• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Subwoofer Time Alignment

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,085
Likes
738
Location
Canada
Uh, we don't seem to understand each other. Low pass filter at channel 4 should also be bypassed:

View attachment 77474

Instead of it you should be using XO point at 80Hz set on the sub here:

View attachment 77476

Please do that an than run 0-20khz sweep on the left and right channel without timing reference.
Yes, I've realized that and had already made changes, but I discovered the Left monitor DSP setttings thing... so. Yeah.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
I know what the discrepancy now is, QMuse. The left monitor's internal 60Hz 12dB/oct HP has been enabled all this while!:facepalm:

Must have forgotten to turn it off when doing tests some days before... I'll get back to this. Will just have to confirm other things!

*60Hz
Ah ok, that explains some things. I suggest you run a check on both mains DSP to make sure all such things are disabled.

What actually worries me is difference in the shape of the IR graphs in the 30-150Hz sweeps between mains and sub. I must say I was expecting them to be much more similar. I think we should repeat the measurements for mains and sub in the 50-110Hz range with no XO applied to the mains (either in miniDSP or in mains DSP) and subs XO set to max frequency.
 
Last edited:

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
Something I don't know how to explain though is the ff:

View attachment 77467

The step for the sub in the drivers summation looks as though it has moved forward in time when all settings are, in fact, the same, only xo has been changed.
XO settings affect phase hence they are visible in step response. Red one looks better but sub part still needs to be shifted app 4ms to the left to avoid having that strange peak at 4ms.

Since your sub has a simple passive XO which cannot be bypassed the idea how to do it is simple: use your sub XO to set low pass for the sub at 80Hz and apply high pass at 80Hz in miniDSP along with appropriate delay to the mains. Any other filtering in miniDSP and in DSP of your speakers hsould be turned off. I suspect the delay of 13ms is wrong and that some filters are still active and that is why we are seeing this peak at 4ms and misalignement of sub part of the step.

Let's try with 10ms delay of mains delay and inverted sub channel, but be sure only high pass of mains is active in miniDSP, 80Hz is set at sub and everything else is turned off in miniDSP and in main's DSP.
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,085
Likes
738
Location
Canada
Much appreciate the help!

Besides forgetting to switch off the left monitor's internal 60Hz HP filter, the sub's internal phase/delay was actually set to MAX.

I mentioned before how I had to reach out blindly (due to placement) to change settings behind the sub... all the while it was the other way around.

Moreover, I just realized that I cannot use the sub's internal crossovers so I have to bypass it is as much as possible. Why? Well, I use this single sub for four other setups -- 2 multi-channel and 2 stereo-only modes for the sit-stand desk that is situated directly above the sub. I know, it's all more complicated than most people would want to bother with... setting the sub's delay back to zero, however, instead of max is okay -- should not take too much work in recalibrating the desk setups aforementioned.

Now, making sure everything was set at baseline, and disabling EQ around the xo region I performed REW's delay calculation as I described previously in the ff. posts:

https://www.audiosciencereview.com/forum/index.php?threads/subwoofer-time-alignment.9575/post-475013

https://www.audiosciencereview.com/forum/index.php?threads/subwoofer-time-alignment.9575/post-477157

REW said anywhere between 13.6 and 13.8 ms between the left and right channels, so I chose 13.7ms as the average.

1597147858044.png


1597147870541.png


Looks close enough, and no big suck-out around the xo region. Phase of just the subs and speakers, GD, etc. all look okay.

Making my life easy, I think I'm going to take that as it is... and only add a few linearizing PEQs around the xo region.

*I kept the delays for channels 1 and 2 in miniDSP constant at 13.7ms for the measurements above (mdat attachment). Channel 4 (subwoofer) was always 0ms. No phase/polarity inversion -- doing so made things worse.
 

Attachments

Last edited:

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
Much appreciate the help!

Besides forgetting to switch off the left monitor's internal 60Hz HP filter, the sub's internal phase/delay was actually set to MAX.

I mentioned before how I had to reach out blindly (due to placement) to change settings behind the sub... all the while it was the other way around.

