• 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!

Science on low frequency quality?

jsilvela

Senior Member
Joined
Dec 1, 2022
Messages
440
Likes
401
Location
Spain
I've been consuming a lot of information on low frequency reproduction, room modes, correction via DSP, brand wars etc.

Something keeps bugging me. For all the effort expended on LF, I wonder if the science really backs it up.

I have the first edition of Floyd Toole's book. Reading that, and lots of material from ASR and elsewhere, I learn that flat FR measured in an anechoic chamber correlates with subjective preferences when listening tests are done double blind.
Is there research like the above specifically for low frequency? Do we have solid evidence on what correlates with subjective preference in the <100Hz range?
Are there studies like the Toole studies specifically about subwoofers? And how about subwoofer integration? That seems arcane too.

The assumption in Toole's book, the work of Welti, Geddes etc. is to try to achieve a flatter response in LF, and more even spatially.
But they open a big can of worms. Not only do you need to get >1 subwoofer, you need to place them carefully, do mic readings, generate individual filters for each sub, etc.
Not for the faint of heart.
If you're not willing to do that kind of homework, is >1 sub still recommended? Does it at least serve to even out the FR a bit and the spatial variation a bit? a lot?

Also, I see from the Psychoacoustic filter in REW that for low frequencies, human hearing seems to be a lot less sensitive.
In that case, is a flat FR still the goal? Can some deviations from flatness pass unnoticed?
How flat is flat enough?
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
In Toole's work small monitors with response to about 100 hz had an almost 100% correlation with their model of preferred sound. It was less with full range speakers. My guess is that is mainly because full range speakers in a real room are uneven at low frequencies. Even in their test circumstances. So at least it seems to me a really flat result is going to be better and preferred.

How important is that? I think they surmised from their research that bass response was responsible for 30% of preference. So an outsized importance despite our relative insensitivity to bass. When you think about it, nulls and resonances in a room can result in huge differences in response. Some instruments can nearly disappear in nulls and others can be heavily overblown masking other frequencies even with our reduced sensitivity at sub 100 hz frequencies. So something of a dichotomy that we are insensitive to bass yet in smaller domestic spaces the response is usually by far the most uneven in the whole frequency range.

So one fortunate circumstance is we stop hearing position around 80 hz and below. So one sub EQ'd is fine. You cannot EQ out nulls, but we are more sensitive to resonances and peaks. So just flattening peaks is very helpful. Optimally having more subs, fewer nulls etc is better. Or alternatively having a large enough room all these effects are pushed way down in frequency which is not usually the case in domestic sized listening rooms.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
Do we have solid evidence on what correlates with subjective preference in the <100Hz range?

If we put resonances and nulls aside it can vary wildly from person to person and on different material. Personally my preference seems to align with the science but with my partner he keeps saying that drums pound his head whenever I set it to what most people would consider 'Neutral' and if I start cutting these frequencies even he thinks the sound VERY quickly starts to sound flat and without dynamics.

The music industry has evolved to very strict tolerances where if you don't statistically belong to where the majority of people belong it's rather hard to have a good experience. Instruments are tuned a certain way at the end of the day.
 
OP
jsilvela

jsilvela

Senior Member
Joined
Dec 1, 2022
Messages
440
Likes
401
Location
Spain
If we put resonances and nulls aside it can vary wildly from person to person and on different material. Personally my preference seems to align with the science but with my partner he keeps saying that drums pound his head whenever I set it to what most people would consider 'Neutral' and if I start cutting these frequencies even he thinks the sound VERY quickly starts to sound flat and without dynamics.

