• 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!

Bass and subwoofers

On multiple occasion I have tried to explain how it works in my room. It's about not supporting the modes, you have to consider floor bounce, you need to have multiple bass radiators distributed in the vertical plane for left and right channel and strategically placed sub(s). You also need to have corresponding delays across the frequency band with regards to the listening position. Phase also largely matters. Mains either cardioid or sealed, sub also sealed (in order to have net displacement and pressurization).

Not having cardioid mains/subs I had to choose half of wavelength of 110Hz for mains spacing (three bass radiators distributed vertically per channel), with sub in the middle, so quarter of the frequency. The entire system also quarter of the frequency from the front wall. No high pass to the mains, with 200Hz lowpass to the sub.

This is vector average of the floor bounce, showing +/-3dB and magnitude falling like a brick above 200 and bellow 55Hz. Also no more than 23 degrees of excess phase 110Hz and bellow:

Vector average mag-phase.jpg


Wavelet of the measurement, note the early decay, peak energy curve, also what happens at 110Hz and bellow:

Vector average ground.jpg


There's no sub bass at the floor underneath my seating position. So where is it? At ear level, delayed to where the main listening position is (10-12ms, for 10-12feet distance), close to the rear wall, where particles have nowhere to move:


01.jpg


Raw magnitude response at ear level, MLP, prior to any EQ, showing as little as 0.5 dB difference in sub bass for the left and right channel (purple/pink). Vector average of the two in red and summation of both channels in cyan:

05.jpg


Phase/mag summation across the full spectrum:

07.jpg


If there's decorrelation in low frequencies, I want to hear it, and I do.
 

Attachments

  • 05.jpg
    05.jpg
    284.2 KB · Views: 60
...rooms are different and what we observe in one of them is different than another. (1)
...enclosed spaces mess up the timing. (1)

...I wonder whether your uncorrelated pink noise was played at the same SPL in both channels... (2)

We already know from other studies (like those of your colleague Makivirta) that sensitivity to group delay is poor for longer wavelengths... (3)

1) Indeed. The four experiments therefore only probed how sensitive humans are to (random) inter-aural LF fluctuations in level and time.

We considered the topic entirely from a perceptual and physiological perspective, hence the term. However, the low frequency limit to AE, suggested in Test D, should be taken with a grain of salt. Random stimuli at very low frequency really don’t give the auditory system much to lock on to. With natural sound or music, there would normally be inter-aural VLF patterns instead, making AE audibility and joy potentially extend as low as the reproduction system is capable of.

2) L and R had the same SPL, in the correlated as well as in the uncorrelated samples. “Uncorrelated” refers to time domain.

3) Right. Such experiments are valid, and Floyd Toole provides compatible references in Sound Reproduction (3rd edition). However, the world turns upside down, physiologically and perceptually, when it is about inter-aural timing rather than absolute timing.
 
The story about Lorin Maazel in Thomas's paper reminded me of something:
[...] A spatial processing system with dozens of loudspeakers had been installed. Maestro Maazel was unhappy with the hall acoustically, and only satisfied when a slight amount of (slow) LF inter-channel fluctuation was also added, resembling what he was used to. [...]
Some time ago, I was looking at some low-passed music tracks on a (software) XY oscilloscope. Several tracks from Dance Fever by Florence + the Machine have seemingly very deliberate inter-channel phase modulation in the subwoofer range. The clearest (although somewhat brief) example is the intro to track 8 (Cassandra):


Apologies for no audio, but I didn't want to wrestle with the copyright detection :). Another good example is in track 1 (King), starting around 2:40:


Both of these have an approximately linear phase 10th order (60dB/oct) Butterworth low pass applied. Filter magnitude and phase:
bw10_filter.png
 
Last edited:
@Thomas Lund This is excellent, appreciate the discussion. I have always preferred full range mains (particularly cardioid) to monitor plus sub setups and now I have a better understanding as to why! Even a monitor (or bookshelf) plus sub (monopole, dipole, etc) that supports <50Hz or so I do not experience the same level of “AE” (nice to have a term to ascribe to my experience) as true full range towers (floorstanders). Good stuff, thank you.
This aligns closely with my own experiences (and many others who optimize their systems using my scripts) with the 'LFE + Main' mode, where the front mains play full range without being crossed over, while properly time aligned subwoofer(s) play along with them low-passed at some FR/GD optimized frequency. Initially, there was significant debate about this approach, with many arguing that every speaker must be crossed over with a subwoofer to prevent potential damage at high volumes. However, 99% of the time, this mode was eventually preferred and sounds audibly better.
 
