Pushing the speakers back for flatter bass

[tuga]

I was experiencing a somewhat 'hollow-sounding' bass with my new speakers – there was a lot of sub-bass but a chunk of the mid- and upper-bass was missing (see blue curve below, dip between 80Hz and 200Hz) – and after reading about the effects of boundary inteference (SBIR) and how the Dutch&Dutch 8c deals with it, I decided to push back the speakers as much as possible and then EQ the peaks off, hoping that the wide dip in the low bass would be reduced in width and depth.
I used REW's room simmulator and an online SBIR calculator to determine the starting point and after many measurements and position adjustments the listening spot and speakers ended up not far from the theoretical optimal location, and, above all, with a significantly more balanced response in the bass and improved clarity.

Blue initial, red final location. Dotted raw, solid EQ'd.

 

[Roog]

Nice one, I have found REW a great tool because it explains very clearly the issue you are hearing And that has to be the starting point, know what is wrong and look for a solution to it. As opposed to open wallet and buy some more boxes/cables/rubber feet Over and over

 

[AndyU]

There is an article on Sonarworks’ blog which explains the rationale behind your strategy:

Should You Place Speakers Against the Front Wall? - Sonarworks Blog

Does putting a speaker near the front wall increase the amount of bass or you should move them away from the front wall? Here's everything you need to know!

www.sonarworks.com

 

[windhoek]

I adjusted the attenuation knobs on my KLH Model 5s yesterday as I suddenly became aware that it was too treble focussed: I'd been listening to the album London Calling and it was blindingly obvious that the vocals were lacking meat. So I set the attenuation knobs to Lo (-3db applied to treble and upper mids) and sure enough, everything sounded much more convincing, much more lifelike. Tone controls or attenuation knobs on speakers? Hell yeah!

 

[tuga]

There is an article on Sonarworks’ blog which explains the rationale behind your strategy:

Should You Place Speakers Against the Front Wall? - Sonarworks Blog

Does putting a speaker near the front wall increase the amount of bass or you should move them away from the front wall? Here's everything you need to know!

www.sonarworks.com

That's a good piece.
This diagram by Neumann indicates that the worst position for the woofer is between 0.80m and 2.00m (bottom left table), and shows the advantage of using subwoofers in minimising SBIR effects.

 

[tuga]

Also worth reading this Dutch & Dutch 8c review.

 

[Tony L]

It is important to recognise different speakers are designed and voiced for different room positions. Far better IMHO to pick the right speaker for the placement than pick a wrong one and try to force it to work with aggressive EQ. Things like baffle step, full/half/quarter-space placement are key here. They are designed-in.

Some modern active speakers can be switched between locations in a way passive devices can’t, some even correcting for different or poor location, but my instinct is to recognise the original design intent and start from that position.

 

[barryb]

I had a similar 50hz mode I could not tame to my satisfaction in my room. Ultimately I chose bass limited speakers (LS3/5A) which really I should have went with from the start, if only I could have overcome my prejudice. The point-source effect and even-handed response more than make up for lack of bass, which in my room was an all-encompassing distraction.

Does the bass dip now present a new issue? I find the kick in a lot of stuff is around here. I'm now listening to some Daniel Avery, headbanger stuff to some, and don't feel wanting. The illusion is real!

As ever, for most of us, speaker choice and siting is series of compromises, and what you can live with.

 

[AndyU]

Floor bounce is also significant, typically creating a dip around 200Hz, and much harder to fix.

Floor/Ceiling Reflection Calculator

 

[ToTo Man]

I was experiencing a somewhat 'hollow-sounding' bass with my new speakers – there was a lot of sub-bass but a chunk of the mid- and upper-bass was missing (see blue curve below, dip between 80Hz and 200Hz) – and after reading about the effects of boundary inteference (SBIR) and the Dutch&Dutch 8c deals with it, I decided to push back the speakers as much as possible and then EQ the peaks off, hoping that the wide dip in the low bass would be reduced in width and depth.
I used REW's room simmulator and an online SBIR calculator to determine the starting point and after many measurements and position adjustments the listening spot and speakers ended up not far from the theoretical optimal location, and, above all, with a significantly more balanced response in the bass and improved clarity.

