subwoofer amplifier 'rules' and phase

[terry j]

hi all

new to the game, seriously thinking of an LLT but that's for another thread.

Often when reading about subs and integrating them with the mains etc, I come across the importance of phase and it's ability to help integrate nicely with the mains. (normal or reverse, or if fancy some varying degree). This of course is usually related to plate amps, which also have a crossover control.

Then I come across the use of high powered pro amps to use on the sub, of whatever ilk, which naturally have no means of phase control or means of changing the xover point.

Simply put, can someone explain the importance or otherwise of phase - or anything else for that matter - in the integration of the sub with the mains??

lots of love

terry

[Otto]

Hi terry,

Well, in general, we want things to be in phase such that we don't cancel out frequencies at crossover points. Signals are 180 degrees out of phase if Signal A and Signal B are at opposite ends of their amplitude range at the same point in time. Consider a sine wave that alternates at frequency = f from 1 to -1 "units". If Signal A is at +1 when Signal B is at -1, they are 180 out of phase. If everything were perfect, they would theoretically cancel one another out, such that there was no sound.

If Signal A is at +1 and at the same time Signal B is at +1, they are in phase (0 degrees out of phase, or 0 degree phase shift). They will add to each other, and if everything was physically perfect, there would be a 3 dB increase in SPL (as opposed to there being only Signal A or Signal B).

Find a piece of scrap paper and draw out a sine wave -- any sloppy old rendition will do. Then draw another one just below it and "add" them -- either on the paper or in your head. Then draw one and below it draw one that's just the opposite of it. "Add" them and see how they cancel. The picture really helps.


So we've considered signals that are 180 degrees out of phase and signals that are 0 degrees out of phase. Because phase can be altered by other crossovers, distance between speakers, speaker placement, room effects and whatever else, it's not likely to have perfect 0 or 180 out of phase conditions. We can also have an infinite number of phase shifts between these two examples. So a 0/180 degree phase switch is one step toward resolving this problem, but for best results a fully variable phase control is much better.

If the two signals are out of phase, they can be detrimental to one another. If they are in phase, the transition between the sub to the mains will be smooth, rather than having a dip or other filtering effect. I believe you'd be able to see this in an REW plot.

Now, I've been lazy in playing with phase, because, as you point out, the big pro amps (like my Behringer EP2500) don't have any way to adjust phase. I'm also in the midst of finding some new speakers, so I won't make any effort to do so until that is complete. There are pre-built devices out there that will adjust phase, and there is DIY stuff out there as well. I have doing a DIY external phase control for my system.

I've also been kinda lazy on really analyzing the transition between the sub and the mains. At this point I certainly don't know the status of sub/main phase at the crossover frequency (60 Hz). I've been thinking lately that I might have a little too much low frequency content (<25 Hz?) and that I should tweak with my BFD. Perhaps I should measure phase at the same time. If I can borrow an oscilloscope from work, I believe I would be able to measure phase using a sine wave from REW.

Anyway, that's it for now. Lemme know what you think...

[Hakka]

Otto wrote:

Now, I've been lazy in playing with phase, because, as you point out, the big pro amps (like my Behringer EP2500) don't have any way to adjust phase.

I believe you can use the subwoofer distance setting in the receiver/prepro to adjust phase in a setup like this.

[SteveCallas]

Correct, you can adjust phase digitally with distance settings. Start with the actual distance you are from your sub, then play with it until the FR of your crossover range measured at the seat ends up with the flattest response.

[terry j]

thanks all,

understand about cancellation etc of the waves, and can see that the major factor would be the varying distances from set up to setup, and the plate amp manufacturer has no control over that and must therefore provide adjustment for it, unlike in say the mains, wherein the relative postions of the drivers are fixed.

I suppose theoretically phase probs can exist between the mains, ie different frequencies will reinforce or cancel as a function of the distance between the two.

Guess that's what 'threw' me, the first time I'd come across phase control when using pro amps was Ottos answer above, but that is more a function of my newness than anything else.

Just so I fully get it conceptually, the easiest way to see this is with something like REW, and look for the smoothest response.

One step further, if say I only had normal or reverse, further fine tuning could then be achieved by moving the sub a bit, and observing the change in response? This could conceivably alter the gain of the sub ( further from the corner for example), which would show as a smoother ( or not ) graph, but with the sub portion of the graph higher ( or lower ) than the mains, and then twiddle with the gain knob. Hope that was clear.

Due to the frequencies involved, if indeed the sub moving can function as a pseudo phase control, we wouldn't really be talking an inch or so making a difference would we???

