Under many circumstances, optimizing frequency response and optimizing transient response is a zero sum game. Here’s what I mean: Much of the time and effort that goes into subwoofer setup is spent optimizing frequency response. That is typically done in one of two non-mutually-exclusive ways: Sub placement or digital EQ. Either of these approaches can result in a much flatter frequency response, making the bass sound less bloated. But both of these approaches can result in a sub-optimal transient response, due to the time misalignment between the sub and the mains. That makes the bass sound slow. To elaborate… If placement is used to optimize the sub’s frequency response, then the sub often winds up closer to or farther from the listener than the mains, thereby misaligning the sub relative to the mains. Alternatively, if digital EQ is used to optimize the sub’s frequency response, then a processing latency is introduced into the sub’s the signal path, but not into the main’s signal path, thereby misaligning the sub relative to the mains. Hence either approach to optimizing frequency response can disturb the system's transient response, making the bass sound slow (or “out of sync”). On the other hand… If you set out to optimize transient response, you will usually place the subs on a plane very close to that of the mains. Now the bass no longer sounds slow. But, due to room modes, this kind of placement often results in a highly uneven frequency response. Now the bass sounds bloated again...and so on. The result of all this is that, when trying to optimize both frequency response and transient response, you often have the experience of chasing your tail. That is what I mean when I say that, under many circumstances, optimizing frequency response and optimizing transient response is a zero sum game. I do believe that there are solutions to this problem, e.g., the use of software that independently controls the EQ and the delay of both the sub and the mains. But most audiophiles who use subs don’t seem to avail themselves of that kind of solution, which can result in their systems sounding bloated or slow. Bryon |
Thanks for the responses so far. In the OP, I wasn't trying to suggest that all subs sound bloated or slow. I was trying to identify some reasons, WHEN subs do sound bloated or slow, WHY that is the case.
Undoubtably, some of those reasons involve the design of the subwoofer itself, as Shadorne indicated. But I have noticed that, even with a well designed sub, seamless integration into a 2 channel system is very challenging. In the OP, I was proposing a theory that explains why sub integration into 2 channel systems so often fails, namely that it tends to result in one of the following scenarios:
1. Good transient response but poor frequency response. Hence "bloated."
2. Good frequency response but poor transient response. Hence "slow."
The reason why the above scenarios are so common, I believe, is because the measures taken to optimize frequency response are often bad for transient response, while the measures taken to optimize transient response are often bad for frequency response.
Has anyone else had this experience during sub setup?
Bryon |
01-23-11: Plato
If the subs are not located in the same horizontal plane as the main speakers (and close to them) as is usually the case when they are placed in corners or close to the front wall, then unless you have subs with fully variable phase controls of at least 0 to 180 degrees, you'll need to do extensive trial and error to get the phase relationship correct between the subs and the sats. And yes, that's pretty much as hard as it sounds. Hi Frank - I agree with you that finding the optimal phase setting is a major challenge of subwoofer setup. But I would add that, IME, phase adjustment alone is often not an adequate way to time align the sub with the mains. The reason was stated by Cbw723: I'm not sure the automatic setup system work any better. They are primarily EQ programs, and do nothing for the transient response. And, frankly, it would be very hard for them to fix transient response because that requires playing the mains and sub(s) at the same time and (usually, because of sub placement and processing delays) delaying the MAINS. The phase control on a sub won't do this either and, to the extent that it is useful at all, can only work when the sub is closer to the listener than the mains. I agree with Cbw723 that neither phase adjustment nor EQ (whether manual or automated) will adequately address the time alignment problems created by placing the sub farther from the listener than the mains, which is probably the most common setup in the typical audiophile listening room. The reason phase adjustment cannot adequately address the time alignment problems of this setup is that, by placing the sub farther from the listener than the mains, the speakers that need to be delayed for proper time alignment are the MAINS, and NOT the sub. Obviously, the phase adjustment on the sub can do nothing for this. The common use of digital EQ on the sub just makes the time alignment problem worse (even though it can dramatically help frequency response), because it introduces a processing latency that further delays the sub's output relative to the mains. IME, for setups in which the sub is farther from the listener than the mains, the only way to time align the system is to be able to DELAY THE MAINS. But, judging from the systems here on A'gon, very few audiophile systems have this capability. For systems that do not have the capability of delaying the mains, time aligning the sub with the mains requires that the sub be placed more or less coplanar with the mains. But placing the sub coplanar with the mains might not result in the best frequency response. Which brings me back to my initial observation in the OP that... Under many circumstances, optimizing frequency response and optimizing transient response is a zero sum game. IMO, the way to defeat the zero sum nature of this game is to: 1. Place the sub(s) to get the best frequency response (varies from room to room) and fix