and it totally contredict what all acoustician recommend to NOT do: to try to solve time domain problem with EQ. ITs recommended to not use EQ over 100hz. so 100hz to 300hz, where bass clarity and fullness is.
However, treatment is useless under 50-60-70hz unless building pressure based bass traps and this is where subwoofer can help.
You're not getting it because it isn't exactly intuitive. Buy reducing the peak via eq DOES reduce the decay time. It DOES address ringing.
Don't cling to the assumption that frequency adjustments can't alter time behavior - ringing, decay time, whatever you want to call it. It CAN and does at low frequencies in domestic sized rooms because of the physics of how long waves in small boxes behave.
can you provide more info about this and perhaps measurements that shows the before and after EQ????
RHosh: what you are saying here as been totally contredict at gearslutz. id like you to read this:
"
But you CANNOT EQ the room response above the critical frequency fc of a non-minimum phase system. (I can see a few freak already)
In other words, you are on the wrong path - one that was once considered almost routine until the big quantum leap in acoustics 'hit', where sound began to be evaluated with respect to time.
If folks want to wear out an old topic that was 'hot' circa 1989, fine, but you cannot EQ non-minimum phase environments. (Sorry folks, but unless you like to do pole-zero analysis - assuming that is not another technoid term for you - you are going to have to take some of this on faith. For others who like that sort of stuff, i have some papers that address aspects of this if you will PM me; but you are forewarned... ;-)) )
This is a topic that has been debated ad nauseaum - and it is also the reason you no longer see literally racks of ATA cases holding 30-60 channels of EQ in concert SR rigs anymore that used to be standard!)
To use EQ successfully requires a special case that assumes a room where the problems are NOT caused by the superposition (combination) of signals that are out of ABSOLUTE phase. This CANNOT be resolved in the frequency domain.
The best you can do, if ALL has been done that can be done to the room, is to evaluate the phase response of the room transfer function relative to the frequency response. I know of 3 programs that do this, and you can rent someone to do it for you, but you cannot rent a 'tool'. And this is the absolute last step in tuning a room AFTER all of the acoustic treatment is successfully optimized. And you still may not be able to do it. It is NOT a shortcut to make up for a less than optimal room!!!
If the phase is 'flat' in the region of a frequency anomaly you wish to correct, you can apply an inverse parametric correction - but ONLY in the minimum phase region! If you try to do so in areas of non-minimum phase, you will change the phase along with the magnitude in a non-1 to 1 manner - and create additional problems elsewhere - in other words, you simply move the problems around and perhaps create additional one.
If their are still issues with speaker room interaction, then the best you can do is to EQ the direct signal emanating from the speaker ONLY. But This will not resolve an speaker-room issues.
Sorry for what seems to be techno-babble, but it is correct. And there is more info if one wants to delve into this. But the easiest thing if you are not prepared for the math - and lots of it - is to simply accept it. As I mentioned, this has been hotly debated and researched...but its now an 'old' topic in terms of the the debate.
This is a consequence of the new time based models - and an old once-common approach that has died a hard death - and is still too pervasive.
---------------------------------
Oh, and to anticipate those who will jump in and claim to have fixed a room mode with PEQ. Great. But in a small acoustical space (as defined by Schroeder) the frequency where the wavelengths of sound are longer then the room dimensions - or conversely, the freq where the wavelengths are smaller than the room dimensions - called the critical frequency, fc.
And the critical frequency is a point - a dividing line if you will, where the models of sound behavior change. Wavelengths of sound longer then the room dimensions where modal behavior typical of the sound as a pressure wave model ceases, and where sound as a particle model begins - where the sound begins to act as a billiard ball and assumes the behavior of discrete specular reflections.
In a soundfield with many reflections that combine with a direct source, you generally speaking, have a non-minimum phase environment. Whereas the pressure zone of the modal response can be argued to be minimum phase. There is a limit to this validity, but it allows limited success in the modal region. But that is NOT an optimal solution as the notion that it is truly minimum phase is not exactly valid either...
And now I suspect that everyone is confused - even if you weren't before.
What can I say - there are some major paradigm shifts from the old models of acoustics. Scary and apple cart upsetting for some, and exciting for others - it all depends on your perspective...
Hey, don't blame me , you can start by blaming Dick Heyser. ;-))))
Or you can start looking at events from the time perspective instead of the derivative frequency perspective! THAT is where you solve the preponderance of issues that manifest themselves in the frequency domain.
A non-minimum phase response is produced by multiple reflections from room boundaries superimposing themselves upon the direct signal from the loudspeaker. In such cases, every different listening position will also receive a different balance of direct and reflected signals from each boundary. And in such a situation, no casual filter can correct the complex spatially dependent response anomalies by equalizing the direct signal speaker response.
I am intentionally NOT trying to address the concept of minimum phase mathematically! It has been rather adequately expressed that most do not desire this information, nor does it help them to understand the phenomena. If one wants to pursue that avenue, there are many more optimal avenues. And a ‘logical’ description attempting to describe a necessarily complex relationship of multiple factors is not the way to do it. But for the purposes of the forum, it must suffice. So debating the use of a single term used loosely within an albeit simplified logical description in lieu of a comprehensive mathematical definition rather misses the point. But you are of course welcome to make it, and I hope it contributes to the practical application of the all-to-common but unproductive attempts to EQ speaker-room interaction in the realm dominated by summed specular reflections.