Anyways, thanks Otto and all, I can see it quite a bit better now.

lots of love

terry

ps just re-read Steves post, re changing the 'distance' to the sub. I guess I missed that in my rave above, cause if I move the sub ( and not change the signal) that's the same as not moving the sub ( but changing the signal distance)

In both those cases would not we hear the bass as either delayed or advanced, or are the 'innaccuracies' so small we wouldn't hear them.

[Otto]

Hey terry,

terry j wrote:

thanks all,
understand about cancellation etc of the waves

Yeah, I wasn't sure exactly sure where to start, so I started with the basics.

Just so I fully get it conceptually, the easiest way to see this is with something like REW, and look for the smoothest response.

I think it'll show up, but I can't say first-hand how apparent it will be. Still need to play with that myself. It may be useful to use pure tones at the crossover point to "narrow in" on the area in question once your sweeps start looking pretty good.

One step further, if say I only had normal or reverse, further fine tuning could then be achieved by moving the sub a bit, and observing the change in response? ...

Due to the frequencies involved, if indeed the sub moving can function as a pseudo phase control, we wouldn't really be talking an inch or so making a difference would we???

That's right. If I did all the math right, a 50 Hz tone will have a wavelength of ~21.1 ft. So, yeah, moving inches may not really do a whole lot.

ps just re-read Steves post, re changing the 'distance' to the sub. I guess I missed that in my rave above, cause if I move the sub ( and not change the signal) that's the same as not moving the sub ( but changing the signal distance)

That's right. I wasn't even thinking of using the distance setting to change phase, but that's right. So what happens there when the distance from the sub to the listener is greater than the distance from mains to listener? The preamp delays the signal coming from the mains such that sub "starts in" a little earlier, so that its sound output hits you at the same time as the mains' output? One way or another, it's playing with the time at which the signal "starts" on a per-speaker (per-channel) basis. In my book, this is pretty much the definition of phase.

Now, I think it's possible to have phase contributions from "other things" like digital and passive filters. Assuming that, it's possible that things are somewhat out of phase even before considering placement. That is, even if placement between mains and sub were coincident, they would still be out of phase from the start. Indeed, you could still play with distance to adjust phase, but it becomes more confusing in that the "distance" value on your screen doen't really mean "distance" anymore. We would need to convert "distance" into "degrees offset". At that point, it's probably just easier to do it empirically, especially since most of us aren't going to be able to measure phase at the listening position (or does REW have that tucked in somewhere? Could I use the impulse response measurement somehow?)

In both those cases would not we hear the bass as either delayed or advanced, or are the 'innaccuracies' so small we wouldn't hear them.

Let's consider a 50 Hz tone. Again, if I did the math right, this has a period of 20 mS and a wavelength of 20.1 ft. Now assume that your mains and sub are totally in phase when placed next to each other. But most of us don't stack our mains and subs, and the sub usually ends up in the corner for a number of reasons. Now suppose that your listening position is 10 ft from your mains and 15 ft from your sub. All else being equal (no distance adjustment, assuming no other phase problem contributions from filters, etc.), your mains and sub are out of phase. The sub signal will "lag" the main signal. Like you are discussing, there are (at least) two ways to fix this -- move the sub or play with distance.

If we move the sub, and we have no other controls to adjust phase, we will have to move it either back to the main position, or we would have to move it further away from the mains, for a total distance of 21.1 ft. And now you have a 20 mS delay between your mains and sub, but you have phase coherence. I don't know how the human brain will detect 20 mS at that frequency, but that's what you'd have.

Now, if we leave the mains at 10 ft and leave the sub at 15 ft, and play with distance in the preamp's control screen, I believe that the preamp will intentionally lag the mains' signal. This leads to the sub generating its output ~4.7 mS before the mains. In this case, there should be no "lead" or "lag" at the listening position. Yeah, the preamp delayed the start of the mains signal, but you don't know that, and it doesn't really matter.

OK OK OK. I quit for now! That's too much typing for this early in the morning (can't sleep at 4am? Come to the Shack!).

[brucek]

Then I come across the use of high powered pro amps to use on the sub, of whatever ilk, which naturally have no means of phase control

Just swap the speaker leads on the driver. It will be 180 degrees out of phase...

brucek

[---k---]

Wow Otto, lots of good stuff there.

Another way to adjust the phase on your processor to get the good phase aligment is using the Avia disc. The have some wide band pink noise tests specifically for this. You play the pink noise, which will be sent to the mains and the sub. Then as this is playing, you adjust your sub distance until you get the maximum output, as read on your meter. Theoritically, at the max output all the waves will be lined up and matched, and at any lesser output there is some cancellation. It isn't perfect, but is quick and easy.