The music industry has evolved to very strict tolerances where if you don't statistically belong to where the majority of people belong it's rather hard to have a good experience. Instruments are tuned a certain way at the end of the day.
hm, yes, this is something I keep seeing. A lot less agreement on LF, a lot more room for subjectivity.
That, in my mind, makes a lot of the advice here to go for multi-sub, apply MSO etc. kind of fragile.
I.e. you might expend effort and money to "go with the science" and then find you don't like it.
The same is always true of course, but at ASR you see routine advice for people to go for Genelec / Revel / Kef etc, without much caveating.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
hm, yes, this is something I keep seeing. A lot less agreement on LF, a lot more room for subjectivity.
That, in my mind, makes a lot of the advice here to go for multi-sub, apply MSO etc. kind of fragile.
I.e. you might expend effort and money to "go with the science" and then find you don't like it.
The same is always true of course, but at ASR you see routine advice for people to go for Genelec / Revel / Kef etc, without much caveating.

I don't like you using my words to justify buying speakers with resonances and/or lack output capability. The brands that you mentioned are popular here because they minimize resonances and are transparent about their output capability, not because of some ASR cult behavior.
 
OP
jsilvela

jsilvela

Senior Member
Joined
Dec 1, 2022
Messages
440
Likes
401
Location
Spain
I don't like you using my words to justify buying speakers with reasonances and/or lack output capability. The brands that you mentioned are popular here because they minimize resonances and are transparent about their output capability.
Oh, I was not trying to twist your words at all. Sorry.
EDIT: and certainly I don't want to "justify buying speakers with resonances"!
Let me try to clarify what I meant.
For mid-high frequencies, "the loudspeaker is in charge", and the science seems solid about flat on-axis, good directivity index, mild downward slope in-room.
Those brands get recommended at ASR because they go with the science. The belong to the Toole school, if you will.

For low frequency, I see a lot less clarity. For <100Hz, the room intervenes.
You often get get told to get at least 2 subs, and use MSO / REW. Perhaps Dirac, Audyssey XT32.
But it all seems to require spending on >1 sub, possibly an AVR for bass management. Possibly a newer AVR because the model you have has an outdated version of MultiEQ/ whatever.
And tweaking quite a bit.

That's why I made the distinction.
 
Last edited:

ZolaIII

Major Contributor
Joined
Jul 28, 2019
Messages
4,069
Likes
2,409
_20230117_120100.JPG
So it's pretty much how you hear it (not imagine, like or whatever). We still don't know (scientific) why it varies so much nor there is a consensus, educated guess is state of our individual hearing (which still needs to be empirically proven on sufficient number of samples).
 

sigbergaudio

Major Contributor
Audio Company
Forum Donor
Joined
Aug 21, 2020
Messages
2,639
Likes
5,397
Location
Norway
First of all, in discussions like this it helps to be precise:

flat is generally used to indicate a neutral slope.
smooth is generally used to indicate few/small deviations from the intended target curve (which may or may not be flat).

With that taken care of:

I don't think there's much dispute about the fact that a smooth frequency response is desirable in the entire frequency range. Below 100hz the response is very often NOT smooth due to room interactions, and DSP and / or multisubs are effective ways of dealing with that. A single sub with DSP is better than no DSP. Two subs with no DSP is better than one sub with no DSP. I think anyone who listen to a system with a well behaved bassrange (smooth response in 20-200hz area) will greatly prefer this to the alternative.

With regards to the slope, we don't have tons of science, but we have a endless amounts of in-room measurements from various forums, we have the Harman study on preferred in-room curve, and we also have a Genelec study who checked the in-room response in several hundred studios.

Summing up all of that data, we can reasonably conclude most people prefer a slope where the bass area (20-100hz) are somewhere in the region of 5-10dB higher than the midrange/treble area (2-10khz). Some outliers with both darker and brighter slopes can be found, but the majority falls within that range. The "right" slope depends both on personal preference and acoustics in the room. We also see some slopes that are gradually falling throughout the entire frequency band, and some curves where the lift doesn't really start until 100-200hz. I suspect a lot of the latter versions is due to limitations in the setup - people don't really have full DSP capabilities, and basically have to create the slope by increasing the gain of their subs.
 

bodhi

Addicted to Fun and Learning
Forum Donor
Joined
Nov 11, 2022
Messages
973
Likes
1,371
hm, yes, this is something I keep seeing. A lot less agreement on LF, a lot more room for subjectivity.
That, in my mind, makes a lot of the advice here to go for multi-sub, apply MSO etc. kind of fragile.
I.e. you might expend effort and money to "go with the science" and then find you don't like it.
The same is always true of course, but at ASR you see routine advice for people to go for Genelec / Revel / Kef etc, without much caveating.