This aligns closely with my own experiences (and many others who optimize their systems using my scripts) with the 'LFE + Main' mode, where the front mains play full range without being crossed over, while properly time aligned subwoofer(s) play along with them low-passed at some FR/GD optimized frequency. Initially, there was significant debate about this approach, with many arguing that every speaker must be crossed over with a subwoofer to prevent potential damage at high volumes. However, 99% of the time, this mode was eventually preferred and sounds audibly better.
Can you expand on "sounds audibly better"? How was it evaluated?
 
This aligns closely with my own experiences (and many others who optimize their systems using my scripts) with the 'LFE + Main' mode, where the front mains play full range without being crossed over, while properly time aligned subwoofer(s) play along with them low-passed at some FR/GD optimized frequency. Initially, there was significant debate about this approach, with many arguing that every speaker must be crossed over with a subwoofer to prevent potential damage at high volumes. However, 99% of the time, this mode was eventually preferred and sounds audibly better.

It will depend on the room. It's pretty common to have severe dips below 100hz in the main speakers, and this cannot be fixed by correctly placing subwoofers if the mains aren't highpassed. So it is a game of compromises, choosing what is deemed more important.
 
It will depend on the room. It's pretty common to have severe dips below 100hz in the main speakers, and this cannot be fixed by correctly placing subwoofers if the mains aren't highpassed. So it is a game of compromises, choosing what is deemed more important.
Stereo bass and depth is the two things I would never exchange so it's down to preference too,despite the compromises.
 
Stereo bass and depth is the two things I would never exchange so it's down to preference too,despite the compromises.
That is what compromise is, chosing something over something else. :) The best compromise will be different for different persons as you say.
 
That is what compromise is, chosing something over something else. :) The best compromise will be different for different persons as you say.
It's down to ways to address these compromises too.
I mean I can fix my room so to be decent (degree of decency varies of course) down low with some treatment,it's doable while also having AE.
It's impossible though the other way around with mono bass,etc.

(and make bass modules,I'll say it again! :p )
 
It will depend on the room. It's pretty common to have severe dips below 100hz in the main speakers, and this cannot be fixed by correctly placing subwoofers if the mains aren't highpassed. So it is a game of compromises, choosing what is deemed more important.
I actually make it work ok here. I do a bit like Geddes, where the mains are playing full range - max a first order HP - to remove the very lowest frequencies, which the speaker cant reproduce anyway.
I then add 4 subwoofers at different locations with different cross-overs, overlapping with the mains. This smooths out the frequency response pretty well.
Around and in the seating position, it never gets much worse than this - none smoothed:
KEF coax + bas + sub - lytte.jpg

The measurement is with 4 subwoofers and two mains at the same time, but only one midrange/tweeter, which is why you see a dip between 500-1000Hz.
Around 2-300Hz, I think it's the back wall canceling the woofer's response out.
Woofers are HP with a first order at 60Hz and two subs are crossed at 80Hz and two others at 110Hz, with a LR96dB LP.
Subwoofers are all around the apartment living room, being as random as possible, but still connected two on the right to the right mains channel, and two on the left to the left mains channel.
So in essence, that makes 6 bass sources.
 
I actually make it work ok here. I do a bit like Geddes, where the mains are playing full range - max a first order HP - to remove the very lowest frequencies, which the speaker cant reproduce anyway.
I then add 4 subwoofers at different locations with different cross-overs, overlapping with the mains. This smooths out the frequency response pretty well.
Around and in the seating position, it never gets much worse than this - none smoothed:
View attachment 421995
The measurement is with 4 subwoofers and two mains at the same time, but only one midrange/tweeter, which is why you see a dip between 500-1000Hz.
Around 2-300Hz, I think it's the back wall canceling the woofer's response out.
Woofers are HP with a first order at 60Hz and two subs are crossed at 80Hz and two others at 110Hz, with a LR96dB LP.
Subwoofers are all around the apartment living room, being as random as possible, but still connected two on the right to the right mains channel, and two on the left to the left mains channel.
So in essence, that makes 6 bass sources.

I would expect quite a bit more even response below 100hz with 4 subwoofers. What software is this measured with?

