questions about understanding subwoofer integration

[Chimiel]

Hi,

I'm trying to understand the conditions for good sub integration.

Maybe you can help me out with my understanding of 'phase correction';
A lot of people talk about phase correction, integrating a sub.
Is this nothing more or less then correcting delay times of the sub, when it’s distance to the listening position, differs from the distance of the monitors ?
I heard other explanations about unwanted cancellation or amplification of soundwaves, caused by phase differences,
but have a hard time understanding this, cause this seems unavoidable using multiple sound sources and a 'global' listening position.

Is good sub - monitor integration dependent only on 3 things ?;
1 a flat response in the crossover region,
2 correction of different delay time, caused by different distances from sound sources (or reflections) ?
3 matching SPL of subs and monitors ?

Kind regards,

Michiel

 

[ppataki]

This is what I do and I guess it matches well with what you described above

Setting subwoofer delay

I know this topic has been discussed before but I would like to provide a new approach (or ask if that makes any sense at all) I have tried different methods to set sub delay but all of them resulted in a different delay figures.....no consistency whatsoever Using acoustic timing reference in...

audiosciencereview.com

 

[ppataki]

And forgot to mention another post from the same thread:

Setting subwoofer delay

I' NOT a room-correction expert... Phase differences will ONLY be an issue in the narrow band around 80Hz where the regular woofer and sub are working together (where the soundwaves might be partially or completely out-of-phase relative to each other). Assuming there is no "different DSP...

audiosciencereview.com

I am not sure what type of microphone you are using to measure delay but if you prefer USB mics then you shall definitely go for the UMIK-2 instead of the UMIK-1 (the former has its own clock source hence it produces consistent and precise delay measurements)

 

[terryforsythe]

2 correction of different delay time, caused by different distances from sound sources (or reflections) ?

Also, different crossover topologies have different phasing characteristics. Moreover, the group delay of the speakers and the subwoofer(s) may be different at the crossover frequency. Adjusting the time delay is helpful to address those issues as well.

As an example, if the subwoofer is further from the listening position than the speakers, intuition may be that time delay should be added to the speakers. But, if the group delay of the speakers is higher than the subwoofer in the crossover frequency range, it could be that, even with the subwoofer being further from the listening position, the time delay still needs to be added to the subwoofer.

If the group delay of the speakers at the crossover frequency is close to that of the subwoofer and the speakers are closer to the listening position, it may be that time delay is needed for the speakers (I have not experienced this scenario myself). If you do not have a way to adjust time delay for the speakers, you can choose a crossover topology that increases the phase delay of the speakers at the crossover frequency, invert the subwoofer signal and apply time delay to it, or experiment with both.

In any case, use REW or something similar to measure while you make the time delay and/or crossover adjustments.

 

[ppataki]

Also, different crossover topologies have different phasing characteristics.

Very true
In order to mitigate that I only use linear phase crossovers - that will eliminate the phase-related impact

 

[terryforsythe]

In order to mitigate that I only use linear phase crossovers - that will eliminate the phase-related impact

I presume with FIR filters. Is that correct?

I would like to experiment with that, but my miniDSP Flex HTx, which I use for my system with the subwoofer, does not appear to support FIR filters. I'm hoping a firmware update comes through that allows user generated FIR filters to be used.

 

[ppataki]

I presume with FIR filters. Is that correct?

I use the thEQorange VST plugin

See some comparison measurements here

 

[mdsimon2]

I presume with FIR filters. Is that correct?

I would like to experiment with that, but my miniDSP Flex HTx, which I use for my system with the subwoofer, does not appear to support FIR filters. I'm hoping a firmware update comes through that allows user generated FIR filters to be used.

I've been meaning to explore the IIR subtractive delay filter concepts described in this thread ->https://www.diyaudio.com/community/...s-a-global-phase-linearization-by-fir.393435/, but haven't had a chance to do so.

