Subwoofer Time Alignment

[waynel]

Read through this thread with interest and have a basic question:

Assuming that the sub phase and level are adjusted properly to give a smooth frequency response (and then perhaps smoothed out further with DSP room correction) , does time alignment matter? Is there any evidence that the effect of time alignment over what can be represented in a frequency response plot is audible?

Thanks
Wayne

 

[ernestcarl]

Ok, I see. Can you plz repeat these 4 measurements, this time with 50-130Hz sweep and add 4ms delay to the mains.

View attachment 77432

I presume you still wanted to keep the 'acoustic timing reference' channel output set to 0ms so I did just that.

4ms was added only to the channel actually being measured while the reference channel remained set to 0ms.

e.g.

REW's timing delay calculation all fall under 0ms with the fronts set to 4ms delay.

I suppose you could say the phase restricted between 80-100Hz looks better matched...

Not quite so below that though.

 

[dasdoing]

Is there any evidence that the effect of time alignment over what can be represented in a frequency response plot is audible?

check this one out. it should be audible for everyone. what sounds better though you decide (cause the closer you get to the perfect step response the more pre-ringing you get) https://www.audiosciencereview.com/...xess-group-delay-corrected.14489/#post-445969

 

[QMuse]

I presume you still wanted to keep the 'acoustic timing reference' channel output set to 0ms so I did just that.

4ms was added only to the channel actually being measured while the reference channel remained set to 0ms.

e.g.
View attachment 77450

REW's timing delay calculation all fall under 0ms with the fronts set to 4ms delay.

I suppose you could say the phase restricted between 80-100Hz looks better matched...
View attachment 77451
Not quite so below that though.

Well, this looks much better, so let's do the XO work and we'll fine tune phase later if necessary.

Set XO point on your sub at 80Hz, leave phase at 0.

Leave 4ms delay on both mains.
Set XO for mains at 80Hz on the high pass filtr like shown on the pic, except use LR24, not BW.
Bypass the low pass filter as shown on the pic.

Make 0-20kHz sweep of your left channel (that would now mean left+sub) without timing reference. Make the same for right channel.

 

[ernestcarl]

All drivers summed and using the original xo AND conventional LR4 at 80Hz using different delays:

neon green and cyan are Left and/or Right only with sub.

I'd say from these plots, the original delay of 13ms looks the flattest and most even i.e. still the best compromise?

Back to where I started, I guess.

 

[QMuse]

All drivers summed and using the original xo AND conventional LR4 at 80Hz using different delays:

View attachment 77455
neon green and cyan are Left and/or Right only with sub.

View attachment 77456

I'd say from these plots, the original delay of 13ms looks the flattest and most even i.e. still the best compromise?

Back to where I started, I guess.

Let's stick to the standard procedure how this is done and do things step by step. Can you plz post those 2 sweeps I asked in my previous post?

Also, be sure to setup XO on the SW side and on the miniDSP side as I described.

 

[ernestcarl]

Sure

 

[QMuse]

Sure

You're right with the 13ms, I aligned impulse responses wrongly.

Here's the overlay of the IRs with delay set to 0.

As you can see main highest peak is at 1ms and SW highest peak is at 14ms, hence mains should be delayed by 13ms. So plz set delay to 13ms and measure again both channels with sweeps.

 

[ernestcarl]

You're right with the 13ms, I aligned impulse responses wrongly.

Here's the overlay of the IRs with delay set to 0.

View attachment 77461

As you can see main highest peak is at 1ms and SW highest peak is at 14ms, hence mains should be delayed by 13ms. So plz set delay to 13ms and measure again both channels with sweeps.

Here you go.

 

[QMuse]

Here you go.

Something is still not right. Are you sure you're using 80Hz XO at SW side and LR24 at 80Hz with HP filter at mains side with miniDSP? Can you also check that all other XO parameters are disabled in miniDSP, especially LP filter on mains channels and all filters on SW channel?

 

[ernestcarl]

Something is still not right. Are you sure you're using 80Hz XO at SW side and LR24 at 80Hz with HP filter at mains side with miniDSP? Can you also check that all other XO parameters are disabled in miniDSP, especially LP filter on mains channels and all filters on SW channel?

Oh, you meant using 80Hz LR4 with 13ms delay...

The attached measurements are using the above settings.

 

[ernestcarl]

In case it may be of importance, I used 512k length for these and not 2M setting in REW.

 

[ernestcarl]

Something I don't know how to explain though is the ff:

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.

 

[QMuse]

Oh, you meant using 80Hz LR4 with 13ms delay...

View attachment 77464

The attached measurements are using the above settings.

No, that is not what I meant. Both of those filters on the channel 4 driving sub should be bypassed and XO on sub should be set at 80Hz using XO knob on the rear panel of the sub.

On both channels driving mains you should enable high pass at 80Hz with LR24db/oct and disable low pass (as I showed on the picture I posted few posts earlier).

Also, please always use 2M sweep length.

 

[QMuse]

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.

Something is horribly wrong here. I suggest you remove all filters (PEQ and FIR) from miniDSP. Plz leave only high pass LR24 filter at 80Hz with channels 1 & 2 which are driving your mains. Bypass HP and LP filter at channel 4 which is driving your sub. Leave also delay setting for the mains.

80Hz XO point for the sub should be configured at the sub as you can't turn it off.

Do you have the possibility to invert polarity of the mains within miniDSP?

 

[QMuse]

Something is horribly wrong here. I suggest you remove all filters (PEQ and FIR) from miniDSP. Plz leave only high pass LR24 filter at 80Hz with channels 1 & 2 which are driving your mains. Bypass HP and LP filter at channel 4 which is driving your sub. Leave also delay setting for the mains.

80Hz XO point for the sub should be configured at the sub as you can't turn it off.

Do you have the possibility to invert polarity of the mains within miniDSP?

This is what I think you should do:

- remove all filters (PEQ and FIR) from miniDSP
- enable high pass LR24 filter at 80Hz with channels 1 & 2 which are driving your mains, bypass low pass filter on these channels
- set delay to 10ms for mains

- bypass HP and LP filters at channel 4 which is driving your sub
- invert polarity of channel 4 driving your sub

- 80Hz XO point for the sub should be configured at the sub

Use 2M sweeps in the 50-110Hz range to measure left and right channel, without timing reference.

 

[ernestcarl]

Polarity can be easily changed within miniDSP, yes.

I have not been using FIR filters only ordinary PEQs. Pretty sure the ff. L&R + sub measurements are as you requested.

 

[ernestcarl]

I know what the discrepancy now is, QMuse. The left monitor's internal 60Hz 12dB/oct HP has been enabled all this while!

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

 

[waynel]

check this one out. it should be audible for everyone. what sounds better though you decide (cause the closer you get to the perfect step response the more pre-ringing you get) https://www.audiosciencereview.com/...xess-group-delay-corrected.14489/#post-445969

Thank you. That is interesting. I was thinking more of controlled studies.

 

[QMuse]

Polarity can be easily changed within miniDSP, yes.

View attachment 77471

I have not been using FIR filters only ordinary PEQs. Pretty sure the ff. L&R + sub measurements are as you requested.

Uh, we don't seem to understand each other. Low pass filter at channel 4 should also be bypassed:

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

Please do that an than run 0-20khz sweep on the left and right channel without timing reference.