Blue initial, red final location. Dotted raw, solid EQ'd.

Nice result! Have you noticed any deleterious effects on soundstage depth with the new position?

What software was used to produce the above graph, and how much smoothing was applied?

 

[tuga]

Nice result! Have you noticed any deleterious effects on soundstage depth with the new position?

What software was used to produce the above graph, and how much smoothing was applied?

Unfortunately the speaker are now in the alcoves and the upper-midrange smoothness has been slightly affected by the mantelpiece but I have not noticed any other adverse effects.
The Kef coaxial mid-tweeter has a narrow directivity above 350Hz which probably helps.

I've used REW for this plot and psychoacoustic smoothing.

I set the smoothing to 1/12 Octave (to avoid an overly small listening spot) and the EQ filters are generate automatically to a 3dB flatness target with 0dB individual and overall boost.
I then tweak each filter manually to avoid high Q / narrow-bandwidth filters below 300Hz.

 

[tuga]

It is important to recognise different speakers are designed and voiced for different room positions. Far better IMHO to pick the right speaker for the placement than pick a wrong one and try to force it to work with aggressive EQ. Things like baffle step, full/half/quarter-space placement are key here. They are designed-in.

Some modern active speakers can be switched between locations in a way passive devices can’t, some even correcting for different or poor location, but my instinct is to recognise the original design intent and start from that position.

Regardless of how a speaker is designed the SBIR effects will always take place so avoiding the 80-200cm range is a good idea. Even though my current living room is 6.45m deep I am unable to have the speakers farther than 2m out into the room.
Some speakers have been designed with built-in acoustic full/half/quarter-space placement compensation but many studio monitors have electronic/EQ compensation.

Floor bounce cancellation depends on the height of the woofer and the distance to the listener and is not a problem if one can move the listening spot. My speakers have two woofers equidistant to the coax driver so the floor bounce effects are somewhat minimised.

According to the specification HQPlayer performs DSP with 64/80/128-bit and arbitrary precision floating point processing and a PCM-equivalent sample rate of 768kHz, far more accurate than any electronic or most DSP EQ devices.

More worrying in my view is interaction with side walls, and that is where a cardioid design like the Dutch & Dutch or a dipole excell.
And we still have to contend with the room effects at the listening spot which vary with positioning.

For those not familiar with full/half/quarter-space placement gain this diagram will help (I think that the gain happens below 200Hz):

 

[tuga]

I had a similar 50hz mode I could not tame to my satisfaction in my room. Ultimately I chose bass limited speakers (LS3/5A) which really I should have went with from the start, if only I could have overcome my prejudice. The point-source effect and even-handed response more than make up for lack of bass, which in my room was an all-encompassing distraction.

Does the bass dip now present a new issue? I find the kick in a lot of stuff is around here. I'm now listening to some Daniel Avery, headbanger stuff to some, and don't feel wanting. The illusion is real!

As ever, for most of us, speaker choice and siting is series of compromises, and what you can live with.

I have dealt with the 52Hz peak using EQ and by moving the speakers very close to the front wall I no longer have a wide dip between 80Hz and 200Hz (see solid red trace above).
Having lived with large 3-way speakers in (acoustically) reasonable rooms I could never go back to small standmounts.
My speakers have their ports tuned too high to 38Hz so I may in the future add a couple of subs.

EQ-ing a room-generated peak corrects both the frequency/amplitude and the time/decay response:

 

[barryb]

Ah, I had misread FR in my haste - looks good indeed. What speakers are you running the days?

 

[tuga]

Ah, I had misread FR in my haste - looks good indeed. What speakers are you running the days?

Kef Reference 3s. I bought a somewhat bruised, second had pair for a reasonable price.