Moreover, I just realized that I cannot use the sub's internal crossovers so I have to bypass it is as much as possible. Why? Well, I use this single sub for four other setups -- 2 multi-channel and 2 stereo-only modes for the sit-stand desk that is situated directly above the sub. I know, it's all more complicated than most people would want to bother with... setting the sub's delay back to zero, however, instead of max is okay -- should not take too much work in recalibrating the desk setups aforementioned.

Now, making sure everything was set at baseline, and disabling EQ around the xo region I performed REW's delay calculation as I described previously in the ff. posts:

https://www.audiosciencereview.com/forum/index.php?threads/subwoofer-time-alignment.9575/post-475013

https://www.audiosciencereview.com/forum/index.php?threads/subwoofer-time-alignment.9575/post-477157

REW said anywhere between 13.6 and 13.8 ms between the left and right channels, so I chose 13.7ms as the average.

View attachment 77561

View attachment 77562

Looks close enough, and no big suck-out around the xo region. Phase of just the subs and speakers, GD, etc. all look okay.

Making my life easy, I think I'm going to take that as it is... and only add a few linearizing PEQs around the xo region.

*I kept the delays for channels 1 and 2 in miniDSP constant at 13.7ms for the measurements above (mdat attachment). Channel 4 (subwoofer) was always 0ms. No phase/polarity inversion -- doing so made things worse.
Well, hate to dissapoint you, but it doesn't actually look ok. Judging by the first graph IR of your mains is inverted in polarity and it peaks at app 8ms. Judging by the 2nd graph impulse peak has shifted to app 23ms with positive polarity. That would imply you need to invert your mains and add app 15ms to them. It would be interesting if you can repeat the measurements with that variant before calling it a day. :)
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,085
Likes
738
Location
Canada
Goodness, what time is it there? Didn't expect you to reply so quick. :p

I think I understand what you mean by the impulse for the mains appearing inverted in relation to the sub. I have indeed noticed this quite a while back, and it is the complete opposite of what I have with my KH120s. This did have me wondering for quite a bit, but I got over that quickly -- I find other things more important. My best guess is it may have something to do with the cable wiring or the RCA to XLR adapters I've been using which might have inadvertently caused the inversion.

The issue is, inverting things so that they look correct 'impulse-wise' actually causes cancellation in the FR. That is the reason why I've kept it as it is... also it may seem as though in my last post I seem to rely too much on REW's calculated delays, but I actually don't. I always prefer to look at the phase relationship between the sub and mains and make sure they closely match as much as possible around the xo area while adjusting delays. Also, in this particular couch/sofa setup, I also find that I need the xo here to be rather steep (48db/oct).

Anyhows, I've changed my mind about REW's calculated 13.7ms average delay suggestion -- yet again -- and moved it to 18ms. Simply put, the phase is now much closer aligned, and I get about the least amount of cancellation, if any.

Whatever the cause of the inversion, the audible effect is likely negligible -- in fact, I have never heard anything unusual in actual musical material -- it's not as if only one of the L-R channels has the wrong polarity, causing 'phasiness' in the sound. BUT, the huge cancellation that appears in the amplitude of the FR when I invert the polarity, though, that is definitely an audible issue. My priority is the latter.
 
Last edited:

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
Goodness, what time is it there? Didn't expect you to reply so quick. :p

I think I understand what you mean by the impulse for the mains appearing inverted in relation to the sub. I have indeed noticed this quite a while back, and it is the complete opposite of what I have with my KH120s. This did have me wondering for quite a bit, but I got over that quickly -- I find other things more important. My best guess is it may have something to do with the cable wiring or the RCA to XLR adapters I've been using which might have inadvertently caused the inversion.

The issue is, inverting things so that they look correct 'impulse-wise' actually causes cancellation in the FR. That is the reason why I've kept it as it is... also it may seem as though in my last post I seem to rely too much on REW's calculated delays, but I actually don't. I always prefer to look at the phase relationship between the sub and mains and make sure they closely match as much as possible around the xo area while adjusting delays. Also, in this particular couch/sofa setup. I also find that I need the xo here to be rather steep (48db/oct).

