More than one sub (long-winded but you can skip to the end)
Jun 1, 2017 19:00:59 GMT -6
mulmany, ericn, and 2 more like this
Post by duke on Jun 1, 2017 19:00:59 GMT -6
I was going to post this in jeromemason's “PAIR of subwoofers” thread, but by the time I was done my post had pretty much turned into a full-blown hijack, as you can see from its length. If you're short on time, just skip down to the last paragraph.
Having totally failed to master the technical intricacies of digitally quoting on this site, I hope you will forgive me for reverting to analog:
jeromemason wrote: “Just wondering the thoughts of some that either have or had dual subs going...
“I read an article where two guys were pitching using 4 or more subs in a room...”
Briefly (or in retrospect, perhaps not!), here is a look at the situation with a single sub in a home studio room, and then we'll look at how an additional sub (or subs) can make an improvement. Imagine you have one subwoofer in your room. Its interaction with the room boundaries results in a rather drastic peak-and-dip pattern. You can move the sub and change the peak-and-dip pattern, but you cannot eliminate it by positioning alone. You can mitigate it by a few dB with large bass traps. You can equalize it to be fairly smooth in one location, but in the process you will have made it worse elsewhere. And it is possible that the response still will not be sufficiently smooth at both ears. Quoting jazznoise:
“In small rooms, the issue is that room modes can be so big that they can cause different tones to 'sit' left or right. Interaural time delay as a listening mechanic doesn't work well for bass, it's intensity and phase at low frequencies that allow your brain to attribute directionality. If your right sub puts out a 50 Hz tone which nulls to your right but rings to your left, the localization is gonna be messed up.”
In a small room with a single sub, there's a good chance you'll have at least one frequency region where the situation jazznoise describes applies, and EQing that single sub won't fix it. It's an acoustic problem, and arguably is best addressed with an acoustic solution.
At low frequencies, sub + room = an inseparable system, from the standpoint of perception. The ear/brain system literally cannot hear the sub apart from the room. We cannot detect the presence of bass energy from less than one wavelength, and we cannot determine the pitch until we have heard several cycles (wavelengths). Mentally compare your sub's distance to your various room boundaries with how long low-frequency wavelengths are, and you'll see that by the time you actually hear the low frequency energy, the room's effects are already in full swing.
Back to the room-induced peak-and-dip pattern for a moment. Smooth bass is “fast” bass, because the peaks literally take longer to decay. To put it another way, speakers + room = a minimum-phase system at low frequencies, which means that if we smooth the frequency response, we have simultaneously smoothed the time-domain response, and vice-versa. In-room peaks spoil the pitch definition because, as the energy decays over time (and remember the peaks take longer to decay), the spectrum actually shifts in the direction of the peaks.
If these room-induced peaks and dips were very numerous and bunched up close together, it might look terrible to the eye on a graph, but perceptually it would in effect be a “continuum”, because the ear/brain system tends to average them out (and indeed this is what happens with the reflections at midrange and high frequencies – perceptually they merge into a continuum). But if the peaks and dips are further than about 1/3 octave apart, the span is too great for the ear's averaging mechanism to come into play, so they stick out like sore thumbs. One counter-intuitive implication of this is, the problem isn't that we have too many room modes at low frequencies – it's that we have too few! (Bigger rooms have greater model density at low frequencies, which is why they usually have more natural-sounding bass; the smaller the room, the fewer and further apart the peaks and dips in the bass region.)
If we look at a set of equal-loudness curves, we'll see that they bunch up south of 100 Hz or so. This is telling us that a change in SPL in the bass region makes a disproportionately large change in perceived loudness. For example, a 5 dB change in SPL at 50 Hz makes approximately as much of a difference in perceived loudness as a 10 dB change at 1 kHz!
So to recap, we have three problems that can arise from using a single sub in most rooms:
1. Poor pitch definition due to room-induced peaks;
2. The peaks (and to a lesser extent the dips) perceptually stick out like sore thumbs because they are not only large but few and far apart; and
3. The ear/brain system has a heightened sensitivity to SPL changes south of 100 Hz, so the subjective detriment of these peaks and dips is greater than we would infer simply from eyeballing our in-room frequency response curves.
At the risk of oversimplifying, what usually happens in practice is, we turn down the level of the subwoofer until the peaks are tolerable. At that point it's usually a net improvement, but arguably still far from ideal.
So let's bring in a second subwoofer, and put it in a very different location from the first one. This second subwoofer generates a room interaction peak-and-dip pattern every bit as bad as the first one, but the peaks and dips are at different frequencies! And this difference happens throughout the room. So the net result is a significant smoothing of the frequency response. The only way that adding a second subwoofer could result in response just as bad as the first sub would be, if their in-room response curves were identical. And unless they are in the same place, that will not happen.
