Manually time-aligning subwoofer(s) to mains - how to

[SoundsGood64]

If you set proper time delays on L and R (from #1), then you should already have L and R phase aligned, shouldn't you?

The time delays for main speakers (L, R, center, surrounds) were done on a midrange sweep 300-5000hz. To preserve stereo imaging, it's probably important to maintain those time delays. When I go down to bass region though there's the phase difference at 80hz you can see between L and R in post 217. Time alignment in a wide mid range of 300-5000hz (probably) doesn't mean phase alignment at 80hz. Maybe the room is affecting phase alignment at 80hz.

 

[janbth]

Hi, must disagree. All sound is traveling waves...some just happen to be big enough to cause in-room resonances, but they are still traveling waves.

Well, sound propagation is described by a wave equation with sources, receivers, and boundary conditions. It’s the boundary conditions that are responsible for the room modes, i.e. an equation with boundary conditions where the sources or receivers are set to zero. It is mathematically possible to describe the solutions to this equation as a sum of waves traveling in opposite directions, but one traveling wave on its own is not a solution. You need to sum them, and then you get solutions in the form of standing waves.

Imo, room correction via DSP might be legitimately viable, when and if we take a rooms impulse response, and invert that.

A room’s measured impulse response includes reflected sounds, and is not in general invertible.

 

[janbth]

The time delays for main speakers (L, R, center, surrounds) were done on a midrange sweep 300-5000hz. To preserve stereo imaging, it's probably important to maintain those time delays.

Why do you sweep only 300-5000 Hz, and not full range?

When I go down to bass region though there's the phase difference at 80hz you can see between L and R in post 217. Time alignment in a wide mid range of 300-5000hz (probably) doesn't mean phase alignment at 80hz. Maybe the room is affecting phase alignment at 80hz.

OK, yes, I think this may be plausible, since L and R are located in different positions in the standing wave patterns of the room modes.

 

[DonH56]

Can you recommend one that stores X session results for later retrieval? Not pricey.

I do not know, sorry. Check Amazon? Doubt any cheap ones do; that requires memory and usually comes in more expensive logging meters. Probably easier to write the trials down on paper or in a spreadsheet or whatever. Or use REW. I have a very old Radio Shack SPL meter, and a UMIK-1 for use with REW, plus and Earthworks measurement mic (M30) if I want to drag out a preamp and all the extra stuff (usually not; REW + UMIK-1 does more than my previous $10k acoustic analysis rig and with far less effort -- not as much dynamic range, but more than enough for 99% of what I do with it).

Alternatively, reco software that makes use of my UMIK-2, ideally via Android, but windoze or linux otherwise?

Again, not for Android, sorry. I know there are many apps as well as a few add-ons for phones so hopefully somebody who uses them will chime in. Make sure it goes low enough in frequency; some (many?) phones are optimized for the vocal band so do not pick up bass well. For Windows, REW is free and very powerful, though with a learning curve, but there is lots of online help and it will work with your UMIK-2. That is what I would do.

REW room acoustics and audio device measurement and analysis software

REW is free software for room acoustic measurement, loudspeaker measurement and audio device measurement

www.roomeqwizard.com

 

[Keith_W]

Thanks Keith, but not for me. I'm an old goat who's made too many measurements for too long, and have become a bit cynical as to how well they reflect reality (in the manner we hope). So not trusting measurements all that far, I'm just not willing to peer down deep into them. I used to do that, lord knows I used to....

You are absolutely right. I too am sceptical that some measurements taken by amateurs (i.e. most of us) do not reflect reality. The solution isn't to throw the baby out with the bath water, it's to learn how to take those measurements properly.

 

[gnarly]

You are absolutely right. I too am sceptical that some measurements taken by amateurs (i.e. most of us) do not reflect reality. The solution isn't to throw the baby out with the bath water, it's to learn how to take those measurements properly.

Fully agree, the solution is learning how to do them right.

To that end, I've tried to become proficient with REW, Smaart, ARTA, Crosslite+, and adding LinFIR to the mix.
I've had 8 days of in-person training with Smaart. And currently participate in a weekly international Crosslite+ training video conference.

It's funny though, the more I learn about measurements, the more I realize their limitations. Or to be more precise, the narrowness of their applicability, and their appropriateness to act on.
Kinda wish sometimes I was back in the day when I'd happily apply a FIR filter to any and every quasi-anechoic DIY speaker/driver measurement I made;
or slap Dirac, or some other room-correction program, into my DSP to correct in-room measurements Lol.

 

[Keith_W]

It's funny though, the more I learn about measurements, the more I realize their limitations. Or to be more precise, the narrowness of their applicability, and their appropriateness to act on.