Anyhows, I've changed my mind about REW's calculated 13.7ms average delay suggestion -- yet again -- and moved it to 18ms. Simply put, the phase is now much closer aligned, and I get about the least amount of cancellation, if any.

Whatever the cause of the inversion, the audible effect is likely negligible -- in fact, I have never heard anything unusual in actual musical material -- it's not as if only one of the L-R channels has the wrong polarity, casusing 'phasiness' in the sound. BUT, the huge cancellation that appears in the amplitude of the FR when I invert the polarity, though, that is definitely an audible issue. My priority is the latter.
If it is not too much hussle for you I would kindly ask you to invert ploarity of the mains and set mains delay to 28ms and re-measure the L+R+sub graph you posted above, just for the sake of the experiment. I am interested to see IR and step response. :)

You can be sure that if you manage to get the timing right FR cancellation will also be at minimum. The opposite is not true.

P.S. My country, Croatia, is in GMT+2 time zone, so it's 16:53 as I'm writing this. :)
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
if you want me to create a phase correction filter in DRC for you to try out send me a full range mdat with a delay where the excess group delay is closest to 0 in that 42Hz-ish range
There is absolutely no point in correcting the phase before getting the time alignement between mains and sub right.

Besides that, correcting sub's phase in the middle of it's working band (around 42Hz) will only introduce pre-ringing (which is audible) on account on reducing GD (which is not audible).

The point of time aligning the sub is to get main's and sub phase as close as possible around XO point, as @KSTR pointed out several times. Sub's phase doesn't need correction below XO. Main's phase above XO may need to be corrected only if there is a phase cancellation between left and right main due to different positioning, but this has to be done very gently to avoid pre-ringing.

Btw, from what I have seen mains woofer and tweeter are time aligned by internal DSP as their step response is textbook perfect.
 

dasdoing

Addicted to Fun and Learning
Joined
May 20, 2020
Messages
844
Likes
427
Location
Salvador-Bahia-Brasil
There is absolutely no point in correcting the phase before getting the time alignement between mains and sub right.

Besides that, correcting sub's phase in the middle of it's working band (around 42Hz) will only introduce pre-ringing (which is audible) on account on reducing GD (which is not audible).

The point of time aligning the sub is to get main's and sub phase as close as possible around XO point, as @KSTR pointed out several times. Sub's phase doesn't need correction below XO. Main's phase above XO may need to be corrected only if there is a phase cancellation between left and right main due to different positioning, but this has to be done very gently to avoid pre-ringing.

Btw, from what I have seen mains woofer and tweeter are time aligned by internal DSP as their step response is textbook perfect.
I hope he sends me my request to prove my concept is right.
anyways, what do you think sounds better, 10% preringing of 2ms, or 100% delay of 15ms of all the bass? the former you wont even realize. if you maintain excess group delay above the zero line you wont have audible pre-ringing because the peak is not coming too early.
have you tryed out my sound examples? noone seams to have noticed the pre-ringing https://www.audiosciencereview.com/...xess-group-delay-corrected.14489/#post-445969

also, it's impossible to align step response at listening position without visable(!) preringing. look at the corrected step response in the link I gave you. you can't get a better step response then that. the waves afterwards are reflections. also, notice that this "requires" preringing. it's just the fact that the bass takes longer to come to peak
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
I hope he sends me my request to prove my concept is right.
anyways, what do you think sounds better, 10% preringing of 2ms, or 100% delay of 15ms of all the bass? the former you wont even realize. if you maintain excess group delay above the zero line you wont have audible pre-ringing because the peak is not coming too early.
have you tryed out my sound examples? noone seams to have noticed the pre-ringing https://www.audiosciencereview.com/...xess-group-delay-corrected.14489/#post-445969

also, it's impossible to align step response at listening position without visable(!) preringing. look at the corrected step response in the link I gave you. you can't get a better step response then that. the waves afterwards are reflections. also, notice that this "requires" preringing. it's just the fact that the bass takes longer to come to peak
Time alignement of the sub and main is important as with adding the sub you are basically adding a driver to the mains (well, sort of). This will still not result in a perfect step response because of the phase error that will be introduced with LR24 crossover at 80Hz between sub and mains. If that phase error is corrected (and this can easilly be done with rePhase) you will end up with perfect step response.

