Subwoofer Time Alignment

[dasdoing]

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

that's exactly what I am talking about. also at listening position we have diferent (peak) delays at diferent frequencys.
if you align to the start of the waves you get (peak) delay in the low end.
If you align the peaks you get pre-ringin.
what is better? well, the least amount of delay you can get before you hear the pre-ringin. saying pre-ringin must be 0 is a concept I can't understand. delay is clearly audible; why ignore it?

 

[QMuse]

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

Here is IR, step and overlay of both of my right speaker. Sweep is 2M, 30-150Hz, FDW of 6 cycles was applied.

IR

Step:

Overlay:

And here's step without FDW:

Judging by these 4 graphs, step with appropriate FDW seems to be the best choice for time aligning job.

 

[ernestcarl]

Indeed it is, my bad. Here's the right channel step, it looks even weirder.

View attachment 77621

Well, I did mention that I switched the sub's internal delay back to zero and kept it as such.

For what is effectively the last settings I made and with no PEQ's around the xo region other than xo shaping filters the ff. would be more representative, I believe:

I think the weirdness of the speaker's impulse and step in the above can partially be explained by SBIR -- they are against the wall and almost touching i.e. very much room and placement dependent.

If we filter the response to, say, 80Hz, which is quite close to the actual xo at 76Hz:

I'm not sure they really look all that different to me.

 

[QMuse]

If I do the same with @ernestcarl initial measurements (posted here) this is what I'm getting. From this one may conclude that mains and subs are already pretty well time aligned here:

FDW of 5 cycles, Left channel purple, sub green (right is same as left, so not shown for clarity):

 

[QMuse]

Well, I did mention that I switched the sub's internal delay back to zero and kept it as such.

For what is effectively the last settings I made and with no PEQ's around the xo region other than xo shaping filters the ff. would be more representative, I believe:

View attachment 77628

View attachment 77629

I think the weirdness of the speaker's impulse and step in the above can partially be explained by SBIR -- they are against the wall and almost touching i.e. very much room and placement dependent.

If we filter the response to, say, 80Hz, which is quite close to the actual xo at 76Hz:

View attachment 77630

View attachment 77631

I'm not sure they really look all that different to me.

What do you mean when you say "filter the response to 80Hz"?

 

[ernestcarl]

What do you mean when you say "filter the response to 80Hz"?

It's in the Filtered IR tab

 

[ernestcarl]

If I do the same with @ernestcarl initial measurements (posted here) this is what I'm getting. From this one may conclude that mains and subs are already pretty well time aligned here:

FDW of 5 cycles, Left channel purple, sub green (right is same as left, so not shown for clarity):

View attachment 77632

Don't quite know yet how to draw plots with FDW, but if this was from that original post then the subwoofer would have still been at max phase/delay internally. The 11.7ms time delay was the average of what REW calculated.

 

[ernestcarl]

If I do the same with @ernestcarl initial measurements (posted here) this is what I'm getting. From this one may conclude that mains and subs are already pretty well time aligned here:

FDW of 5 cycles, Left channel purple, sub green (right is same as left, so not shown for clarity):

View attachment 77632

Ah, okay, I see, this is the left, right speakers and subwoofer:

*latest setting

 

[QMuse]

Don't quite know yet how to draw plots with FDW, but if this was from that original post then the subwoofer would have still been at max phase/delay internally. The 11.7ms time delay was the average of what REW calculated.

Yes, that was the first measurement you posted so I also believe sub would have been at max phase/delay internally. Still, judging by step response overlay between mains and sub such setting needed only 2.5 ms delay to mains.

Can you please post 50-110Hz measurements for L, R and sub with your current delay settings?
Turn off your XO settings and set sub XO frequency to max.

What would also be interesting is to see step and IR of your L and R channels made with full sweep when XO is active (so when sub is working with mains).

P.S. I don't use Filtered IR.
Using FDW is easy, you apply it to selected measurement via Tools/IR Windows menu and then you work normally as you would without it. It is usefull when working with phase as it eliminates reflections so phase/step graphs are presented clearly.

 

[QMuse]

Ah, okay, I see, this is the left, right speakers and subwoofer:
View attachment 77638
*latest setting

This looks perfect!

Is it with XO's active or disabled?