transient response problems with DELAY. This assumes you can independently delay the sub(s) and the mains, which as I mentioned above, doesn't seem to be a common capability in audiophile systems. -OR- 2. Place the sub(s) to get the best transient response (i.e. roughly coplanar with the mains) and fix frequency response problems with EQ. But to the extent that the EQ introduces processing latency, you will have to move the sub(s) CLOSER to the listener than the mains. Again, this doesn't seem to be a common arrangement in audiophile systems. -OR- 3. In light of (1) and (2), I have recently come to the conclusion that the most effective way to optimize both frequency response and transient response is to be able to independently control BOTH THE EQ AND THE DELAY of both the sub(s) and the mains. That allows you to correct for room modes (better frequency response) and time align the various speakers (better transient response). Of course, I could be wrong about this. And... This all this assumes that the time alignment of the sub(s) with the mains affects the transient response of the system in ways that are (a) audible, and (b) not reducible to changes in frequency response. Duke (Audiokinesis) has expressed doubts about that assumption, raising the question: What are the limits to the temporal resolution of human hearing at low frequencies? As I understand it, that is a subject about which there is some controversy. Bryon |
01-25-11: Martykl
There's an element of "When did you stop beating your wife?" to this discussion. It starts with the presumption of guilt (for subwoofers) and asks for proof of innocence. Hi Marty - I regret the title that I chose for this thread for a variety of reasons, including the one you mention. It is was not my intention to imply that subs ALWAYS sound bloated or slow, as I mentioned in a post the same day I started the thread: In the OP, I wasn't trying to suggest that all subs sound bloated or slow. I was trying to identify some reasons, WHEN subs do sound bloated or slow, WHY that is the case. The principal reason I identified in the OP is a conclusion that I arrived at after extensive experimentation in my system and others, namely that optimizing frequency response often comes at the expense of optimizing transient response, since efforts to optimize frequency response usually include non-coplanar placement or EQ, both of which disturb the sub's time alignment with the mains, and hence the transient response of the system. The use of the of the terms "bloated" and "slow" were not intended to be derisive, but rather descriptive. It seems to me that room modes often result in non-flat frequency response that can be described as "bloated," and that time misaligned subs often sound out of sync with the mains, which can be described as "slow." Having said that, I quite agree that bloated or slow bass is not a problem unique to systems with subs. But what IS unique to systems with subs, IMO, is the "zero sum game" that results from scenarios in which efforts to optimize the system's frequency response disturb the system's transient response, and vice versa. The reason this is unique to systems with subs is that, for systems without subs, the time alignment of the system is determined almost entirely by the manufacturer, not the end user, for the simple reason that the woofers are physically attached to the same cabinet as the other drivers. For systems with subs, frequency response and transient response (for bass) are independently controllable parameters. Hence systems with subs have unique challenges, including, IME, the challenge of optimizing both frequency response and transient response. That was the central idea in the OP, and the idea I've been advancing throughout the thread. Part of the confusion here might have been avoided if A'gon didn't limit my initial post to a little over 100 words. I have no idea why that is done, and I have seen many other threads in which the OP was far more than 100 words. FWIW, I use a sub in my 2 channel system, and it sounds neither "bloated" nor "slow." But it took me literally years of periodic experimentation to arrive at that point. The reason, I believe, is the "zero sum game" I've been trying to describe. Bryon |
Is there a moratorium on serious conversation?
Bryon |
BIF - IME, the two simplest and most effective techniques for subwoofer placement that require no equipment other than a SPL meter are: 1. Optimizing frequency response by placing the sub at the listening position, then walking around the room and listening for the location where the bass is *consistent* across low frequencies. You can do this by ear with music, or you can use a SPL meter and low frequency test tones. Once you find the location in the room that has the smoothest frequency response, you place the sub in that location. Now the bass response at the listening position should be in pretty good shape. 2. Optimizing time alignment by maximizing constructive interference at the crossover frequency. This technique was described by Cbw in his post on 1/27: You flip the polarity of either your mains or your subs (whichever is easier), then play a tone at the crossover frequency (e.g., and 80 Hz tone for an 80 Hz crossover). You then adjust the delay (or sub position) to *minimize* the SPL at the listening position. (You're effectively maximizing destructive interference.) When you're done, you flip the polarity back to normal, and you should be very close to time aligned. Both of these techniques are simple and highly effective. Unfortunately, IME, they don't always give you the same answer about the best location for the sub (which is the point I've been trying to make throughout the thread). So, if I were you, I would start with the first technique to find a general location for the sub, then use the second technique to refine the placement. Hope that helps. Bryon |
Telescope - The speed at which electricity travels through a wire is so fast that it makes differences in cable lengths inaudible, since the high speed makes any differences in latency (elapsed time) vanishingly small.
