach (that case linked here) :

1 day ghetto line array build < £200

Bumping my own thread to say these speakers have been in use for a year or so now and they are holding up well. Treble and midrange very clear. Bass is tight and accurate. I’ve played everything I could think of on them. No unpleasant surprises. If I want to impress, I play a female vocalist...

www.diyaudio.com

I suggested the reason his original horizontal orientation of opposing-subs may have been more satisfactory because all 4 woofers were in vertical alignment, vs. vertical orientation of 2 woofers aided by gravity and 2 woofers opposed by gravity (early pics vs. that last pic in the dyiAudio link). Not sure how much effect gravity would have on driver reflexes, but probably “ > 0” ?

I did not attach my opposing-subs so securely to each other, rather both chassis were fastened to a beam frame, and given my heavier and stronger subs specs, this may have been an issue when I tried it. I’ve not discussed it here (ASR) as I did not take measurements. No need - preliminary listening said the opposing-sub approach, for me and my kit, was a “No”.

(1) Stacking both of my forward-facing subs (each receiving stereo signal) vs. (2) using them as separate stereo L / R forward-facing subs, vs. (3) putting them in different parts of the room (again, with each receiving a stereo signal) vs. (4) running them in mounted opposition, all give different sounding results. Subjectivity alert.

Long story short, I agree @Bjorn it is very important to clarify which “multiple sub” use is employed and/or recommended. They must surely produce different results including as measured on plots in a given room - any takers with the gear to test for this or validate via example?

 

[Bjorn]

@Bjorn how much do you weigh the time domain response?
I am assuming that frequency domain can be largely fixed with DSP - but I am not fully comprehending the minimum phase part you mentioned.

I don't want to bring this too off-topic, but the short answer is that I highly value the time domain behaviour both in regards to the speaker design and the room. Amplitude follows the time domain, not the other way around.
When I pointed out the trouble of EQing certain areas of the speaker, that part very much effects the frequency response negatively too. It can get worse in other directions.

About subwoofers. Stacking is considered by many to make it into one single unit, but it's not quite that the straightforward. Below are two graphs. The blue graph shows two subwoofers, placed behind in each front speaker with an Xover of 120 Hz. With the red graph one subwoofer has been added on top of the other (total of three). As we can see, it's not only an improvement and besides a bit less SPL below 40 Hz we have a fairly broad cancellation with center at 72 Hz. Some other areas are improved though. This is quite typical with stacking.

 

[Absolute]

Long story short, I agree @Bjorn it is very important to clarify which “multiple sub” use is employed and/or recommended. They must surely produce different results including as measured on plots in a given room - any takers with the gear to test for this or validate via example?

Just to clarify myself as I see I can easily be misunderstood. With the frequency response equalized to the same in the listening position I've yet to see a significant difference between multi-sub configuration in the overall decay times at low frequencies (50-60 hz and down). I've read the Fazenda-paper exploring the differences between some configurations, but I don't find their method particularly relevant as they never matched the frequency response between the configurations and all the configurations were crossed as high as 120 hz- with no care taken whatsoever for integrating the subs to the mains in the time domain. So many factors that can potentially skew the results.

I experimented quite a lot with active cancellation myself, but it was very hard to reduce the overall decay at any significant factor compared to other configurations due to constructive interference in some areas and destructive in others. Without FIR filtering and steep slopes on the destructive sub it's difficult to achieve only positive contributions. The main target for destruction may yield a good gain in reduction of decay time, but having to compensate with EQ for the part of the range that suffers from unwanted interference gave me no clear overall benefit. At no point have I tried with more than three low frequency sources nor with FIR filtering of individual sources, so keep that in mind.

Based on what I've seen, tried and tested I'm not convinced there's much inherent difference between stuff at the low frequencies. Above 50-60 hz we become highly sensitive to the time domain and the room interaction becomes complicated real fast, so I think the room treatment should start to be effective from around there.
DBA with enough drivers might stretch it by about an octave or so - I think.

Your mileage may vary and all that.