[Ayreonaut]

Otto wrote:

Now, I think it's possible to have phase contributions from "other things" like digital and passive filters. Assuming that, it's possible that things are somewhat out of phase even before considering placement. That is, even if placement between mains and sub were coincident, they would still be out of phase from the start. Indeed, you could still play with distance to adjust phase, but it becomes more confusing in that the "distance" value on your screen doen't really mean "distance" anymore. We would need to convert "distance" into "degrees offset"...

Its not exactly degress of phase, its straight delay.

And you're right that some components induce a delay which can be compensated for.

Sonnie's BFD Guide wrote:

NOTE: Something that you should remember to do when you are setting up your BFD is add a foot to the value you enter for sub distance in your pre/pro or receiver set up. The 1 msec DSP processing delay in the BFD would account for approximately a foot in distance.

[Otto]

Ayreonaut wrote:

Its not exactly degress of phase, its straight delay.

Right, but aren't delay and phase really the same thing? If one signal lags another, they are, by definition, out of phase. If I know a wavelength and how long the signal is delayed with respect to another, I can calculate phase.

And you're right that some components induce a delay which can be compensated for.

The processing time incurs a delay, but digital and analog filtering will also cause phase shifts (filters that are steeper than 1st order, IIRC). I do believe that, in addition to the actual processing time involved when using the BFD as an EQ, we are also changing the phase of the original signal. So if the processing time were 1/4 of a given tone's wavelength, we would be 90 degrees out of phase -- I would call it "output lagging input". In addition, there would be a phase contribution from the filtering itself, such that we will now be 90+x degrees out of phase with the input signal, where x is non-zero and the phase contribution of the filtering.

Same thing at your crossover point in your mains. From mid to tweeter, there's some crossover network, and that circuit can (will) alter the phase at that crossover point. Some believe that this phenomenon is an attribute to be considered when listening to speakers. This phenomenon will occur wherever you have crossover points that are higher than first order.

Now, it might be possible to adjust phase in the digital realm to account for whatever phase shift the processing might have incurred. I'm not really much of a DSP guy, but something like that wouldn't surprise me.

[Otto]

Woke up this morning thinking about this stuff.

I guess I should add that, when applying filters (analog or digital), phase can do strange things when the math becomes complex (i.e., has an imaginary component). The phase shift will change over the frequency range. Like I alluded to in my previous post, there may well be ways to combat this, especially in the digital realm, but I certainly don't have any practice with it.

I still believe that a time delay or distance change will create a phase shift. This is a simple case, but will still cause different phase relationships at different frequencies. Those phase shifts are different that the ones caused by filtering. I just wanted to say that I know that there are some things I'm glossing over or ignoring, and I'd have to go review some textbooks to better understand it.

[Owen Bartley]

I was wondering the same thing a while ago about how to adjust phase on pro amps, and I remember someone bringing up the distance setting. I have a plate amp with variable phase now, so I thought I was all good, and I remember tweaking the phase to get the best output when I set up my sub a couple years ago. Then I read (along with the other posts here) this:

SteveCallas wrote:

Start with the actual distance you are from your sub, then play with it until the FR of your crossover range measured at the seat ends up with the flattest response.

And I realized that my tweaking was probably at best uneffective, and at worst, detrimental. For one thing, I didn't do it at the crossover point, I just put on a CD with tones from probably around 30 - 40Hz and tried to dial it in to the loudest. For another, I did this from about 1 foot away from the sub, while I was leaning over it into the corner so I could reach the dial. So it won't be great at the listening position about 8 or 10 feet away. Finally, did I even have the mains on at the time? I must have... I think... who knows. Looks like it's time to retune. At least I have a project for Saturday. And it might even cure the boominess I've been noticing too.

[brucek]

Right, but aren't delay and phase really the same thing?

I guess you could argue that a wholesale digital delay in time is hardly the same as a phase shift, but phase shift controls do end up as close approximations.

Consider a 10Hz signal that has a calculated period of 100ms versus an 80Hz signal that has a period of 12.5 ms.
If we delay each of these signal by 1msec (1foot), then the phase of each signal is preserved and delayed the same amount and remains a constant, but if I introduce a constant analog phase shift by passing the signals through a filter, the end result is quite different. A degree of phase shift for the 10Hz signal is different than a degree of shift for the 80 Hz signal. But that isn't what phase controls do.

Most proper phase adjustment circuits are of a second order 'all pass filter' design with a low Q, and over the designed frequency range result in a linear phase response, such that they closely mimic a pure time delay. I did a little reading about all pass filters and I think you can safely use time delay as a replacement for phase if you have no control.

brucek