So you have nothing to worry about.
Bryon |
I agree with Marty that, for the greatest control over subwoofer integration, you will need some way of measuring and correcting the frequency response of the system. There are both hardware and software approaches to measurement/correction. Both can yield excellent results. Personally, I use a software approach to measurement, namely Room EQ Wizard, and a hardware approach to correction, namely Meridian Room Correction. IME, the ability to measure and correct the system's frequency response dramatically increases your control over sub integration. But if your system doesn't allow for this, the techniques I described in my previous post will get you a good part of the way toward excellent bass. Good luck. Bryon |
I rarely post on tweaks, but I just discovered one that is relevant to this thread: I placed 3 Black Diamond Racing Jumbo Pucks under my sub, which is resting on a three inch maple platform. This, in effect, replaced the soft feet of the sub with extremely hard feet. The result: An audible increase in bass articulation. FWIW. Bryon |
04-21-11: Dbphd
One of the most reliable phenomena in psychoacoustics is what is know as the masking level difference (MLD). Present a mid-freuquncy sinusoid in correlated noise to both ears and adjust the level until it becomes inaudible; flip the phase of the sinusoid in one ear, and the signal pops up as much as 15 dB, depending on frequency. Dbphd - My understanding is that BMLD is a measure of the difference between the thresholds of detection for… 1. Auditory stimuli in which the signal and noise are the same phase and level. ...and… 2. Auditory stimuli in which the signal and noise are different in phase and/or level. It’s unclear to me how the existence of BMLD affects my observations about subwoofer time alignment. Maybe you can elaborate. At the time, the data suggested the auditory system doesn't preserve timing as such up the neural chain, but may convey such information by the more central areas excited. Again, it’s unclear to me how this bears on my observations about subwoofer time alignment. I’m not saying it doesn’t. I just don’t see the connection. Maybe you can say more. Having said that, I do have a few thoughts... As you are probably aware, there has been a great deal of neuroscientific research over the past 20 years using fMRI and PET scans. Much of that research has been directed at correlating various brain regions with perceptual, linguistic, and motor abilities/deficits. A significant amount of that research has been devoted to auditory perception, including the brain regions that correlate with temporal processing and temporal resolution. I’m under the impression that the neuroscientific research on temporal resolution has discovered a number of neural correlates for auditory temporal resolution, corroborating earlier measurements of temporal resolution obtained through psychoacoutic experiments (e.g. gap detection). Here is an excerpt from an article in the journal Cerebral Cortex: We used positron emission tomography to examine the response of human auditory cortex to spectral and temporal variation…Results indicated that (i) the core auditory cortex in both hemispheres responded to temporal variation, while the anterior superior temporal areas bilaterally responded to the spectral variation; and (ii) responses to the temporal features were weighted towards the left, while responses to the spectral features were weighted towards the right. These findings confirm the specialization of the left-hemisphere auditory cortex for rapid temporal processing, and indicate that core areas are especially involved in these processes. Here is an excerpt from a paper by researchers at UCI and University of Toronto: Our findings of M100 modulation by the shortest gap (2 ms) tested are also in good accord with animal studies of auditory cortical temporal acuity, where gap detection thresholds have been measured using electrophysiological methods to record activity in single or cluster units. A key result of those studies is that the firing patterns of neurons in auditory cortex reflect minimum detectable gap thresholds that are similar in scale (at 2-10 ms) to thresholds measured psychophysically in human [4, 5, 13]. Our MEG findings reported here provide evidence for a similar level of temporal resolution to brief (2ms) discontinuities in sounds in the synchronized neural response of tens of thousands of neurons in secondary auditory cortical fields [10], reflecting neural response properties at the population level in auditory cortex. Research like this leaves me with the impression that human temporal resolution is highly sensitive, not only when measured psychoacoustically, but also when measured neurologically. Of course, it does not *necessarily* follow that subwoofer time alignment is audible. But it does seem to indicate that my statements about subwoofer time alignment are not *disproved* by the current state of neuroscientific research. Bryon |