We are dealing with a phenomena that by definition is mathematical, relating to how amplitude and phase track to each other. It has no relevance to the quantity of phase change. Therefore, a minimum phase system is essentially one where every change in the amplitude response has a corresponding change in the phase response and visa versa. When the restoration of ‘flatness’ of either domain does not restore the other to flatness, the systems is essentially non-minimum phase, and this cannot be corrected by a causal filter in the form of an EQ.
And this does not even begin to address the concept of ‘excess’ phase deviation caused by the summation of many time shifted signals as occurs with the sum of many crossover filters.
The reference to flatness was made with respect to the sum (superposition) of the phase of multiple signals. Thus indicating that the destructive combination of various sources relative to their absolute phase was not at play – and thus presented a minimum phase condition whereby the system may effectively be EQ’d – IF, and Only if, such a situation is present!
And as such a situation is indeed rare, and cannot be determined without adequate measurement and display of multiple interactive relationships, for all practical purposes for most, one cannot EQ the speaker room response. One can only EQ the direct signal emanating from the speaker - and this will not correct superposed speaker-room anomalies.
Edit: Thanks Frank!
Yes, to reiterate. You do not correct issues that manifest themselves as frequency response issues in the frequency domain. You deal with them in the TIME domain. And this is accomplished via signal synchronization - to the degree possible.
And,to extrapolate this out a bit for the more practical applications useful in a studio, I will link again to an article that all should keep handy - or simply remember with respect to speaker position in a room (bounded space); noting that the anomalies 'come with' the choice. And it is also a reason that horn loaded 'corner placement and CORRECTLY mounted and isolated 'soffit/wall' mounting is a preferred technique.
Note: This also accounts for SBIR as well!
How Boundaries Affect Loudspeakers
And one more note: Please notice how the placement causes a necessary deviation from the free-field anechoic response you so often see in the marketing brochure! Buyer beware! And please consider your speaker placement carefully. It is easier to avoid problems than to correct for them - as much fun as many of us have doing exactly that! ;-)
https://www.gearslutz.com/board/stud...ding-acoustics/453854-eq-monitors.html
The fundamental issue that few are addressing is that changing the setting on an EQ unit does not necessarily result in a 1:1 change in the corresponding phase of the signal.
A precise definition of minimum phase is a detailed mathematical concept involving positive real transfer functions, i.e., transfer functions with all zeros restricted to the left half s-plane.
Likewise, the relationship can be determined empirically by examining the relationship between the amplitude and phase of the signal by examining the Hilbert transform between the impulse and doublet response and examining the relationship for regions of a 1:1 relationship which indicates a minimum phase relationship.
To ignore this fundamental relationship is to twiddle knobs thinking one is making a desired difference when instead they are destructively modifying relationships elsewhere.
I know few are interested in understanding this aspect, as after all, I am constantly reminded that folks simply want answers and results. So with that, I'll bow out of this topic.
For what it’s worth, for those who desire the ability to evaluate such a response (without bothering to understand the underlying concepts), there are a few analysis platforms which provide such analysis, and can even generate precise parametric equalization settings for the minimum phase regions. But one wonders as to the value of discussing such aspects as it is a bit more involved than the ETC response, and such a discussion is predicated upon understanding some of the component relationships thereof, especially as we all know how questionable some here feel about the discussion of such an 'advanced' and 'esoteric' function…
https://www.gearslutz.com/board/studio-building-acoustics/521064-main-eq-curve-questions.html
Last edited:
From post 54: http://www.diyaudio.com/forums/subwoofers/284221-multiple-subs-vs-bass-traps-6.html#post4565335
There are a couple of things going on here. Let's look at the procedure used.
A measurement at the listening position is made in Acourate, saved, exported, and imported into REW. The Q/gain/frequency of the PEQ filter is played with until it hits the target properly. You can use the waterfall view in the EQ window in REW and the up/down arrows in the filter window to fine tune. See REW manual on how to do that.
Then in REW the equivalent IIR filter is generated, exported, and imported back into Acourate and applied as a prefilter. Meaning, the IIR filter is applied before Acourate processes the measurement and generates the FIR correction filter, with the IIR filter applied, along with the FDW parameters for amplitude and excess phase correction. By default, Acourate does not boost frequencies and has a proprietary psychoacoustic filter applied.
In other words, two correction filters have been applied to the measurement to achieve the result where the frequency response is flat, yet significant ringing has been removed. Or at least below one's auditory masking threshold: https://en.wikipedia.org/wiki/Auditory_masking
Auditory masking is another important topic to look at in world of psychoacoustics https://en.wikipedia.org/wiki/Psychoacoustics as people sometimes spend quite a bit of time focusing on acoustic issues that are masked by the main signal, i.e. music and could spend better time fixing the main audio signal arriving at the ears and less time on aspects that are not so audible. The precedence or Haas effect is another important concept as well: https://en.wikipedia.org/wiki/Precedence_effect
As can be seen by the measurements in post 54, this technique works. Sereval others on various DRC forums are achieving the same results.
Did you read jj's presentation? This is the first step into better understanding what our ears hear in small room acoustics.
About precedence effect:
"
So one wonders how to mitigate LF “early reflections”, not in all small acoustical spaces, but specifically in “LEDE rooms”…
That would be a legitimate concern were room modes somehow different in an LEDE topology as compared with any other small acoustical space topology. So it seems that one is reaching a bit in the attempt to find fault. And this reach is certainly exceeded if one attempts to attribute such issues to a particular acoustical response model.