In general, the in-room smoothness increases proportionally with the number of subs we have distributed around the room. So two subs will be roughly twice as smooth as one, and four subs will be roughly twice as smooth as two. And this improved smoothness holds up throughout the room, which means that if you still need to EQ to fix a remaining problem, chances are you are fixing a global problem instead of fixing a local one but making things worse elsewhere. (I realize the mix position is by far what matters most, but when the time comes to “impress the customer”, really good bass throughout the room is not going to hurt.)
Now when we measure the in-room response at low frequencies, the peaks and dips will be smaller, more numerous, and bunched up closer together (more like what we'd see in a significantly larger room). The subjective improvement will probably be greater than we would infer from eyeballing our before-and-after curves, for two reasons: The ear/brain system's one-third-octave averaging mechanism can probably come into play; and just as the ear/brain system is disproportionately sensitive to SPL variations south of 100 Hz, it is also disproportionately appreciative of the improvements we will have made.
Let me point out one more implication of all this which is somewhat counter-intuitive: From a perceptual standpoint, it is the trailing edge of the notes (the in-room decay times) that matters most. A little bit of smearing of the initial arrival times from multiple subs at low frequencies doesn't matter because the ear/brain system doesn't react fast enough to detect it. But the ear/brain system is very good at hearing amplitude changes at low frequencies (remember the equal-loudness curves bunching up), so we want to get the amplitude as smooth as we reasonably can. And in doing so, we will also be getting the decay times as even as we reasonably can.
So even though it doesn't make sense at first glance, several small subs intelligently distributed will generally give you better perceived transient response than a single uber-sub.
The good news for those of you who already have a sub is, multiple subs do not need to all be identical. If your current sub goes plenty low enough, you probably add get one or more less expensive subs and still see significant benefits from the smoothing you'll get in the modal region. For those of you who don't have subs but are thinking about getting one or more, if you opt for “more”, you can get away with smaller subs than you might have otherwise been thinking of. And you don't have to buy them all at once.
One final argument for maybe putting something like this on your down-the-road wish-list: Humans are wired to have a positive emotional response to low frequencies moreso than to mid and high frequencies. So if you want to enjoy your incredibly cool jobs even more...
For those of you who took my advice and just skipped to the last paragraph, in general the more subs you have (and set up intelligently), the better the in-room bass will be, both from a frequency response standpoint and from a time-domain standpoint. In effect, you will be using the room's acoustics to work with you instead of against you, both acoustically and psychoacoustically. The net result will be, perceptually you will have pretty much gotten the room out of the way, so you'll hear what's on the recording down low more clearly and therefore make better mix decisions. And when the time comes that you need your system to make a good impression on a customer, better is better.
Having totally failed to master the technical intricacies of digitally quoting on this site, I hope you will forgive me for reverting to analog:
jeromemason wrote: “Just wondering the thoughts of some that either have or had dual subs going...
“I read an article where two guys were pitching using 4 or more subs in a room...”
Briefly (or in retrospect, perhaps not!), here is a look at the situation with a single sub in a home studio room, and then we'll look at how an additional sub (or subs) can make an improvement. Imagine you have one subwoofer in your room. Its interaction with the room boundaries results in a rather drastic peak-and-dip pattern. You can move the sub and change the peak-and-dip pattern, but you cannot eliminate it by positioning alone. You can mitigate it by a few dB with large bass traps. You can equalize it to be fairly smooth in one location, but in the process you will have made it worse elsewhere. And it is possible that the response still will not be sufficiently smooth at both ears. Quoting jazznoise:
“In small rooms, the issue is that room modes can be so big that they can cause different tones to 'sit' left or right. Interaural time delay as a listening mechanic doesn't work well for bass, it's intensity and phase at low frequencies that allow your brain to attribute directionality. If your right sub puts out a 50 Hz tone which nulls to your right but rings to your left, the localization is gonna be messed up.”
In a small room with a single sub, there's a good chance you'll have at least one frequency region where the situation jazznoise describes applies, and EQing that single sub won't fix it. It's an acoustic problem, and arguably is best addressed with an acoustic solution.
At low frequencies, sub + room = an inseparable system, from the standpoint of perception. The ear/brain system literally cannot hear the sub apart from the room. We cannot detect the presence of bass energy from less than one wavelength, and we cannot determine the pitch until we have heard several cycles (wavelengths). Mentally compare your sub's distance to your various room boundaries with how long low-frequency wavelengths are, and you'll see that by the time you actually hear the low frequency energy, the room's effects are already in full swing.
Back to the room-induced peak-and-dip pattern for a moment. Smooth bass is “fast” bass, because the peaks literally take longer to decay. To put it another way, speakers + room = a minimum-phase system at low frequencies, which means that if we smooth the frequency response, we have simultaneously smoothed the time-domain response, and vice-versa. In-room peaks spoil the pitch definition because, as the energy decays over time (and remember the peaks take longer to decay), the spectrum actually shifts in the direction of the peaks.