My bottom line is that when I optimize my subs for smoothest FR in a co-planar arrangement it sounds no better (or worse) than when the subs are FR optimized for placement 2 1/2 feet behind the mains...I'm pretty confident (can't be 100% sure, though) that no one would characterize the bass that I'm currently producing as either bloated or slow... So, my experience just runs counter to your theory. I don't doubt your experiences, Marty, and I thank you for sharing them. FWIW, the theory I proposed in the OP wasn't intended to be an exhaustive account of the variables that determine good subwoofer integration. I have no doubt that there are many things I have failed to consider. In fact, that was the principal reason I started the thread - to explore some of those things with people who, like me, use subs in their 2 channel systems. And FWIW, I'm not trying to convert anyone to a theory that I only tentatively believe. I apologize if I appear that way. Bryon |
01-23-11: Audiokinesis
Both the bloat and the slowness you describe are primarily frequency response problems; at low frequencies the ear has very poor resolution in the time domain so it doesn't really pick up the transient response in and of itself. This is precisely what I was hoping to discuss. Duke - As I mentioned, my theory about frequency response and transient response tending to be a zero sum game is just that - a theory. Which is a way of acknowledging that I very well may be wrong. Having said that, the theory that optimizing frequency response often comes at the expense of transient response, or vice versa, explains many of my experiences with subwoofer setups using frequency response software. Of course, there may be other explanations for my experiences. IME, small differences in sub placement on the z axis of the room often result in significant audible changes in bass response, EVEN WHEN there are little or no measurable differences in frequency response. Since the audible changes don't seem attributable to differences in frequency response, I eventually concluded that they were attributable to differences in transient response, or more precisely, differences in the system's time alignment, where 'alignment' is understood in terms of the arrival times of the system's various drivers at the listening position. That conclusion is consistent with the audible changes being a function of z axis placement. Encouraging that conclusion is the impression that, for some z axis sub placements, the bass sounds "in sync," whereas for others, it sounds "out of sync." Again, this is sometimes the case even when two placements have a similar or identical measurable frequency response. On the basis of these experiences, I concluded that, when two different z axis sub placements measure similarly in frequency response, but one of them sounds significantly better than the other, the reason is because the better-sounding placement has a more ideal time alignment with the other speakers in the system. Again, I recognize that this conclusion could be a mistake. Bryon |
01-24-11: Shadorne
Spatial issues are not that important in the LF as we are talking only a few milliseconds with waveforms that are 50 milliseconds long at 20 Hz. Shadorne - I agree that, at 20Hz, which is a wavelength of over 50 feet, and which corresponds to a time scale of around 50 ms, a few milliseconds of time misalignment is probably no big deal. But what about at 80Hz, which is a wavelength of about 14 feet, and which corresponds to a time scale of around 14 ms? Let's say the delay introduced by EQ is something like 2-4 ms, and that the delay introduced by placing a sub in the corner, when the mains are out in the room, is something like 3-6 ms. In systems that combine corner placement with EQ, you might have a delay between 5-10 ms. All of these numbers are an appreciable fraction of 14, the number of milliseconds that corresponds to a wavelength of 14 feet and a frequency of 80Hz. IMO, this math illustrates that the use of EQ and/or the use of corner sub placement could introduce delays that result in significant time misalignments. As to whether those time misalignments are audible, I can only say that, from my experiences with sub setups, differences of LESS THAN A FOOT are easily audible, even in circumstances where there is little or no difference in frequency response. Transient response is important but spatial positioning of a sub is not the big deal you make it out to be - as long as sub and speakers are within a few feet of being the same distance to the listener then no big deal. You may be right that I am exaggerating the importance of time alignment at low frequencies, though it is not my intent to exaggerate. My view about the importance of time alignment at low frequencies resulted from about 2 years of periodic experimentation with placement, EQ, and delay. But I suppose the view I'm advancing might be considered unconventional. Though, if it is unconventional, then why do several different manufacturers of full range speakers make an effort to time align their drivers, INCLUDING the woofers? It seems to me that the importance (or lack thereof) of time alignment at low frequencies essentially depends on the human limits of temporal resolution at low frequencies. The limits of temporal resolution has been studied in depth, where estimates place it on the order of magnitude of MICROseconds, with some estimates placing it as low as 5 μs. But these studies tend to measure higher frequencies. I do not know the human limits of temporal resolution at lower frequencies. If they are anything even vaguely approaching those of higher frequencies, then it would lend credence to the idea that time alignment is audible at low frequencies, making time alignment an important consideration in subwoofer setup. Bryon |
01-25-11: Martykl