In which case, perhaps you should look at my book and suggest some changes. I can give you editing rights. That book is free, it's for the community, and I have been working on it solo. It would be better if it had more than one editor ... otherwise it reflects all of my biases and that can't possibly be good.

 

[gnarly]

In which case, perhaps you should look at my book and suggest some changes. I can give you editing rights. That book is free, it's for the community, and I have been working on it solo. It would be better if it had more than one editor ... otherwise it reflects all of my biases and that can't possibly be good.

Your book is a very generous gift to the community. Very, Very Nice!

I figure I would not make the best collaborator, as my focus is primarily that of a DIY speaker builder, and secondarily that of a large-space system tuner
Putting that together with no longer emphasizing the stereo illusion, aka sweet spot listening ....I just don't have much interest in fine-tuning small rooms.
Especially given I have a hard time accepting the efficacy of room-tuning, other than to a specific spot (that I don't care about)

And i have to admit I may have become jaded as to the value of measurement optimization, once a certain level of speaker & room performance is met.
Seems akin to pixel peeping in digital photography. I think the time to quit measuring and simply get some tone controls we like, comes a lot sooner than realize, until we've been around the block a few times.

So please let me bow out gracefully from trying to offer suggestions for your book. In exchange i will do my best not to let my specific focus, or doubts about the validity of measurements in small rooms, lead to comments that dampen others' enthusiasm toward DSP room-corrrection.

 

[john61ct]

Personally I think fresh "iconoclastic counterpoints" are good for learners, all opinions welcome.

Not talking about you "collaborating" with Keith, but maybe your own thread laying out your framework rationally

Then when you have contributions and comments in other threads, you can crosslink to give newcomers a context

 

[SoundsGood64]

The REW alignment tool works very well? This is what I do:

1. Take a sub measurement with XO in place with a tweeter or loopback timing ref.
2. Take a sweep of both main speakers (with HPF in place, if any) with the same timing ref.
3. Open the alignment tool, load both graphs, then fiddle with phase alignment until there are as few dips as possible (ignore the peaks). Also pay attention to the impulse graph, you want the sub to the right of the main impulse. You also don't want any delay greater than half a period. If you need more, consider inverting the polarity.
4. Click on the "Make aligned copy" on each measurement, then vector sum them.
4. Dial those settings in and take a verification measurement. Compare it to your vector sum. If it's the same, you have successfully completed the alignment procedure. You may proceed to EQ.

Following this protocol from Keith_W:

The crossover is in place at 80hz with a surround speaker as timing reference. When sweeping L+R (with sub on and crossover active) Inverted polarity gave higher SPL at the crossover frequency of 80hz, so I chose inverted polarity, as shown in figure 'sub_polarity.jpg'.

Here (phase-overlay-7.9ms-delay-LRsub.jpg) is the phase vs. frequency overlay for three separate sweeps: 1) L+R together without sub (blue), 2) subwoofer only with no delay (red), and 3) subwoofer only with 7.9ms delay (green). You can see that the phase of L+R is aligned with the sub when sub is delayed by 7.9ms at the 80hz crossover point.

Also shown is the alignment tool with L+R together without sub (blue) and the subwoofer only with 7.9ms delay (green), both the phase view and the impulse view.

Next, for the verification step, I made 'aligned copy' for L+R together without sub, and separately sub with 7.9ms delay, and then took the vector average (blue). I then took a new measurement of "L+R", which because of the active crossover also plays the subwoofer with 7.9ms delay (red), showing both SPL and unwrapped phase.

It would appear the graph shows the vector average does not align with the measurement at 80hz. I think what this may mean is that when L,R, and sub sweep together (red) there is interaction among the 3 speakers that cannot be accounted for when L+R (no sub) and subwoofer are swept separately and then vector averaged. (Possibly more on that in a future post...) Based on the phase overlay and alignment tool views, it looks like I have successfully aligned L+R and subwoofer with 7.9ms delay, but would like comments here as I'm not sure given potential trouble in this verification step.

 

[levimax]

Following this protocol from Keith_W:

The crossover is in place at 80hz with a surround speaker as timing reference. When sweeping L+R (with sub on and crossover active) Inverted polarity gave higher SPL at the crossover frequency of 80hz, so I chose inverted polarity, as shown in figure 'sub_polarity.jpg'.
View attachment 550243


Here (phase-overlay-7.9ms-delay-LRsub.jpg) is the phase vs. frequency overlay for three separate sweeps: 1) L+R together without sub (blue), 2) subwoofer only with no delay (red), and 3) subwoofer only with 7.9ms delay (green). You can see that the phase of L+R is aligned with the sub when sub is delayed by 7.9ms at the 80hz crossover point.
View attachment 550244

Also shown is the alignment tool with L+R together without sub (blue) and the subwoofer only with 7.9ms delay (green), both the phase view and the impulse view.
View attachment 550245