If these room-induced peaks and dips were very numerous and bunched up close together, it might look terrible to the eye on a graph, but perceptually it would in effect be a “continuum”, because the ear/brain system tends to average them out (and indeed this is what happens with the reflections at midrange and high frequencies – perceptually they merge into a continuum). But if the peaks and dips are further than about 1/3 octave apart, the span is too great for the ear's averaging mechanism to come into play, so they stick out like sore thumbs. One counter-intuitive implication of this is, the problem isn't that we have too many room modes at low frequencies – it's that we have too few! (Bigger rooms have greater model density at low frequencies, which is why they usually have more natural-sounding bass; the smaller the room, the fewer and further apart the peaks and dips in the bass region.)
If we look at a set of equal-loudness curves, we'll see that they bunch up south of 100 Hz or so. This is telling us that a change in SPL in the bass region makes a disproportionately large change in perceived loudness. For example, a 5 dB change in SPL at 50 Hz makes approximately as much of a difference in perceived loudness as a 10 dB change at 1 kHz!
So to recap, we have three problems that can arise from using a single sub in most rooms:
1. Poor pitch definition due to room-induced peaks;
2. The peaks (and to a lesser extent the dips) perceptually stick out like sore thumbs because they are not only large but few and far apart; and
3. The ear/brain system has a heightened sensitivity to SPL changes south of 100 Hz, so the subjective detriment of these peaks and dips is greater than we would infer simply from eyeballing our in-room frequency response curves.
At the risk of oversimplifying, what usually happens in practice is, we turn down the level of the subwoofer until the peaks are tolerable. At that point it's usually a net improvement, but arguably still far from ideal.
So let's bring in a second subwoofer, and put it in a very different location from the first one. This second subwoofer generates a room interaction peak-and-dip pattern every bit as bad as the first one, but the peaks and dips are at different frequencies! And this difference happens throughout the room. So the net result is a significant smoothing of the frequency response. The only way that adding a second subwoofer could result in response just as bad as the first sub would be, if their in-room response curves were identical. And unless they are in the same place, that will not happen.
In general, the in-room smoothness increases proportionally with the number of subs we have distributed around the room. So two subs will be roughly twice as smooth as one, and four subs will be roughly twice as smooth as two. And this improved smoothness holds up throughout the room, which means that if you still need to EQ to fix a remaining problem, chances are you are fixing a global problem instead of fixing a local one but making things worse elsewhere. (I realize the mix position is by far what matters most, but when the time comes to “impress the customer”, really good bass throughout the room is not going to hurt.)
Now when we measure the in-room response at low frequencies, the peaks and dips will be smaller, more numerous, and bunched up closer together (more like what we'd see in a significantly larger room). The subjective improvement will probably be greater than we would infer from eyeballing our before-and-after curves, for two reasons: The ear/brain system's one-third-octave averaging mechanism can probably come into play; and just as the ear/brain system is disproportionately sensitive to SPL variations south of 100 Hz, it is also disproportionately appreciative of the improvements we will have made.
Let me point out one more implication of all this which is somewhat counter-intuitive: From a perceptual standpoint, it is the trailing edge of the notes (the in-room decay times) that matters most. A little bit of smearing of the initial arrival times from multiple subs at low frequencies doesn't matter because the ear/brain system doesn't react fast enough to detect it. But the ear/brain system is very good at hearing amplitude changes at low frequencies (remember the equal-loudness curves bunching up), so we want to get the amplitude as smooth as we reasonably can. And in doing so, we will also be getting the decay times as even as we reasonably can.
So even though it doesn't make sense at first glance, several small subs intelligently distributed will generally give you better perceived transient response than a single uber-sub.
The good news for those of you who already have a sub is, multiple subs do not need to all be identical. If your current sub goes plenty low enough, you probably add get one or more less expensive subs and still see significant benefits from the smoothing you'll get in the modal region. For those of you who don't have subs but are thinking about getting one or more, if you opt for “more”, you can get away with smaller subs than you might have otherwise been thinking of. And you don't have to buy them all at once.
One final argument for maybe putting something like this on your down-the-road wish-list: Humans are wired to have a positive emotional response to low frequencies moreso than to mid and high frequencies. So if you want to enjoy your incredibly cool jobs even more...
For those of you who took my advice and just skipped to the last paragraph, in general the more subs you have (and set up intelligently), the better the in-room bass will be, both from a frequency response standpoint and from a time-domain standpoint. In effect, you will be using the room's acoustics to work with you instead of against you, both acoustically and psychoacoustically. The net result will be, perceptually you will have pretty much gotten the room out of the way, so you'll hear what's on the recording down low more clearly and therefore make better mix decisions. And when the time comes that you need your system to make a good impression on a customer, better is better.