You seem to be beating around the bush here. Maybe you are asking if smooth bass response all the way to infrasonics is always preferable or if uneven bass response with lacking extension caused by inferior equipment and/or environment could actually be "as good" or even better?

Answer to that question is "no". I don't think there is much ambiguity about this.
 
OP
jsilvela

jsilvela

Senior Member
Joined
Dec 1, 2022
Messages
440
Likes
401
Location
Spain
You seem to be beating around the bush here. Maybe you are asking if smooth bass response all the way to infrasonics is always preferable or if uneven bass response with lacking extension caused by inferior equipment and/or environment could actually be "as good" or even better?

Answer to that question is "no". I don't think there is much ambiguity about this.
Not exactly.
The issue I'm going for in this thread is more of:
The "average" (loaded term I know) advice at ASR is to get multi-sub, use MSO, apply room EQ, have drivers that go to 20Hz etc.
The "beginner package" for low frequency is steep. Does it really make such a difference? Is it borne out by the science? Blind tests?
 

sigbergaudio

Major Contributor
Audio Company
Forum Donor
Joined
Aug 21, 2020
Messages
2,639
Likes
5,397
Location
Norway
Not exactly.
The issue I'm going for in this thread is more of:
The "average" (loaded term I know) advice at ASR is to get multi-sub, use MSO, apply room EQ, have drivers that go to 20Hz etc.
The "beginner package" for low frequency is steep. Does it really make such a difference? Is it borne out by the science? Blind tests?

You (or others) are given this advice when asking about what?

And make a difference compared to what?
 

bodhi

Addicted to Fun and Learning
Forum Donor
Joined
Nov 11, 2022
Messages
973
Likes
1,371
Not exactly.
The issue I'm going for in this thread is more of:
The "average" (loaded term I know) advice at ASR is to get multi-sub, use MSO, apply room EQ, have drivers that go to 20Hz etc.
The "beginner package" for low frequency is steep. Does it really make such a difference? Is it borne out by the science? Blind tests?

Get two passable subs like SVS SB-1000 and UMIK-1. Then look at MSO tutorial video on Youtube and replicate the steps exactly, no tweaking.

That is something like 1k€/$ and let's say one Saturday afternoon of work. After this you are 80% there and can quickly A/B the results.
 
OP
jsilvela

jsilvela

Senior Member
Joined
Dec 1, 2022
Messages
440
Likes
401
Location
Spain
You (or others) are given this advice when asking about what?

And make a difference compared to what?
ah, good clarifications.

when asking about what?
It happens quite often even with beginners. You see recommendations for, say, some affordable bookshelf speakers with a good rating by Amir.
And are told a sub would be recommendable with them to get LF. Then someone points out that in your small living room, you'll surely need room EQ, and they consider 2 subs to be bare minimum, and MSO would help. Not going to say I see this every time, but I do see it often.

Compared to what?
Exactly. Well, say I bought a sub, placed it carefully and applied room correction. Would getting a second/third sub + MSO be a night and day difference? Subtle improvement?

I've seen the argument that if you only care about one listening position, then 1 sub + EQ is good enough.
But I think many of us listen to music + watch movies/tv in the same space. In a sofa, say, sometimes lying down with a laptop, sometimes sitting on the floor or another chair etc.
So, "only one listening position" sounds a bit restrictive to me.
Anyhow, that was the gist. How much of a relative improvement is it over 1 sub + DSP?
 

FrantzM

Major Contributor
Forum Donor
Joined
Mar 12, 2016
Messages
4,337
Likes
7,734
Hi

You seem to be beating around the bush here. Maybe you are asking if smooth bass response all the way to infrasonics is always preferable or if uneven bass response with lacking extension caused by inferior equipment and/or environment could actually be "as good" or even better?