Of course, the primary issue is the behavior of waves and how they manifest themselves spatially and temporally. And ‘within’ this behavior we must stop and remember what information is of primary use in localization.
AS Ethan has pointed out on more than a few occasions, both modal reinforcement and specular polar lobing is a result of constructive and or destructive interference respectively.
LF modal standing waves do not form room reinforced polar lobing anomalies per se (although they can certainly also display this behavior in the near field due to the interaction of spaced sources at wavelengths greater than ¼ wavelength spacing), as they instead form room boundary reinforced resonant standing waves – which we commonly refer to as modes.
Specular reflections, consisting of energy whose wavelengths are smaller than the incident boundary dimensions reflect and behave specularly. And by virtue of superposition physically manifest themselves as frequency dependent polar lobing. It is this spatial polar lobing that manifests itself as the characteristic pattern called comb filtering in the measured frequency response. (If not already familiar with this, one might want to note that the polar lobing is what is physically manifest. Comb filtering is not an actual physical phenomena- it is simply a description of a periodic pattern exhibited within a measured frequency response. In other words, polar lobing is ‘real’; comb filtering is not.)
This characteristic mix of both modal and specular energy reflection is characteristic behavior of the entire subset within acoustics defined by/as “small acoustical spaces”. It is interesting that this issue is now posited not as an issue of concern to the entire class of bounded small acoustic spaces, but rather as an issue pertaining in particular to the LEDE acoustical response.
If one were aware of this fact, there would be no need to focus on a particular small room response with respect to the issue of room modes when the behavior and treatment with some combination of porous and resonant traps is pertinent for all small acoustical spaces.
The LEDE concept includes resonant standing wave mitigation as does any other acoustical room model of which I am aware, thus damping the delayed resonant mode energy and allowing the direct signal to remain predominant.
As for the second question, the simplest answer is that one should know that our ear-brain is not dependent upon LF information 200 Hz and below for localization cues.
But as some may not be aware of this, let's jump over into the area of psycho acoustics for a moment to get a glimpse of (part of) what is happening.
Research has shown the important function of the pinna – the outer structure of the ear (see Carolyn ‘Puddie’ Rodgers’ seminal research in particular) acting as an important directional filter, which along with reflections off the torso and shoulders, have an important contributory effect on perceptual comb filtering with regard to ones perception of direct signals as well as their localization.
The pinna imposes spectral modifications (comb filtering) dependent on the angle of the received incident energy and with the perceived time/phase and intensity differences between the arriving signals in each ear – as are measured via interaural cross correlation measurements – forms an important portion of the localization and imaging process.
Thus the physical structure and orientation of the outer ear helps to filter the differences in time and level between the signals as they are received in each ear – what we call inter-aural cross-correlation (IACC).
{A slight detour through yet another looking glass into psycho-acoustics land...
An interaural cross correlation measurement tells us, frequency band by frequency band, how similar two signals are.
This is commonly done using a dummy head (OK, just for the record, I am trying my best to be nice…) or more practically, by placing a simulated human head with microphones in its ‘ears’, or a SASS mic, or best, Meade Killion designed ‘In-The-Ear’ (ITE) mics by Etymotic) in a space and connect the outputs of the microphones to a platform such as ARTA or EASERA that measures cross-correlation, then we are able to make and analyze IACC measurements.
And to deviate for just a moment longer with more information than I suspect anyone cares about… This filtering action of the pinna can be characterized as a transfer function, with one part of this transfer function describing the angle dependent sound attenuation which is referred to as the Interaural Intensity Difference (IID), and when measured in dB is referred to as the Interaural Level Difference (ILD). Another part of the transfer function describes the angle dependent time difference between the arrival of a sound at the left and right ears due to their separation (which typically ~ 8 inches in adults). This is called the Interaural Time Difference (ITD) (not to be confused with Beranek’s Initial Time Delay), or when measured as the phase shift of a sinusoidal tone, the Interaural Phase Difference (IPD).
Thus, in theory, if the sound source is located directly in front of the receiving ‘head’, and the room is exactly symmetrical, and the head is on the line of symmetry, then the sound at the two ears will be identical. Therefore, the IACC will have a high value at all frequencies. In reality, the signals will never be identical, therefore you will have a varying correlation value at different frequencies.
And yes, I know we are ignoring the inner ear contributions, but we will quit before anyone’s head explodes….}
OK, we’re back from the side trip into psycho-acoustics land… (and you had thought acoustics was weird... ;-)
So what does all of this have to do with LF localization – or that lack of it?
At very low frequencies, the wavelength of the sound in air is much much larger than the ~17 cm / 8 inch distance between the ears/microphones, and therefore the pressure at the two microphones will be almost identical.
The lower the frequency, the more similar the two signals and the higher the correlation. Thus, as the mind is dependent upon the differences in order to gauge localization, the lower the frequency, the less useful they are for localization purposes.
BTW, these characteristics can be specifically evaluated and incorporated into measurements in platforms such as ARTA and Easera and in measurements such as the polar(3D) ETC.
As far as the effective frequency range of human hearing, we often think in terms of the typically stated as 20 Hz to 20 kHz. But this is not as simple as it appears, as the functional aspects vary dramatically over that range. Decreasing in frequency below 1000 Hz, the hearing threshold increases dramatically from 0dB at 1000 Hz to required levels of ~90 dB at 20 Hz! One can frequently see the practical feature that attempts to address this issue in the form of a loudness contour control which boosts low frequency gain signals at low average volume levels.