 

[hvbias]

Many of the speakers that we see scoring high in spinoramas are in my opinion quite mediocre. Place them in a room, and the response in part of the frequency range will not be particular even and there are other issues like intermodulation distortion and thermal distortion. Fixing uneven response with "room correction" doesn't work very well, and the distortion issues can't be solved either. Correcting parts of the uneven response with treatment is difficult and sometimes impossible.

Why would the bolded part occur?

 

[Gringoaudio1]

Why would the bolded part occur?

That would happen to any speaker. The room causes the uneven FR at the sitting position. Best to start with a speaker capable of a good FR.

 

[TimW]

Why would the bolded part occur?

One issue that has always bothered me is the floor-bounce null. You can put bookshelf speakers with an excellent spinorama like the KEF R3 or March Audio Sointuva on a stand and in-room you may get a suck out around 100-200 Hz depending on stand height. This suck out is audible and right where you want solid bass punch. Tower speakers with multiple bass woofers have multiple null frequencies which average out to a smoother in-room response. So while the tower may not have as good of a spinorama as the bookshelf, it may measure better in-room.

 

[gnarly]

That would happen to any speaker. The room causes the uneven FR at the sitting position. Best to start with a speaker capable of a good FR.

Sure, all speakers will have some degree of floor bounce.

That said, two classes of speakers that excel at mitigating floor bounce are: (that I have experience with)

-Line array type speakers......including floor to ceiling line arrays, CBT line arrays; and rightly or wrongly, i put electrostats, planars, tall ribbons, etc into the line array family too.
Ime, all these have all significantly reduced floor bounce vs more traditional speakers.

-Larger horns / waveguides. Their various wall angles/curvature designs, control directivity considerably lower in frequency than conventional speakers.
The larger the horn/waveguide the lower the frequency of maintaining pattern control.
My experience with large conical horns with vertical patterns in the 60 degree zipcode, is that they work very well at mitigating bounce down to their patten control frequency.

Neither of these two classes of speakers seem likely to appear in Amir's reviews....both large, and often bulky.
(I've often wondered how well the Klippel rig could handle them....even if they could be brought to it.)

 

[Bjorn]

Why would the bolded part occur?

Floor bounce is one reason as mentioned. Secondly, they tend to loose directivity quite high in frequency. A great speaker design should be constant down the Schroeder frequency. Another cause is how the speakers measure vertically. A wide vertical directivity causes a more uneven frequency response and get's worse with the traditional lobing issues.

 

[Bjorn]

(I've often wondered how well the Klippel rig could handle them....even if they could be brought to it.)

A large horn needs to be measured in farfield. The CBT can be measured up close.

 

[gnarly]

A large horn needs to be measured in farfield.

Do you have a good rule of thumb you use, for determining far-field?
I've always just gone with the old 3x longest speaker dimension, wondering how well it works when applying it to horn mouth dimensions.

 

[Bjorn]

Do you have a good rule of thumb you use, for determining far-field?
I've always just gone with the old 3x longest speaker dimension, wondering how well it works when applying it to horn mouth dimensions.

This is an area I'm also wondering about. While I noticed that beam width became narrower and closer to simulation when going from 0.6m to 2.9m of a 1.1m wide midrange horn, 2.9m still seemed to be too close.

I recently learned that Dr Eugen Patronis said to use 10x the largest dimension of the DUT. But it will also depend whether the DUT emits high frequency from its entire frontal surface or not. If it doesn't and which is the normally the case, half the distance (5x of the largest dimension) is likely good enough.

This is something I have to experiment with and also with different speaker designs.

 

[gnarly]

This is an area I'm also wondering about. While I noticed that beam width became narrower and closer to simulation when going from 0.6m to 2.9m of a 1.1m wide midrange horn, 2.9m still seemed to be too close.

I recently learned that Dr Eugen Patronis said to use 10x the largest dimension of the DUT. But it will also depend whether the DUT emits high frequency from its entire frontal surface or not. If it doesn't and which is the normally the case, half the distance (5x of the largest dimension) is likely good enough.

This is something I have to experiment with and also with different speaker designs.

Thanks for that.

It is ironic to me, how it often takes more distance to get into the far-field for HF, than for low frequencies. Ironic, given conventional wisdom....