Answer to that question is "no". I don't think there is much ambiguity about this.
Not exactly.
The issue I'm going for in this thread is more of:
The "average" (loaded term I know) advice at ASR is to get multi-sub, use MSO, apply room EQ, have drivers that go to 20Hz etc.
The "beginner package" for low frequency is steep. Does it really make such a difference? Is it borne out by the science? Blind tests?
I am with @bodhi here.
Don't waver .. The body of work is solid on low frequencies. Please re-read @siegberd reply too.
Geddes advocates 3 subwoofers, you can use more but the benefits are slim in his opinion. Geddes multisub are placed in "quasi-random" positions. One in the front corner (usually the most potent and lowest reach subwoofer), the others .. somewhere on the opposite wall, between the listening postion and the plane of the mains speakers, the last one behind the listening position or on another wall , slightly elevated if possible... much smaller subwoofer can be used for that purpose...

Welti-Devantier-Toole-/ Harman preconize 4 subwoofers at the middle of the walls of a rectangular room.
'
In most cases you need a DSP component for multi-subwoofers. The consensus is the miniDSP 2x4 HD. It cost around $260.00 plus applicable taxes on Amazon.
Now for the subwoofers, t depends on the room. I currently use 2 very inexpensive ($250 each) subwoofers in a 3.3 x 5 x 2.8 meters W x L x H concrete room. I am very pleased. I reach 19 Hz -3 dB. I subscribe to the Harman curve BTW. You, then present this multi-subwoofer as one subwoofer to the AVR for integration in a MCH system

Two subwoofers properly configured already take you very far. 3 or more is ideal. There is a lot of FUD about subwoofers. We'll debate some more but the best tool to choose a subwoofer for your room is @sweetchaos Subwoofer Comparison Tool. Essential tool BTW.

.....
Get two passable subs like SVS SB-1000 and UMIK-1. Then look at MSO tutorial video on Youtube and replicate the steps exactly, no tweaking.

That is something like 1k€/$ and let's say one Saturday afternoon of work. After this you are 80% there and can quickly A/B the results.
Was about to write a long reply...
Make it $1250 because of the miniDSP 2x4
Slam dunk! @bodhi ... Although I wouldn't call the SB-1000 "passable" mine, the Dayton Audio SUB-1500, however, fit the qualifier :)

Peace.
 

sigbergaudio

Major Contributor
Audio Company
Forum Donor
Joined
Aug 21, 2020
Messages
2,639
Likes
5,397
Location
Norway
ah, good clarifications.


It happens quite often even with beginners. You see recommendations for, say, some affordable bookshelf speakers with a good rating by Amir.
And are told a sub would be recommendable with them to get LF. Then someone points out that in your small living room, you'll surely need room EQ, and they consider 2 subs to be bare minimum, and MSO would help. Not going to say I see this every time, but I do see it often.


Exactly. Well, say I bought a sub, placed it carefully and applied room correction. Would getting a second/third sub + MSO be a night and day difference? Subtle improvement?

I've seen the argument that if you only care about one listening position, then 1 sub + EQ is good enough.
But I think many of us listen to music + watch movies/tv in the same space. In a sofa, say, sometimes lying down with a laptop, sometimes sitting on the floor or another chair etc.
So, "only one listening position" sounds a bit restrictive to me.
Anyhow, that was the gist. How much of a relative improvement is it over 1 sub + DSP?

If you

A) are mainly concerned with one main listening position (as opposed to several)
B) have some options available for where to locate your subwoofer (so you can put it where it gives you the smoothest raw response)
C) Have automatic room correction with a decent resolution in the bass (so not the very cheapest of AVRs)

my opinion is that

1) You are likely to get a more than decent result even with just one subwoofer + DSP.
2) No, you do not need MSO.