 

[ToTo Man]

EQ-ing a room-generated peak corrects both the frequency/amplitude and the time/decay response

I recall having a rather heated discussion with our old chum Keith Cooper about this. The way I view it, reducing a modal peak with EQ does not dampen the ringing per se. The mode still rings out for the same amount of time but, because it's starting off from a lower amplitude, a larger portion of the reverb 'tail' is pushed below the noise floor of the room, so the ringing reaches the point of inaudibility quicker. Even if you went crazy with EQ and levelled out the peaks so much that you effectively has a completely flat FR response at the listening position, the waterfall graph will still show more ringing from the modal frequencies, especially the low order modes. AIUI, the only true way to dampen ringing is to absorb its energy, like how a mat damps the resonance of a drum when placed on its skin, either with copious amounts of passive or active bass absorption,

 

[tuga]

I recall having a rather heated discussion with our old chum Keith Cooper about this. The way I view it, reducing a modal peak with EQ does not dampen the ringing per se. The mode still rings out for the same amount of time but, because it's starting off from a lower amplitude, a larger portion of the reverb 'tail' is pushed below the noise floor of the room, so the ringing reaches the point of inaudibility quicker. Even if you went crazy with EQ and levelled out the peaks so much that you effectively has a completely flat FR response at the listening position, the waterfall graph will still show more ringing from the modal frequencies, especially the low order modes. AIUI, the only true way to dampen ringing is to absorb its energy, like how a mat damps the resonance of a drum when placed on its skin, either with copious amounts of passive or active bass absorption,

I agree, you can only treat ringing and also cancellation with acoustic treatment.
But from my experience by bringing down the peak then the decay tail generally also falls back reasonably 'in line' with the decay of surrounding frequencies making it perceptually (almost) innocuous. Perhaps I should have put decay in inverted commas.

 

[ToTo Man]

@tuga, when you say that you use low-Q filters when flattening out the bass peaks, can you give me an example of some Q values and why you choose low Q values instead of high Q? My approach is to look at the low frequency response with 1/48 smoothing and use a Q that's low enough to encompass the full width of the peak but high enough to minimise attenuation of adjacent frequencies.

The EQ AU plug-in I'm using at the moment has an 'octave width' parameter instead of 'Q'. The narrowest octave width that can be chosen is 0.06 and the widest is 3. The previous EQ AU plug-in I used has a 'Q' parameter and the narrowest Q that can be chosen is 20 and the widest is 1. The filters I use to reduce my 51.5Hz and 112Hz peaks both have a width of 0.16 octave. As you can see in the 3rd graph, a 0.16 octave width corresponds to Q=19.

 

[tuga]

@tuga, when you say that you use low-Q filters when flattening out the bass peaks, can you give me an example of some Q values and why you choose low Q values instead of high Q? My approach is to look at the low frequency response with 1/48 smoothing and use a Q that's low enough to encompass the full width of the peak but high enough to minimise attenuation of adjacent frequencies.

The EQ AU plug-in I'm using at the moment has an 'octave width' parameter instead of 'Q'. The narrowest octave width that can be chosen is 0.06 and the widest is 3. The previous EQ AU plug-in I used has a 'Q' parameter and the narrowest Q that can be chosen is 20 and the widest is 1. The filters I use to reduce my 51.5Hz and 112Hz peaks both have a width of 0.16 octave. As you can see in the 3rd graph, a 0.16 octave width corresponds to Q=19.

My understanding is that high Q filters are effective only in a very specific/narrow area, that's why I avoid them.
If you apply EQ over an MMM measurement this problem should be minimised, I find it easier to take a sweep measurement and then apply some smoothing. I started EQ'ing over 1/24 octave but later changed to 1/12.

Initially I would not tick the "Allow narrow filters below 200Hz" option in the "Filter Tasks" menu but now I let it create the filters freely and then change the Q value manually to a maximum of 7.5, sometimes 10, and also the amount of attenuation. But as you rightly point out this will affect adjacent frequencies which is why I also play with the attenuation value and sometimes even the centre frequency.

 

[tuga]

Interesting piece on audioXpress:

Boundary Effects on Sources at Low Frequencies

VC Spotlight: Boundary Effects on Sources at Low Frequencies

Christopher J. Struck, (CJS Labs) writes about Boundary Effects on Sources at Low Frequencies

audioxpress.com