EDIT: I see the dips are not as deep as I first thought, so yes this is pretty good. :)
 
I would expect quite a bit more even response below 100hz with 4 subwoofers. What software is this measured with?

EDIT: I see the dips are not as deep as I first thought, so yes this is pretty good. :)
Thank you... I'm always learning, and sometimes you just need to throw it out there, to get some fresh perspective, like when you have friends over and let them spill the beans on their thoughts :D
I use a Groundsound DCN28 and the accompanying software Xoverwizard.
https://groundsound.com/dcn28.php
https://groundsound.com/download.php

I can't really fault it. I use a Roland Quad capture sound card, calibrated cross-spectrum microphone, almost reflection free sound boom, and the software uses loop-back for better overall accuracy and phase precision.
 
Can you expand on "sounds audibly better"? How was it evaluated?
By 'sounds audibly better,' I’m referring to the clear distinction between poorly controlled, boomy bass and well-balanced, properly tuned bass in a room. The difference is unmistakable and does not require formal testing to perceive. In fact, if bias were influencing the evaluation, it would more likely lean toward favoring the opposite outcome in this specific scenario.
 
It will depend on the room. It's pretty common to have severe dips below 100hz in the main speakers, and this cannot be fixed by correctly placing subwoofers if the mains aren't highpassed. So it is a game of compromises, choosing what is deemed more important.
Yes room modes would persist in all scenarios but you can improve it in the summation response with correct time alignment. With XO's you will need to leave that dip with the sub.
 
By 'sounds audibly better,' I’m referring to the clear distinction between poorly controlled, boomy bass and well-balanced, properly tuned bass in a room. The difference is unmistakable and does not require formal testing to perceive. In fact, if bias were influencing the evaluation, it would more likely lean toward favoring the opposite outcome in this specific scenario.

Once you hear it, you cannot unhear it. :)

I see no other way in hearing the effect reliably, other than having little to no audible difference in between the channels at those frequencies:

L-R sub.jpg


This is why it works so well outdoors, and is quite hard to get it to work indoors.
 
  • Like
Reactions: MKR
By 'sounds audibly better,' I’m referring to the clear distinction between poorly controlled, boomy bass and well-balanced, properly tuned bass in a room. The difference is unmistakable and does not require formal testing to perceive. In fact, if bias were influencing the evaluation, it would more likely lean toward favoring the opposite outcome in this specific scenario.
As I keep saying, I wish I would get $1 for every person who says that and then fails a proper double blind test.
 
As I keep saying, I wish I would get $1 for every person who says that and then fails a proper double blind test.
I don’t think that approach would make you rich, and I genuinely wish you could step away from that rhetoric for a moment to truly listen to your systems as they were designed to be experienced.
 
I don’t think that approach would make you rich, and I genuinely wish you could step away from that rhetoric for a moment to truly listen to your systems as they were designed to be experienced.
Thanks, but as I designed my own system, I might actually know better than you how they were designed to be experienced. And as an engineer I know how fallible our ears (or rather the stuff between the ears) are, so I know not to use them for determining objective reality. Other people prefer to live in a subjective perception of what they believe reality is. I think my approach is referred to by the term "science" (that also appears in the name of this forum), whereas your approach is usually referred to as "belief".
 
Thanks, but as I designed my own system, I might actually know better than you how they were designed to be experienced. And as an engineer I know how fallible our ears (or rather the stuff between the ears) are, so I know not to use them for determining objective reality. Other people prefer to live in a subjective perception of what they believe reality is. I think my approach is referred to by the term "science" (that also appears in the name of this forum), whereas your approach is usually referred to as "belief".
Just a note,are you sure you want to address OCA the same way we address snake oil stuff?
See his signature.

As we saw,the effect is real,unless you want to invalid Lund's paper too.I believe he's a proper engineer specialized in the field and Genelecs are nice.
 
  • Like
Reactions: MKR
Just a note,are you sure you want to address OCA the same way we address snake oil stuff?
No matter who you are, I want to know on what evidence your claims are based.
See his signature.
Not sure how calling yourself a "Obsessive Compulsive Audiophile" makes any difference this way or that.
As we saw,the effect is real,unless you want to invalid Lund's paper too.I believe he's a proper engineer specialized in the field and Genelecs are nice.
The effect might be real, and yes, Genelecs are nice (I enjoy the ones I have), but you still shouldn't trust subjective, uncontrolled perception.
 
Back
Top Bottom