Thus, without delving onto the timing and relative gain issues involved with localization here, for localization and imaging issues we are concerned primarily with a rather restricted mid-range band of energy as the ears’ greatest sensitivity is normally in the frequency range of from 1 kHz to 4 kHz. (If you want to lower that a bit to ~750 or 800 Hz, I won’t argue.)
But energy below 200 Hz is of little to no practical value for use in localization.
Thus it has little bearing on the Precedence effect. We, quite frankly, have bigger more substantial issues to address.
And the primary manifestation of LF direct and indirect energy is in the form of modal (standing wave) resonance.
A practical example for those who served in the military and/or have been exposed to helicopters, you know how difficult it is to identify the approach vector of an unseen slick for which you only have LF sonic clues (and assuming they are ours, works to our advantage!).
Instead we are more worried about its resonance, as it is perceived as boomy and tends to mask other sounds, which is why we need to control the modal behavior with LF traps.
"
https://www.gearslutz.com/board/6202822-post67.html
"
So one wonders how to mitigate LF “early reflections”, not in all small acoustical spaces, but specifically in “LEDE rooms”…
That would be a legitimate concern were room modes somehow different in an LEDE topology as compared with any other small acoustical space topology. So it seems that one is reaching a bit in the attempt to find fault. And this reach is certainly exceeded if one attempts to attribute such issues to a particular acoustical response model.
Of course, the primary issue is the behavior of waves and how they manifest themselves spatially and temporally. And ‘within’ this behavior we must stop and remember what information is of primary use in localization.
AS Ethan has pointed out on more than a few occasions, both modal reinforcement and specular polar lobing is a result of constructive and or destructive interference respectively.
LF modal standing waves do not form room reinforced polar lobing anomalies per se (although they can certainly also display this behavior in the near field due to the interaction of spaced sources at wavelengths greater than ¼ wavelength spacing), as they instead form room boundary reinforced resonant standing waves – which we commonly refer to as modes.
Specular reflections, consisting of energy whose wavelengths are smaller than the incident boundary dimensions reflect and behave specularly. And by virtue of superposition physically manifest themselves as frequency dependent polar lobing. It is this spatial polar lobing that manifests itself as the characteristic pattern called comb filtering in the measured frequency response. (If not already familiar with this, one might want to note that the polar lobing is what is physically manifest. Comb filtering is not an actual physical phenomena- it is simply a description of a periodic pattern exhibited within a measured frequency response. In other words, polar lobing is ‘real’; comb filtering is not.)
This characteristic mix of both modal and specular energy reflection is characteristic behavior of the entire subset within acoustics defined by/as “small acoustical spaces”. It is interesting that this issue is now posited not as an issue of concern to the entire class of bounded small acoustic spaces, but rather as an issue pertaining in particular to the LEDE acoustical response.
If one were aware of this fact, there would be no need to focus on a particular small room response with respect to the issue of room modes when the behavior and treatment with some combination of porous and resonant traps is pertinent for all small acoustical spaces.
The LEDE concept includes resonant standing wave mitigation as does any other acoustical room model of which I am aware, thus damping the delayed resonant mode energy and allowing the direct signal to remain predominant.
As for the second question, the simplest answer is that one should know that our ear-brain is not dependent upon LF information 200 Hz and below for localization cues.
But as some may not be aware of this, let's jump over into the area of psycho acoustics for a moment to get a glimpse of (part of) what is happening.
Research has shown the important function of the pinna – the outer structure of the ear (see Carolyn ‘Puddie’ Rodgers’ seminal research in particular) acting as an important directional filter, which along with reflections off the torso and shoulders, have an important contributory effect on perceptual comb filtering with regard to ones perception of direct signals as well as their localization.
The pinna imposes spectral modifications (comb filtering) dependent on the angle of the received incident energy and with the perceived time/phase and intensity differences between the arriving signals in each ear – as are measured via interaural cross correlation measurements – forms an important portion of the localization and imaging process.
Thus the physical structure and orientation of the outer ear helps to filter the differences in time and level between the signals as they are received in each ear – what we call inter-aural cross-correlation (IACC).
{A slight detour through yet another looking glass into psycho-acoustics land...
An interaural cross correlation measurement tells us, frequency band by frequency band, how similar two signals are.
This is commonly done using a dummy head (OK, just for the record, I am trying my best to be nice…) or more practically, by placing a simulated human head with microphones in its ‘ears’, or a SASS mic, or best, Meade Killion designed ‘In-The-Ear’ (ITE) mics by Etymotic) in a space and connect the outputs of the microphones to a platform such as ARTA or EASERA that measures cross-correlation, then we are able to make and analyze IACC measurements.
And to deviate for just a moment longer with more information than I suspect anyone cares about… This filtering action of the pinna can be characterized as a transfer function, with one part of this transfer function describing the angle dependent sound attenuation which is referred to as the Interaural Intensity Difference (IID), and when measured in dB is referred to as the Interaural Level Difference (ILD). Another part of the transfer function describes the angle dependent time difference between the arrival of a sound at the left and right ears due to their separation (which typically ~ 8 inches in adults). This is called the Interaural Time Difference (ITD) (not to be confused with Beranek’s Initial Time Delay), or when measured as the phase shift of a sinusoidal tone, the Interaural Phase Difference (IPD).