View attachment 550246


Next, for the verification step, I made 'aligned copy' for L+R together without sub, and separately sub with 7.9ms delay, and then took the vector average (blue). I then took a new measurement of "L+R", which because of the active crossover also plays the subwoofer with 7.9ms delay (red), showing both SPL and unwrapped phase.View attachment 550248

View attachment 550249

It would appear the graph shows the vector average does not align with the measurement at 80hz. I think what this may mean is that when L,R, and sub sweep together (red) there is interaction among the 3 speakers that cannot be accounted for when L+R (no sub) and subwoofer are swept separately and then vector averaged. (Possibly more on that in a future post...) Based on the phase overlay and alignment tool views, it looks like I have successfully aligned L+R and subwoofer with 7.9ms delay, but would like comments here as I'm not sure given potential trouble in this verification step.

I am a little surprized you would be delaying the subs relative to the mains unless the subs are much closer to the LP than the mains. Most subs have DSP delay of several ms or more and large sub drivers tend to lag a little bit as well so normally you end up delaying the mains realative to the subs.

I have had nothing but trouble with the REW alignment tool below 200 Hz, so if I had conflicting measurements I would discard what the alignment tool says, I am especially skeptical of a need to invert the subs if everything is wired correctly.

I would confirm there is a good reason the subs need to be delayed relative to the mains and go from there. Things can appear phase aligned when they are in multiples of 360 degrees out of phase.

 

[Keith_W]

I forgot to say: REW's timing measurement can be UNRELIABLE (I think I said that in my book) particularly if the signal-noise ratio is poor, or if the subwoofer measurement is affected by room modes. I recommend repeating the measurement 3-4 times to make sure it is delivering consistent results. Discard any measurement which is "off". To be honest, I don't like REW's timing measurement method at all. I prefer another method, but that is not possible in REW.

 

[john61ct]

Can you name the tool(s)?

 

[90sNintendoKid]

I forgot to say: REW's timing measurement can be UNRELIABLE (I think I said that in my book) particularly if the signal-noise ratio is poor, or if the subwoofer measurement is affected by room modes. I recommend repeating the measurement 3-4 times to make sure it is delivering consistent results. Discard any measurement which is "off". To be honest, I don't like REW's timing measurement method at all. I prefer another method, but that is not possible in REW.

I’m guessing a loop back timing reference.

 

[Keith_W]

I’m guessing a loop back timing reference.

It's actually detection of the subwoofer impulse which is unreliable. So it does not matter which timing reference you use. Think about it: on the impulse response, the subwoofer impulse is time stretched and low volume. The algorithm detects the impulse from a certain % of deviation from the baseline: and if the SNR is poor, it may be indistinguishable from a wobbly baseline caused by room background noise. Only human eyes can reliably spot the subwoofer impulse. And since you have a brain, you can decide to reject the measurement and do it again.

Some time ago I asked John Mulcahy why I can't manually measure the distance between the tweeter timing chirp and the subwoofer impulse to manually determine the delay. Actually I did measure it, but it made no sense. Manually measuring the delay did not correlate with REW's automatic delay measurement at all.

Anyway, this is why I say that we need to repeat timing measurements a few times to ensure consistency. It's very simple, get REW to take 5 measurements, then walk away and have a snack. Come back and look at them - if they are all the same, you have taken a valid timing measurement. If there are outliers, reject them. If none of them are the same ... well, you have a problem!

 

[OCA]

Following this protocol from Keith_W:

The crossover is in place at 80hz with a surround speaker as timing reference. When sweeping L+R (with sub on and crossover active) Inverted polarity gave higher SPL at the crossover frequency of 80hz, so I chose inverted polarity, as shown in figure 'sub_polarity.jpg'.
View attachment 550243


Here (phase-overlay-7.9ms-delay-LRsub.jpg) is the phase vs. frequency overlay for three separate sweeps: 1) L+R together without sub (blue), 2) subwoofer only with no delay (red), and 3) subwoofer only with 7.9ms delay (green). You can see that the phase of L+R is aligned with the sub when sub is delayed by 7.9ms at the 80hz crossover point.
View attachment 550244

Also shown is the alignment tool with L+R together without sub (blue) and the subwoofer only with 7.9ms delay (green), both the phase view and the impulse view.
View attachment 550245

View attachment 550246


Next, for the verification step, I made 'aligned copy' for L+R together without sub, and separately sub with 7.9ms delay, and then took the vector average (blue). I then took a new measurement of "L+R", which because of the active crossover also plays the subwoofer with 7.9ms delay (red), showing both SPL and unwrapped phase.View attachment 550248

View attachment 550249

It would appear the graph shows the vector average does not align with the measurement at 80hz. I think what this may mean is that when L,R, and sub sweep together (red) there is interaction among the 3 speakers that cannot be accounted for when L+R (no sub) and subwoofer are swept separately and then vector averaged. (Possibly more on that in a future post...) Based on the phase overlay and alignment tool views, it looks like I have successfully aligned L+R and subwoofer with 7.9ms delay, but would like comments here as I'm not sure given potential trouble in this verification step.