But correcting sub's phase and GD around 40Hz, and GD is naturally raising as frequency is falling toward 20Hz, is something that should never be done. All subs are showing this effect which is totally inaudible and trying to correct it may only result in audible pre-ringing.

There is a reason why automatic EQ systems have pre-ringing compensation and I would never advise users to try to cheat them as things will only get worse.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
10,845
Likes
10,333
Location
Riverview FL
Wouldn't step response be more useful than band limited Impulse for time alignment?

As frequency decreases, the peaks (impulse time?) comes later and later:

1597168428743.png
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,085
Likes
738
Location
Canada
It may very well be that I am just making an error in the way I've been taking measurements or showing them -- perhaps I've simply not been doing it properly/clearly from the start. One thing is certain though, and that is getting things 'right' is proving to be too much of a time-sink.

Still, I think the way the drivers are positioned and how they interact with the room etc, creates more complications than otherwise would have been...

I can't repost the same IR and step graphs with said settings you suggested @QMuse since I already made some changes and cannot quite remember how to backtrack to that exact point.

The main thing that concerns me, anyway, is satisfied:

30-150Hz
no other PEQs other than xo shaping filters
1597168878258.png


If it's of any interest below are the impulse and step of the same:
1597169043901.png


1597169049047.png


The left is pretty much similar. Summed both L+R with Sub not too different, as well.

I may revisit this topic again, but the music and show/movie backlog I have is growing as we speak!
 
Last edited:

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
Wouldn't step response be more useful than band limited Impulse for time alignment?

As frequency decreases, the peaks (impulse time?) comes later and later:

View attachment 77613
One may indeed find good reasons for such reasoning to be correct, although I believe they both (IR and step) should yield the same result.

Main problem, however, in this case was that the initial measurement of mains and sub that @ernestcarl provided was obviously flawed by misconfigurations in main's DSP, miniDSP XO settings and XO and phase settings in sub (which cannot be bypassed by using LFE input as there is none).

Look for example at step response of the sub:

Capture.JPG


Not bad at first sight, right?

Now comes several "but"s..
Sub was supposed to be measured, together with right main speaker, using left main to provide common timing reference so that delay can be estimated. XO for the channel 4 driving the sub was supposed to be bypassed so that only XO set on sub is active - but it was not. XO at miniDSP on channel 4 was active, with XO on set to unknown parameter.
Phase on the sub was supposed to be set on 0 deg but it was set on 180 deg.

All this makes this measurement unusable.

Now let's look at the right channel step:

Capture_R.JPG


Ok, this one doesn't even look right on the first sight. It should be similar to the sub but instead it's a picture of sheer horror.

It was supposed to be made with XO HP setting on miniDSP channel 2 at 80Hz and delay of 0ms. PEQs on miniDSP should be disabled and all internal DSP settings on the main speaker should be disabled.

But that was clearly not the case, otherwise it would look much more similar to the step of the SW.

So, problem here is not in the methodology (using IR vs using step) but more into making clean precise measurements which would yield reliable and usefull result.

To answer your question, I believe both, IR and step, can be used as both should have single dominant clean peak which can be then compared to the similar peak of the other speaker to precisely determine the delay.

The only way out from this mess that I can see is get the settings right and repeat the measurements. :)
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
10,845
Likes
10,333
Location
Riverview FL
I'd like to see the Impulse and Step of that system using a full-range sweep now that it is all tuned up.

---

Just for grins, I played the five frequencies in my previous post (not the simulated step), and this is how my no-eq JBL LSR 308 make the sound at the listening position (bottom trace):

1597170748285.png
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
Wouldn't step response be more useful than band limited Impulse for time alignment?