Thus, in theory, if the sound source is located directly in front of the receiving ‘head’, and the room is exactly symmetrical, and the head is on the line of symmetry, then the sound at the two ears will be identical. Therefore, the IACC will have a high value at all frequencies. In reality, the signals will never be identical, therefore you will have a varying correlation value at different frequencies.
And yes, I know we are ignoring the inner ear contributions, but we will quit before anyone’s head explodes….}
OK, we’re back from the side trip into psycho-acoustics land… (and you had thought acoustics was weird... ;-)
So what does all of this have to do with LF localization – or that lack of it?
At very low frequencies, the wavelength of the sound in air is much much larger than the ~17 cm / 8 inch distance between the ears/microphones, and therefore the pressure at the two microphones will be almost identical.
The lower the frequency, the more similar the two signals and the higher the correlation. Thus, as the mind is dependent upon the differences in order to gauge localization, the lower the frequency, the less useful they are for localization purposes.
BTW, these characteristics can be specifically evaluated and incorporated into measurements in platforms such as ARTA and Easera and in measurements such as the polar(3D) ETC.
As far as the effective frequency range of human hearing, we often think in terms of the typically stated as 20 Hz to 20 kHz. But this is not as simple as it appears, as the functional aspects vary dramatically over that range. Decreasing in frequency below 1000 Hz, the hearing threshold increases dramatically from 0dB at 1000 Hz to required levels of ~90 dB at 20 Hz! One can frequently see the practical feature that attempts to address this issue in the form of a loudness contour control which boosts low frequency gain signals at low average volume levels.
Thus, without delving onto the timing and relative gain issues involved with localization here, for localization and imaging issues we are concerned primarily with a rather restricted mid-range band of energy as the ears’ greatest sensitivity is normally in the frequency range of from 1 kHz to 4 kHz. (If you want to lower that a bit to ~750 or 800 Hz, I won’t argue.)
But energy below 200 Hz is of little to no practical value for use in localization.
Thus it has little bearing on the Precedence effect. We, quite frankly, have bigger more substantial issues to address.
And the primary manifestation of LF direct and indirect energy is in the form of modal (standing wave) resonance.
A practical example for those who served in the military and/or have been exposed to helicopters, you know how difficult it is to identify the approach vector of an unseen slick for which you only have LF sonic clues (and assuming they are ours, works to our advantage!).
Instead we are more worried about its resonance, as it is perceived as boomy and tends to mask other sounds, which is why we need to control the modal behavior with LF traps.
"
https://www.gearslutz.com/board/6202822-post67.html
Depends on the frequency as to what the authorities recommend. I thought this thread was about subwoofers... is it not?
Well I guess I am confused. I thought the thread was about subs. But here you're talking about 100 - 300 Hz?
This has been studied extensively. Many papers published in JAES by Toole and others are not available to repost here. Nothing I can add to their work will satisfy you.
Nothing in that lengthy quote contradicts what I have said here. Indeed, the first sentence or so defines what Fc is and its relationship to minimum phase behavior below and non minimum phase behavior above. And the rest (which is spot on, btw, if repetitive and unnecessarily obtuse) goes on to explain why specular reflections, which are non minimum phase phonomena, above Fc cannot be corrected by traditional minimum phase IIR filters (all passive and active analog eq/filters save impractical tapped delay lines, and even most digital filter implementations).
But subwoofers generally operate below the Schroeder frequency. Except yours that is, which apparently operates from 100 - 300 Hz.
Last edited:
Okay. the op question is if subs and bass traps do the same thing. And My naswer is absolutely not. as this thread shows, bass traps will increase the upper bass, something that subwoofer and obviously eq cant. OP also opened a thread recently about him complaining about unsatisfactory bass IIRC.
im very curious about subwoofer that could help significantly room resonance under 100hz...where absorption is less and less useful and completely useless basically under 40 or 50hz I think.
Ive read most of Toole papers, theres not many waterfalls graph and indeed it shows improvements when taming HUGE peaks with EQ. However, those graphs are only with one seat; id need to see the behavior of that tamed peak at many seats.
''It seems to me that one limitation that is constantly brought up (and is not really argued by anyone) is the fact that any correction is good only at one location. Given that the more untreated a room is, the higher the Q of the peaks and nulls are in the room, and the more these vary with position in the room, it seems that the best scenario for the use of these kind of "room correction" programs is a room that is already quite well treated with as little variance in frequency response throughout the room and with relatively wide Q peaks and nulls. At this point, EQ could be used to bring down some peaks in the minimum phase region, and the results will not be as variant at different points in the room.''
In gearslutz, theres a attempt that seems successful by using EQ to tame room resonance:
https://www.gearslutz.com/board/7723926-post34.html
But again, what is really missing is what happens when you move the mic 3 feet away from that spot...
"Before we start adjusting an EQ to alter the frequency response, we need to see a response to adjust, so we need to make a measurement. This brings up the first limitation. The measurement is made at a single position, and the frequency response of that measurement is only valid at that position, moving the mic elsewhere and making another measurement will produce a different frequency response. It may be a little different, or it may be (and usually is) a lot different. The changes made by an equaliser in the path to the speaker are the same no matter where we are in the room, so since the response is changing in different positions and the EQ isn't, it stands to reason that the EQ is only going to be good in places where the frequency response is the same as the one we used when setting the EQ.