Correct summation is L + Sub + R + Sub, vector average will count sub only once.
Always generate EP of the measurement, Remove IR delays from the EP and only then compare phase graphs
If you use "Align IRs at cursor" instead of "Align phase at cursor", REW's alignment tool will automatically give you correct sub polarity. Phase alignment cannot invert.

 

[OCA]

Manually measuring the delay did not correlate with REW's automatic delay measurement at all.

For two reasons:

First, the subwoofer’s impulse peak occurs around the frequency where its LPF is set - typically around 250 Hz whereas the speaker’s impulse peak represents the arrival of its highest frequency content (24 kHz). It makes no sense to align the subwoofer’s 250 Hz arrival to the speaker’s 24 kHz tweeter arrival. What you actually want to align is the subwoofer with the speaker’s woofer, which is much closer to aligning their actual IR start times.

Second, REW’s timing is dead accurate, and the relative impulse responses of the sub and speaker will be displayed correctly in the Overlays impulse graphs. The problem is that removing the IR delay i.e. determining the actual IR start for a subwoofer can only be based on detecting the transition from the noise floor. That can be extremely difficult with most subs that have a wobbly start or pre-ringing.

 

[janbth]

It's actually detection of the subwoofer impulse which is unreliable. So it does not matter which timing reference you use. Think about it: on the impulse response, the subwoofer impulse is time stretched and low volume. The algorithm detects the impulse from a certain % of deviation from the baseline: and if the SNR is poor, it may be indistinguishable from a wobbly baseline caused by room background noise. Only human eyes can reliably spot the subwoofer impulse. And since you have a brain, you can decide to reject the measurement and do it again.

This is wrong. This is not how REW uses a timing reference. It doesn’t matter if you measure a subwoofer or anything else, if there is an acoustic timing reference, REW will use that to time align measurements. After all, that is the purpose of timing references. I think you are mixing up this with the «Set t=0 at IR start» method.

From the REW documentation:

• A Timing reference can be selected for the measurement, see Measuring with a timing reference for more on the settings when using a timing reference. When not using a timing reference there are 3 choices for the t=0 position of the measured impulse response: (1) Set t=0 at IR peak will place t=0 at the exact position of the peak of the IR, determined by interpolating an oversampled version of the IR; (2) Set t=0 at IR start will place the t=0 point at the sample before the IR first exceeds 10% of the peak level; (3) Set t=0 from delay estimate will use cross-correlation with a minimum phase version of the impulse to estimate and remove any time delay.

Link: https://www.roomeqwizard.com/help/help/html/makingmeasurements.html#top

Acoustic timing in REW is accurate and reliable if the reference speaker has a tweeter. More from the REW documentation:

Acoustic timing reference​

If an acoustic timing reference is used REW will generate a timing signal on the output that has been selected to act as the reference before it generates measurement sweeps on the channels being measured. The timing signal is a sweep from 5 kHz to 20 kHz lasting about 700 ms. It must be directed to a speaker that can reproduce high frequencies, one that has a tweeter. The level of the timing signal is set relative to the measurement level using the Ref level trim control, it may need to be higher or lower depending on whether the speaker used as the timing reference is further away or closer than the speaker being measured. The timing signal is a high frequency sweep to allow accurate timing, a subwoofer cannot be used as the reference channel. Measurements will have a time delay that corresponds to the difference in their distance from the microphone compared to the distance of the reference speaker - if the reference speaker is further away the delay would be negative. The delay can be shifted using the Timing offset. When an acoustic timing reference is used individual measurements taken from the same mic position will have the same relative timing, allowing trace arithmetic to be carried out on the traces in the All SPL graph. Note that multiple sweeps cannot be used when using an acoustic timing reference.

Same link as above.

 

[JStewart]

wrt subwoofer to woofer/midwoofer acoustic alignment, wouldn’t a comparison of post adjustment predicted vs actual response answer any questions about accuracy?
If they don’t match then go find the problem?

 

[Keith_W]

Here are 3 measurements of my right sub, all with tweeter timing reference. The measurements were taken back to back, and the mic was not moved.

And here are the delays reported by REW. There is a +/- 0.3ms discrepancy between all 3 measurements, despite all of them being correctly aligned on the impulse response.

If we drop a marker at the 7ms mark, where REW has indicated that the subwoofer impulse starts, we can see the problem. Each measurement has a tiny discrepancy in SPL, which is throwing off REW's automatic delay reporting by +/- 0.3ms.