What would happen if you added more subs?
1) With just one sub, while getting a decent result, you're likely to have at least one major dip, potentially a wide one. You may notice this on some tracks if some significant part of the music content happens to hit this dip. For instance some parts of a bass line may be less powerful than the rest, or the bass drum lacks power because it's tuned to this frequency. Some people notice this more than others. Adding a second sub in the right place is likely to fix this, so you have a good chance of getting an almost perfect response across 20-100hz in a limited sweet spot.
2) You are likely to get a smoother response over a wider listening area. This may not be important to you.
3) You will obviously get more capacity / lower distortion, which may be of importance if you play music and especially movies at loud levels.


Real life example:
In my living room I just remodeled, so I currently (and temporarily) only have one subwoofer + DSP. 15-100hz in that room in green below. I also have a dedicated listening room with dual subwoofers + DSP. 15-100hz in red below. Both rooms sound just fine, but as you can see the room with the single sub has a pretty wide dip between 50-60hz. I doubt most "normal" people would notice or care though.
1673962416154.png

1673962483402.png
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
The reason for multiple subs is to smooth the in-room response. You can have peaks typically in the 3~6 dB range where sound waves add, and nulls whilst theoretically infinitely deep can hit -20~-30 dB in a typical room based on what I have measured. The peaks you can usually trim down with EQ, but nulls are too much for EQ and, since they occur because of signals cancelling each other, EQ doesn't really help. If you have a null because direct (D) and reflected (R) sound cancels you get D-R = 0. Apply EQ, raising the amplitude by A, and you raise both D and R so AD - AR = 0 -- no change except you waste a lot of power. You must either move the listening position out of the null, or use subwoofers to counter the null. The latter is the focus of many papers by Geddes, Toole, Welti, and others.

There is a short article on room modes here: https://www.audiosciencereview.com/forum/index.php?threads/room-modes.25433/

One sub can help, if properly placed, for one listening position (and perhaps a limited number of others based upon the room's dimensions and seating positions). You can look at any number of online room mode calculators to estimate the frequencies for a given (typically rectangular) room size and seating position. Additional subs allow you to compensate more modes and/or cover a wider listening area.

My room has "bad" dimensions, nearly exact multiples of length and width, leading to doubled-up nulls. Like many folk, moving the listening position was not a viable option due to physical constraints of the room (door, windows, etc.) I have four small'ish (12") subs strategically placed and aligned to provide nearly flat response across my listening couch. One would not do it, two helped some, but four (I did not try three) were required to get reasonably smooth response. My room is not large and I do not need large subs; four 12" (Rythmik) subs do just fine. I am not using a miniDSP (or MSO) but tweaked things myself (using room measurements, e.g. REW) in my AVP to get decent bass response.

FWIWFM - Don
 
Last edited:

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Regarding science about perceptions of bass and "quality" (which I interpret as subjective impressions, could be potentially described along specific parameters like https://www.aes.org/e-lib/browse.cfm?elib=16323), there are some interesting papers and models that have been discussed here but not very often.

1. Fazenda et al. Subjective preferences of modal control methods JAES 2012: https://www.aes.org/e-lib/browse.cfm?elib=15734 (can download free copy at https://www.avsforum.com/attachments/jaes_v60_5_perception_modal_control-pdf.2273992/)
2. Martens et al, SPATIAL AUDITORY DISPLAY USING MULTIPLE SUBWOOFERS IN TWO DIFFERENT REVERBERANT REPRODUCTION ENVIRONMENTS, ICAD 2005, and the related SUBJECTIVE EVALUATION OF AUDITORY SPATIAL IMAGERY ASSOCIATED WITH DECORRELATED SUBWOOFER SIGNALS ICAD 2002
3. Griesinger (http://www.davidgriesinger.com/), specifically
but can find other papers on his linked page