Small movements of the source of deep bass can audibly affect the perceived sound, but once again I'd be inclined to attribute this to room interactions (rather than increased group delay)... Marty - You, I, and Cbw are all in agreement that small changes in sub placement can result in significant audible changes at the listening position. Where we differ is what we attribute those audible changes to. You attribute them to differences in room interactions, by which you may or may not mean frequency response. Cbw and I attribute them to differences in time alignment, and hence transient response. Cbw suggested a reason to doubt whether the audible differences at the listening position resulting from small changes in sub placement are attributable to room interactions: Marty, I don't think the effect I described can be reduced to room interaction because, as I said in my earlier post, I can achieve the same result by adjusting the delay of the mains relative to the subs without physically moving anything. I have experienced the exact same thing. That is to say, I can produce significant audible changes at the listening position by changing the digital delay on my sub, without altering sub placement. Having said that, I have to acknowledge that changing the delay on the sub DOES change the "room interactions" insofar as it affects the patterns of constructive and destructive interference between the sub and the mains, which are mediated by the size and shape of THE ROOM. So changing the delay on the sub does change its room interactions. Hence it might be reasonable to conclude that differences in room interactions explain the audible changes at the listening position. But even if this is true, it still leaves open the question of whether those audible changes at the listening position are attributable to frequency response or transient response. The reason is because constructive and destructive interference, while being forms of "room interaction," alter BOTH the frequency response AND the transient response of the system. Because of this, concluding that the audible changes at the listening position are attributable to room interactions does not definitively answer the question of whether those audible changes are attributable to differences in frequency response or differences in transient response or both. In my view, the way to determine whether the audible changes at the listening position resulting from small changes in sub placement are attributable to frequency response or transient response is hold one constant, alter the other, and listen for changes. In my reply to Audiokinesis above, I reported just that kind of finding. That is to say, I have had many experiences in which I have moved the sub on the z axis by a small amount and produced significant audible changes at the listening position, EVEN WHEN the measurable frequency response remained the same for both sub placements. It was precisely those experiences that led me to the conclusion that the audible changes at the listening position were attributable to changes in transient response. Bryon |
01-25-11: Cbw723
Bryon, while I agree with much of what you say, I think room interactions are secondary. The perceived coherence of the signal is going to be primarily determined by the direct waves from the speaker and sub to the ear. I'm not sure I agree with this point. I agree with Josh358 that the "perceived coherence" of the signal is significantly affected both by the direct sound and the indirect sound, and my suspicion is that the indirect sound is even more relevant, since the indirect sound is mediated by room modes, which are highly audible. But this is not a major source of disagreement. What IS a major source of disagreement is the importance (or lack thereof) of transient response at low frequencies, and hence the importance (or lack thereof) of time alignment at low frequencies. It just occurred to me that there is a plausible argument about the importance of time alignment at low frequencies that has nothing to do with with improving transient response, but only with improving frequency response, namely: Time alignment between the sub and the mains improves low frequency response by minimizing the patterns of constructive and destructive interference around the crossover frequency. An explanation of why this is so can be read here (see the second section entitled "Timing is Everything"). For reasons of improving frequency response, some manufacturers of subs explicitly advocate the time alignment of the sub with the mains. For example, Rythmik Audio says: In order to get the most out of your subwoofer, it is critical that it is correctly integrated with the rest of the system...
...The simple method is to compensate by changing the speaker distance setting on your receiver. Bass management in HT receivers has a speaker distance adjustment which process the signal on digital domain. If one puts distance of the sub x feet further away than its physical distance relative to other speakers, the HT receiver will put out the signal to the sub x/1000 sec before it puts out signals to other channels. That essentially puts a negative delay on the sub which can be used to reduces the "phase lag" on the sub and therefore reduces the phase difference between the sub and the front speakers. This trick enables us to use the speaker distance as a tool for phase adjustment between subwoofer and front speakers. The point of all this is that there may be compelling reasons to time align a sub with the mains that have nothing to do with transient response, but only with frequency response. Bryon |