The best you can do is to look at the frequency responses measured at many positions in the area where you need the correction to work, figure out which bits of them are sufficiently common, and come up with a compromise EQ setting that helps somewhat in most places and doesn't do too much harm elsewhere. It can help, but it is no magic bullet."
Toole said:
“Room resonances at low frequencies behave as “minimum phase” phenomena, and so, if the amplitude vs. frequency characteristic is corrected, so also will the phase vs. frequency characteristic. If both amplitude and phase responses are fixed, then it must be true that the transient response must be fixed – i.e. the ringing, or overhang, must be eliminated” (Toole, The Acoustical Design Of Home Theaters, 1999)
But again, what is missing is what happens to those tamed peaks with EQ 6 feet from where the measurements was taken?? We need to see the behavior of the after EQ across the room....
If Mathematically equipped one might debate the reasoning Toole use, but ultimately I don't think the answer I/we want lies there. i.e. How wide is the area usefully corrected?
Time for a new test IMO, as time passes an old wisdom may become a new myth. And let's test in a non comparative way, i.e. not Eq vs Treatment.
Id like to see a comparison between using effective pressure based bass traps down to 50-60hz VS to using a EQ to see if theres really improvement in the time domain using EQ at multiple seats. I personally dont believe EQ can come close to taming room resonance as evidently as that
https://www.gearslutz.com/board/7597560-post146.html
So I think we agree about all that I just said and that the solution Toole and Geddes is to use Subwoofer for the minimum phase region.
How do subwoofers tame the room resonance and reduce decay times?
I think youve also mentionned this, but I'm not sure if I understand. Do using multiple subwoofer in room will also help for LF resonance all across the room or simply for frequency response at one exact location?
im very curious about subwoofer that could help significantly room resonance under 100hz...where absorption is less and less useful and completely useless basically under 40 or 50hz I think.
Ive read most of Toole papers, theres not many waterfalls graph and indeed it shows improvements when taming HUGE peaks with EQ. However, those graphs are only with one seat; id need to see the behavior of that tamed peak at many seats.
''It seems to me that one limitation that is constantly brought up (and is not really argued by anyone) is the fact that any correction is good only at one location. Given that the more untreated a room is, the higher the Q of the peaks and nulls are in the room, and the more these vary with position in the room, it seems that the best scenario for the use of these kind of "room correction" programs is a room that is already quite well treated with as little variance in frequency response throughout the room and with relatively wide Q peaks and nulls. At this point, EQ could be used to bring down some peaks in the minimum phase region, and the results will not be as variant at different points in the room.''
In gearslutz, theres a attempt that seems successful by using EQ to tame room resonance:
https://www.gearslutz.com/board/7723926-post34.html
But again, what is really missing is what happens when you move the mic 3 feet away from that spot...
"Before we start adjusting an EQ to alter the frequency response, we need to see a response to adjust, so we need to make a measurement. This brings up the first limitation. The measurement is made at a single position, and the frequency response of that measurement is only valid at that position, moving the mic elsewhere and making another measurement will produce a different frequency response. It may be a little different, or it may be (and usually is) a lot different. The changes made by an equaliser in the path to the speaker are the same no matter where we are in the room, so since the response is changing in different positions and the EQ isn't, it stands to reason that the EQ is only going to be good in places where the frequency response is the same as the one we used when setting the EQ.
The best you can do is to look at the frequency responses measured at many positions in the area where you need the correction to work, figure out which bits of them are sufficiently common, and come up with a compromise EQ setting that helps somewhat in most places and doesn't do too much harm elsewhere. It can help, but it is no magic bullet."
Toole said:
“Room resonances at low frequencies behave as “minimum phase” phenomena, and so, if the amplitude vs. frequency characteristic is corrected, so also will the phase vs. frequency characteristic. If both amplitude and phase responses are fixed, then it must be true that the transient response must be fixed – i.e. the ringing, or overhang, must be eliminated” (Toole, The Acoustical Design Of Home Theaters, 1999)
But again, what is missing is what happens to those tamed peaks with EQ 6 feet from where the measurements was taken?? We need to see the behavior of the after EQ across the room....
If Mathematically equipped one might debate the reasoning Toole use, but ultimately I don't think the answer I/we want lies there. i.e. How wide is the area usefully corrected?
Time for a new test IMO, as time passes an old wisdom may become a new myth. And let's test in a non comparative way, i.e. not Eq vs Treatment.
Id like to see a comparison between using effective pressure based bass traps down to 50-60hz VS to using a EQ to see if theres really improvement in the time domain using EQ at multiple seats. I personally dont believe EQ can come close to taming room resonance as evidently as that
https://www.gearslutz.com/board/7597560-post146.html
So I think we agree about all that I just said and that the solution Toole and Geddes is to use Subwoofer for the minimum phase region.
How do subwoofers tame the room resonance and reduce decay times?
I think youve also mentionned this, but I'm not sure if I understand. Do using multiple subwoofer in room will also help for LF resonance all across the room or simply for frequency response at one exact location?
Last edited:
Clearly there are differences of opinion on some of the finer points.
I would guess that the smoothing of the steady-state frequency response played a significant role in the subjective improvements you have observed, as I've observed similar subjective improvements from the distributed multisub approach. What I don't know is, the net subjective effect of the reduction in reverberation from the CABS approach. I can think of reasons why that might be beneficial, and reasons why it might be detrimental, and I'm not psychoacoustician enough to intuit which trend(s) would dominate perceptually. Hence my earlier post saying that the CABS concept is something I intend to try for myself at some point.