My hope with my large conical horns using a compression driver, is that they represent a very small HF source, not needing much distance....
(Well, small other than all the HF diffraction from my poor straight sided mouth terminations ...haha yikes)

 

[FrankW]

For the way our hearing apparatus works and because of physics, the best way to have smooth bass response in a room, is multiple subwoofers and DSP. Best pass almost mandates DSP. One can debate ad infinitum, the two best known approaches, at this point in time that seem to work and well are The Harman/Toole/Welti/Devantier methos that preconizes subwoofer at the midpoint in a rectangular room, or the less documented Geddes method that goes for 3 subwoofers in a quasi-random positioning, one of those in the front corner... Both work.
One could retort that some speakers are full range and may not need subwoofers, Well this is true, but best bass even with full range requires DSP... we coudl actually call this a mandate. In some ways, the woofer section of those full range are in fact low frequency transducers .. in that case two... Thus multiple subwoofers if it really full range full dynamics.. Stretching it a bit but there is truth to that. Same speakers + external subwoofers, even inexpensive subwoofers would have presented better bass reproduction in most rooms , most of the times.

From reading AVS and other forums, that's been my understanding also, multi-subs, usually around 80hz to have smooth bass and avoid localization. But the linked post you were responding to is interesting, as it seems to be saying something quite different, that localization is not the issue with the standard 80-90hz recommendation, if I'm understanding it correctly.
https://www.audiosciencereview.com/...than-digital-right.37657/page-19#post-1490035
That there is a "flaw" in this idea?? Btw, it that "J_J" James D Johnston?

 

[BDWoody]

From reading AVS and other forums, that's been my understanding also, multi-subs, usually around 80hz to have smooth bass and avoid localization. But the linked post you were responding to is interesting, as it seems to be saying something quite different, that localization is not the issue with the standard 80-90hz recommendation, if I'm understanding it correctly.
https://www.audiosciencereview.com/...than-digital-right.37657/page-19#post-1490035
That there is a "flaw" in this idea?? Btw, it that "J_J" James D Johnston?

JJ's Signal Processing and Psychoacoustics Master Library

We are privileged to have James Johnston ("JJ") as one of our members. As one of the luminaries in audio research, it has been a pleasure to read his past presentations to the local chapter of Audio Engineering Society. JJ has kindly given me permission to post links to his work per below. I...

www.audiosciencereview.com

 

[FrankW]

JJ's Signal Processing and Psychoacoustics Master Library

We are privileged to have James Johnston ("JJ") as one of our members. As one of the luminaries in audio research, it has been a pleasure to read his past presentations to the local chapter of Audio Engineering Society. JJ has kindly given me permission to post links to his work per below. I...

www.audiosciencereview.com

Oh my. Thanks! Still sifting through this thread...

 

[MKR]

32 days to AXPONA! Countin em down!

 

[mkt]

Just to keep things honest. Genelec and I have made some peace. They told me I have a path to complete return at no financial loss.

I’ve decided to rather investigate and work with them rather than return them. Because I really do like them.

I think they were really big on checking all the electronics first and then allowing me this testing.

Can you share how this turned out?

 

[puppet]

Don't know if this particular loudspeaker was mentioned ... and I don't have any idea on the cost but they look interesting. They have a dealer here in the USA and that dealer will be at AXPONA : https://bluesmokesystems.com/

This is the loudspeaker: https://analogdomain.eu/index.php?option=com_content&view=article&id=33Itemid=33

 

[dkinric]

Don't know if this particular loudspeaker was mentioned ... and I don't have any idea on the cost but they look interesting. They have a dealer here in the USA and that dealer will be at AXPONA : https://bluesmokesystems.com/

This is the loudspeaker: https://analogdomain.eu/index.php?option=com_content&view=article&id=33Itemid=33

Looks interesting - 99db sensitivity with 1650+ watt power handling. Nice looking speaker IMO too. Wow, bet these things can put out some SPL. Only checked one source, but these seem to be in the $50k range, which would be well over OPs budget.

 

[puppet]

50K ... ugh.