 

[ernestcarl]

Yes, that was the first measurement you posted so I also believe sub would have been at max phase/delay internally. Still, judging by step response overlay between mains and sub such setting needed only 2.5 ms delay to mains.

Can you please post 50-110Hz measurements for L, R and sub with your current delay settings?
Turn off your XO settings and set sub XO frequency to max.

What would also be interesting is to see step and IR of your L and R channels made with full sweep when XO is active (so when sub is working with mains).

P.S. I don't use Filtered IR.
Using FDW is easy, you apply it to selected measurement via Tools/IR Windows menu and then you work normally as you would without it. It is usefull when working with phase as it eliminates reflections so phase/step graphs are presented clearly.

View attachment 77637

That snapshot makes things clearer.

The only option left alone was the Window Ref Time other options same as your example.

I need to use three filters to just to help shape the curve for the left and right channels since putting a xo simply 'as is' in miniDSP misses my xo target curve (SBIR issues).

 

[ernestcarl]

This looks perfect!

Is it with XO's active or disabled?

That would be active and same as above (only 40-110Hz -- steep xo anyway). Didn't use FDW but just overlayed the measurements as they were.

 

[ernestcarl]

I didn't try to follow the LR 48dB/oct target curve exactly, just as long as it conformed enough.

The xo is quite steep:

 

[ernestcarl]

God, I'm hungry! it's over three hours since I missed lunch... Think I'm going to call it a day!

 

[QMuse]

I didn't try to follow the LR 48dB/oct target curve exactly, just as long as it conformed enough.

The xo is quite steep:
View attachment 77643

View attachment 77644

View attachment 77646

This looks very good. It would also be interesting to see the phase response of 40-110Hz sweep and step and phase response of 20-20kHz sweep.

I need to use three filters to just to help shape the curve for the left and right channels since putting a xo simply 'as is' in miniDSP misses my xo target curve (SBIR issues).

There is something strange happening with step response of your mains but I doubt it has anything to do with SBIR. That shape of step response is simply not "normal". Even when a main speaker itself is not fully phase corrected (such as mine) step response shouldn't be looking like that with 40-110Hz sweep as only speaker's woofer will be playing - it should be looking fairly similar to your sub response.
So usually you don't need 3 "shaping" XO filters as one should do the job just fine.

Anyway, I have created a separate thread to show how this should be done in a "standard" case.

 

[_Bass]

Anyway, I have created a separate thread to show how this should be done in a "standard" case.

THANK YOU!

 

[Deleted member 17820]

Um, this is interesting to me but if time alighment of bass was so important why would anyone use ports.
Don't ports alter the phase and timing of the sub frequencies?

 

[ernestcarl]

This looks very good. It would also be interesting to see the phase response of 40-110Hz sweep and step and phase response of 20-20kHz sweep.



There is something strange happening with step response of your mains but I doubt it has anything to do with SBIR. That shape of step response is simply not "normal". Even when a main speaker itself is not fully phase corrected (such as mine) step response shouldn't be looking like that with 40-110Hz sweep as only speaker's woofer will be playing - it should be looking fairly similar to your sub response.
So usually you don't need 3 "shaping" XO filters as one should do the job just fine.

Anyway, I have created a separate thread to show how this should be done in a "standard" case.

When I have time I'll try to take a "very near-field" measurement of the mains and see if the step looks more 'normal' i.e. minus SBIR.

 

[ernestcarl]

Um, this is interesting to me but if time alighment of bass was so important why would anyone use ports.
Don't ports alter the phase and timing of the sub frequencies?

Don't have a ported sub... But, even if sub frequencies come in a cycle/few milliseconds late, I'd still endeavor to get the phase aligned as much as possible near the xo region to get the least amount of cancellation between the speakers and sub(s) as they sum at the main listening position.

 

[QMuse]

Um, this is interesting to me but if time alighment of bass was so important why would anyone use ports.
Don't ports alter the phase and timing of the sub frequencies?

Time (phase) aligning sub to main speaker is done in the narow frequency range around XO frequency to ensure smooth summing. Port affects phase mainly around port tuning frequency which is typically much lower than the XO frequency. Phase around XO frequency won't be affected if you plug the port or leave it open as this will affect only the sub roll-off region.