The improvement is all across the room. The response variation from one location to another is greatly reduced with a distributed multisub system. Each sub produces its own unique in-room peak-and-dip pattern at any given listening position, and the sum of these dissimilar peak-and-dip patterns is necessarily considerably smoother than any one alone. The only way the sum could be as bad as any one alone would be if they were all identical, and that could only be approximated by clustering the subs, which is pretty much the opposite of what is recommended.
The resulting summed response will have more peaks and dips, and they will be closer together. Recall that the ear/brain system tends to sum peaks and dips that lie within the same critical band, and you'll see that we have the possibility of even greater psychoacoustic benefit than is obvious from eyeballing the curves.
Duke
Last edited:
Looking at a waterfall chart, I'd define ringing as the frequencies which show significantly longer decay times than the rest of the spectrum. Even after EQ'ing down for a flat frequency response, those same frequencies still show longer decay times than the rest of the spectrum, so I'd say they are still ringing. (except on the chart I quoted where 42Hz has been boosted a couple of dB and the ringing has gone down, which didn't make sense to me, but apparently is possible)
Isn't Follgott [sp? 1 'l'?] a Diyaudio member?
Ya, here we go. See wall of death
http://www.diyaudio.com/forums/members/follgott.html
http://www.hannover-hardcore.de/infinity_classics/!!!/Dokumentation SBA-01.pdf
Ya, here we go. See wall of death
http://www.diyaudio.com/forums/members/follgott.html
http://www.hannover-hardcore.de/infinity_classics/!!!/Dokumentation SBA-01.pdf
Last edited:
You are right in that a single sub + EQ is limited in how effectively you can correct modal behavior of room+sub interaction because the corrections are location specific.
The key, going back many many years, is the use of multiple subs. There are better and lesser effective schemes for placement (four corners, midwalls, distributed vertically as well, etc.) that have their own proponents, but in general distributing multiple subs in the room has on its own a large impact on smoothing response (both frequency, and time) because of a more even excitation of room modes (plus in some placement schemes, some active cancellation of dominant modes as energy is intentionally added out of phase).
But once you have multiple subs, EQ has an equally multiplied increase in effectiveness. It gets very complicated... REW plus lots of time and trial and error can produce quite good results. Or you can leave it to some smart system like Harmon's JBL Synthesis products which take many measurements in the room and use an optimization algorithm varying frequency response, delay and phase of each sub independently to reach the best possible response within a defined target seating area.
Yes! Room treatment still has many advantages. A big one is that it is effective everywhere in the room. And it is absolutely necessary above Fc, so if you're already going all in on room treatments to do it right, putting some thought into the low frequency region isn't that much of an extra design effort.
But the obvious disadvantage is that the size of treatments necessary becomes large relative to the room itself, so there is a practical limit to how low you can reasonably address.
But in general, yes, I agree, treatment should come before EQ. IMO, it should come after placement of multiple subs though (well, both are equally necessary, just that treatments are more necessary above Fc and multiple subs more necessary below).
That being said, in my own theater I'm now in the process of designing, it will be heavy on room treatments and likely only utilize stereo subwoofers that are vertically distributed, breaking my own rule/suggestion. This also is for practical limitations, as the subs will be IB in design, and handling multiple distributed manifold locations doesn't seem possible with the physical construction of the house and room. I'm partially combating this with a multiway approach - 1. Frequency of subs likely limited to 60Hz and below, 2. Distributed drivers from 60Hz and above, 3. Extensive room treatments including some physically large absorptive bass trapping, some having dimensions on the order of tens of feet, 4. Complex room shape with more finely distributed modal peaks, 5. EQ
And in the end, I may still find I need to place a few smaller sealed subwoofers elsewhere in or around the room for response tailoring purposes only.
thanks for this thread, I see both the shortcomings and strenght of both bass traps and subwoofer. Since treatment becomes less and less useful (unless VERY large) under 60hz, subwoofer is the way to go and for over 80hz, bass traps and treatment is amazing.
I wonder though how subwoofer deal with room resonance. in posibbly my flawed understanding, it seems that adding more source of subs would seem to add room resonance.
My system uses amphion one18 which are surprisingly flat down very low to 45hz. In my room, I have only a slight problem that isnt too bad between 60 and 80 hz, something a sub couldnt solve due to how low my mains go... but I might try to implement subwoofers again, If I remember it did solve the minor dips at 70hz, but I never quite liked listening to music with my subwoofers.
Say you have an 8" driver in an 8"x8"x64" cabinet, placed at the top of the baffle and there is a pipe resonance along the long dimension of the cabinet. Now change that to be eight 8" drivers spanning the baffle. The behaviour will be such that there would be no difference if you were to put an internal divider in between each driver.
That's an interesting analogy. I would maybe add that without the internal dividers the pipe resonance will still exist.... but the levels at any one point in the pipe are mostly swamped by the output of the nearest driver. The pipe resonances get almost completely masked...
Hey, thanks for this! That's great info....
Nils (Follgott) shows some very good information about permutations of the DoubleBassArray, including some showing it can work really well in irregularly shaped rooms, see these in diyAudio - View Profile: FoLLgoTT
yes they're in German, but Google Translate made it very easy to understand.