I interpret these loosely to suggest that:
A. Although bass reproduction is generally considered as a minimum phase phenomenon, perceptual testing suggest possible benefit of prioritizing modal decay time reduction over simple frequency response approaches (#1 above, also https://www.semanticscholar.org/pap...ivet/c60956bb7ae27df063100c5431f738c7225667ef)
B. Laterally decorrelated bass may positively contribute to spatial perceptions (envelopment or spaciousness) in certain situations with certain listening materials, perhaps with certain recording techniques (this has been discussed a fair amount in the past, but my speculation is that perhaps many stereo classical music recordings might benefit more than, say, pan pot stereo recordings). Since many listening rooms are not perfectly symmetric on the left and right in terms of construction, including window and door placement, reality does not perfectly reflect the mathematical model for modes, so I wonder to what degree sidewall placement of subs might excite room modes a little differently. The other aspect I wonder is whether placing the subwoofers wider than the left and right speakers might act like Blumlein shuffling in the low bass (can see discussed https://www.flyhighwaves.com/studio/stereo-shuffling-new-approach-old-technique/ and https://www.audiosignal.co.uk/Resources/Stereo_shuffling_A4.pdf, but again, spatial perceptions are referenced here)

Two aspect of Geddes' approach (which has been described in different ways at different times but generally involves reasonably full-range speakers run full-range, one ultra-low frequency subwoofer located in a corner, one bass source above the midline of the room, and at least one other subwoofer due to point of diminishing returns) that potentially differ significantly from the Welti approach (which has also evolved somewhat from the typically linked "multisub" paper to https://hometheaterhifi.com/technic...n-interview-with-todd-welti-and-kevin-voecks/):
i. The degree of overlap between the subs and the main channels
ii. My speculation that his approach of multiple decorrelated bass sources may end up incorporating some aspects of A and B above, even if unintentionally so, since highly asymmetric bass source placement may be expected to differentially "drive" room modes, also phase and level of each additional bass source are sequentially adjusted to reduce peaks, which may involve phase cancellation at modal frequencies, essentially acting somewhat like a "sink" or electroacoustic absorber, as well as possibly contributing to what I discussed above in point B. Perhaps @Duke could comment further.

Anyway, since you asked about science and bass quality, I thought the above references might serve as a possible starting point to explore further on your own.

Young-Ho
 

FrantzM

Major Contributor
Forum Donor
Joined
Mar 12, 2016
Messages
4,337
Likes
7,734
The reason for multiple subs is to smooth the in-room response. You can have peaks typically in the 3~6 dB range where sound waves add, and nulls whilst theoretically infinitely deep can hit -20~-30 dB in a typical room based on what I have measured. The peaks you can usually trim down with EQ, but nulls are too much for EQ and, since they =occur because of signals cancelling each other, EQ doesn't really help. If you have a null because direct (D) and reflected (R) sound cancels you get D-R = 0. Apply EQ, raising the amplitude by A, and you raise both D and R so AD - AR = 0 -- no change except you waste a lot of power. You must either move the listening position out of the null, or use subwoofers to counter the null. The latter is the focus of many papers by Geddes, Toole, Welti, and others.

There is a short article on room modes here: https://www.audiosciencereview.com/forum/index.php?threads/room-modes.25433/

One sub can help, if properly placed, for one listening position (and perhaps a limited number of others based upon the room's dimensions and seating positions). You can look at any number of online room mode calculators to estimate the frequencies for a given (typically rectangular) room size and seating position. Additional subs allow you to compensate more modes and/or cover a wider listening area.

My room has "bad" dimensions, nearly exact multiples of length and width, leading to doubled-up nulls. Like many folk, moving the listening position was not a viable option due to physical constraints of the room (door, windows, etc.) I have four small'ish (12") subs strategically placed and aligned to provide nearly flat response across my listening couch. One would not do it, two helped some, but four (I did not try three) were required to get reasonably smooth response. My room is not large and I do not need large subs; four 12" (Rythmik) subs do just fine. I am not using a miniDSP (or MSO) but tweaked things myself (using room measurements, e.g. REW) in my AVP to get decent bass response.

FWIWFM - Don
@DonH56

I knew you were a gentleman, and admit that you didn't rub more salt on the wound.
But... , you failed to mention that you hit 8 Hz, (!!!) in your room and that you have the more than potent, JBL equivalent of the Trinnov Altitude Processor... so ,
Duly noted ...
you are still on my Hate List.

:p
:D

As usual , great info.
Peace.
 
Top Bottom