As frequency decreases, the peaks (impulse time?) comes later and later:

View attachment 77613
Unfortunately I currently don't have sub(s) in my setup so I can't measure them. Do you have them?
It would be interesting to see IR and step measurements of sub and mains, esepcially if they are time aligned, so I would appreciate if you can make them. :)
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,085
Likes
738
Location
Canada
One may indeed find good reasons for such reasoning to be correct, although I believe they both (IR and step) should yield the same result.

Main problem, however, in this case was that the initial measurement of mains and sub that @ernestcarl provided was obviously flawed by misconfigurations in main's DSP, miniDSP XO settings and XO and phase settings in sub (which cannot be bypassed by using LFE input as there is none).

Look for example at step response of the sub:

View attachment 77618

Not bad at first sight, right?

Now comes several "but"s..
Sub was supposed to be measured, together with right main speaker, using left main to provide common timing reference so that delay can be estimated. XO for the channel 4 driving the sub was supposed to be bypassed so that only XO set on sub is active - but it was not. XO at miniDSP on channel 4 was active, with XO on set to unknown parameter.
Phase on the sub was supposed to be set on 0 deg but it was set on 180 deg.

All this makes this measurement unusable.

Now let's look at the right channel step:

View attachment 77619

Ok, this one doesn't even look right on the first sight. It should be similar to the sub but instead it's a picture of sheer horror.

It was supposed to be made with XO HP setting on miniDSP channel 2 at 80Hz and delay of 0ms. PEQs on miniDSP should be disabled and all internal DSP settings on the main speaker should be disabled.

But that was clearly not the case, otherwise it would look much more similar to the step of the SW.

So, problem here is not in the methodology (using IR vs using step) but more into making clean precise measurements which would yield reliable and usefull result.

To answer your question, I believe both, IR and step, can be used as both should have single dominant clean peak which can be then compared to the similar peak of the other speaker to precisely determine the delay.

The only way out from this mess that I can see is get the settings right and repeat the measurements. :)
For clarity, that actually looks like the Left channel you selected to show, not the right step.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
3,124
Likes
2,570
I'd like to see the Impulse and Step of that system using a full-range sweep now that it is all tuned up.
I doubt they will look nicely as IR and step of the right channel doesn't look very promising to me.
But ok, at least FR seems to be summed well. :)
 

dasdoing

Addicted to Fun and Learning
Joined
May 20, 2020
Messages
844
Likes
427
Location
Salvador-Bahia-Brasil
Time alignement of the sub and main is important as with adding the sub you are basically adding a driver to the mains (well, sort of). This will still not result in a perfect step response because of the phase error that will be introduced with LR24 crossover at 80Hz between sub and mains. If that phase error is corrected (and this can easilly be done with rePhase) you will end up with perfect step response.

But correcting sub's phase and GD around 40Hz, and GD is naturally raising as frequency is falling toward 20Hz, is something that should never be done. All subs are showing this effect which is totally inaudible and trying to correct it may only result in audible pre-ringing.

There is a reason why automatic EQ systems have pre-ringing compensation and I would never advise users to try to cheat them as things will only get worse.
we surely can treat our sub as a 3rd driver. but why? if we meassure at LP it only makes sense to align at LP, too.
yes, there is always a little pre-ringingin. but we are not selling nothing here. if we don't hear the pre-ringing, why care?
also step response and predelay are showing basicly the same thing. a perfect triangle in step response is a flat line in excess phase group delay. btw: excess group delay "filters" out the FR effect on phase/delay (and btw, phase and delay are also basicly the same thing). the natrual rise in group delay is not so natural at all lol. I mean, it's a consequence of using drivers to reproduce natural sound, but it is not what is happening in nature at the sound source. many concepts I read here are from the time we had no DSP. and many others are concepts from speaker building. now we have DSP to our disposal, and shoudl use it to get as close to natural reproduction as possible.
have you actualy listened to the sound examples I produced (link)? what is more important, 50 year old concepts pre-DSP or what you are hearing?
 

Similar threads

Top Bottom