Also, looks like for his latest project "wall of death", due to cost issues, he replaced the rear array with a massive amount of rockwool and some minor EQ
Here is a GREAT study he did on DBA permutations:
http://hannover-hardcore.de/infinity_classics/!!!/Alternative DBA-Anordnungen.pdf
Of particular interest is the case study on Page 8 (Seite 8 in German) that shows how a 1/2 DBA with two woofers placed at half-height in the front of room, and 1/4 of the room width away from the side walls, with the same on the back wall, performs nearly exactly like the full DBA with 4 drivers, and gives near optimum response with minimal side-to-side or front-to-back variation anywhere in the room, all the way to 200Hz!
This could greatly reduce or eliminate the need for very large bass traps and absorbent treatments, maybe all that is needed then is deflectors and diffusers.
His stuff makes for a very interesting read.
Last edited:
Yes Jack, Follgott has piqued my interest and I have an idea. That is to make a sub - driver in box, amp and microphone.. a standalone active absorber (eg for the back of the room). Not only would it save processing but it might be more accurate and versatile.
I don't see this approach as strictly necesary insofar as the conventional multisub setup can do the same thing but I like the simple elegance of the double bass array.
Sounds like a neat thing to do with surplus woofers and amps
I don't see this approach as strictly necesary insofar as the conventional multisub setup can do the same thing but I like the simple elegance of the double bass array.
Sounds like a neat thing to do with surplus woofers and amps
Shades of Nelson Pass and more recently Clayton Shaw's Black Hole!
That's definitely something that can work, and work well... but make sure you use TWO of them (see Follgott's comparisons, the single absorber doesn't fare so well)...
However I think the use of the mic can create needless complexity when compared to the simplicity of using a Behringer DCX2496... first off , you'll need to use a mic pre-amp and phantom power prior to your amp for each active bass absorber. Not overly difficult to do, but more pieces and complexity.
Also, this method requires some fairly judicious experimentation and tweaking of the low-pass filter and of the gain levels for the mic/amp combination, as you don't want it to be putting out any spurious signals... if you get too high a roll-off or too much gain, you are likely to create/hear some mid-bass howl or artifacts with this approach.
There is a much simpler way:
I would urge you to look at using the Behringer DCX2496 as it is probably the easiest way to implement DBA, and generates no spuriae. It accepts a digital input with up to 6 independent channels output, with adjustable gain, eq, delay for each.
Here is a simple example of how it could be used:
Supposing a 22ft length to the listening room, with two woofers at the front, the main speakers at half-way point and two woofers in the back...
1) Use the front DBA array with no delay.
2) Add delay to the mains so that it is in perfect sync with the wave from the front of the DBA. In this example, we want to compensate for an 11ft difference, that would be 10msec of delay (11ft/1100 ft/sec)
3) Add delay to the rear DBA woofers, in this case that would be 20mSec delay (22ft/1100ft/sec), invert polarity and adjust level for complete cancellation of the bass pressure wave (at the back wall).
From a purist's perspective I lean much more toward DBA rather than the distributed multi-sub approach, for example w Geddes system each sub gets its own eq curve... whereas with DBA an eq curve may not be necessary at all if the subs have inherently flat response.
Note: If you try this, make sure you are using TWO of them at the back, Follgott's work showed that having only one absorber was a big compromise relative to what can be obtained with two.
It also looks like there isn't all that much extra to be gained by using 4 on front and 4 in back... the ideal solution looks like the "1/2 height DBA" version, comprised of TWO subs in front, and TWO in back (all at 1/2 the room height), placed 1/4 of the room width away from the side-walls
With all of this, the other DBA approach that could be taken is to make a modified version of what Follgott did with his "wall of death" HT project... instead of using the rear DBA woofers, he placed a HUGE amount of rockwool at the back wall (simply stacked the bags along the back wall, floor to ceiling, and did not remove the plastic... it works great!)...
so the modified version would be to use the 1/2 DBA concept with just TWO subs at the front, placed at 1/2 room height, 1/4 of room width from the side walls, and add massive amounts of rockwool at the back.
100 x 50 lb bags of Rockwool? Expensive as heck, and needs to be hidden... yeesh, try and make that LOOK good too! Nah, I'd go for the two rear 1/2 DBA active bass absorbers any day...
Last edited:
100 bags of 50lb rockwool. jeez that extreme! I figure the ultimate maximum a pro studio would put is about 25. Most project ive seen use maximum 10. actually even 10 is a lot. ive use 5 bags in my room, many put 2-3 bags.
AND, for a really serious project, pressure based bass traps would be used to try to treat the very LF (and they need much less rockwool to work), velocity based bass traps cannot reach very low. I dont really see who can use 50 bags of rockwool.
The best method is to use both velocity and pressure based bass trap.
AND, for a really serious project, pressure based bass traps would be used to try to treat the very LF (and they need much less rockwool to work), velocity based bass traps cannot reach very low. I dont really see who can use 50 bags of rockwool.
The best method is to use both velocity and pressure based bass trap.
Last edited:
Did you look at Follgott's "wall of death" project? Sure looks like it worked for him... the measurements sure do show some good performance.... have a look at the photo he posted of his back wall... it is entirely covered in bags of rockwool... looks like at least 100 of them
Maybe he should have called it the "wall of rockwool"
Arrgh, I can't find it.... Maybe the one I saw was from a different thread or person, the recent pix of Follgott's unfinished HT room don't show all that many bags of rockwool... still a whole bunch of them but not as many as I had seen in the other pic.