Seems like a good way to implement linear phase crossovers if you do not have FIR capability but have the ability to implement relatively long delays. As the Flex HT can do 30 ms of delay this could be a good option for you.

Michael

 

[Chimiel]

Also, different crossover topologies have different phasing characteristics. Moreover, the group delay of the speakers and the subwoofer(s) may be different at the crossover frequency. Adjusting the time delay is helpful to address those issues as well.

As an example, if the subwoofer is further from the listening position than the speakers, intuition may be that time delay should be added to the speakers. But, if the group delay of the speakers is higher than the subwoofer in the crossover frequency range, it could be that, even with the subwoofer being further from the listening position, the time delay still needs to be added to the subwoofer.

If the group delay of the speakers at the crossover frequency is close to that of the subwoofer and the speakers are closer to the listening position, it may be that time delay is needed for the speakers (I have not experienced this scenario myself). If you do not have a way to adjust time delay for the speakers, you can choose a crossover topology that increases the phase delay of the speakers at the crossover frequency, invert the subwoofer signal and apply time delay to it, or experiment with both.

In any case, use REW or something similar to measure while you make the time delay and/or crossover adjustments.

Thank you for explaining.
I’m trying to understand phase conflicts. Does this apply to low frequency waves, of a couple of meters in length, for which it is important to connect in the crossover range ? In accordance with the part of the sound waves of other sources ? (subs / monitors). So a top is connected to a top of a different wave, and for instance not a bottom, causing cancellation ?
Cheers !

 

[sigbergaudio]

In reality a phase control is a pretty crude EQ for a limited frequency range. If you have access to a parametric EQ, you might as well just use that, typically with better and more precise results.

Another tip is to reduce inherent phase issues by plugging the port of the speakers (if they are ported).

 

[Keith_W]

Thank you for explaining.
I’m trying to understand phase conflicts. Does this apply to low frequency waves, of a couple of meters in length, for which it is important to connect in the crossover range ? In accordance with the part of the sound waves of other sources ? (subs / monitors). So a top is connected to a top of a different wave, and for instance not a bottom, causing cancellation ?
Cheers !

I began typing out a reply before I realised that I would have to type out half a chapter of a high school physics book. So i'll point you to a link: understanding phase.

 

[Chimiel]

In reality a phase control is a pretty crude EQ for a limited frequency range. If you have access to a parametric EQ, you might as well just use that, typically with better and more precise results.

Another tip is to reduce inherent phase issues by plugging the port of the speakers (if they are ported).

Ah, you mean using a parametric EQ for the crossover frequency, and finding one with the least cancellations / amplifications, instead of fiddling with the phase in a predetermined crossover frequency ? Do some like to do both; first finding the best crossover frequency, determined by postion of monitors, subs and room reflection, and finetune with phase correction, or the other way around ?

 

[sigbergaudio]

Ah, you mean using a parametric EQ for the crossover frequency, and finding one with the least cancellations / amplifications, instead of fiddling with the phase in a predetermined crossover frequency ? Do some like to do both; first finding the best crossover frequency, determined by postion of monitors, subs and room reflection, and finetune with phase correction, or the other way around ?

1. Find the placement where the subwoofer measure the best on its own
2. Set the crossover (testing different ones if need be)
3. Adjust the frequency response with EQ as needed.

When you have done this, there is no need for additional phase adjustment. Low frequencies are minimum phase, so the frequency response follows the time/phase. So if you are able to get a better frequency response using phase, the same can be done with EQ.

 

[terryforsythe]

1. Find the placement where the subwoofer measure the best on its own

This is not always practical, especially in terms of aesthetics. In my family room, for example, there only is one location available for the subwoofer without degrading the room's aesthetics.

When you have done this, there is no need for additional phase adjustment. Low frequencies are minimum phase, so the frequency response follows the time/phase. So if you are able to get a better frequency response using phase, the same can be done with EQ.

I also use EQ, but not alone. The combination of phase adjustment and EQ works better for the crossover transition between the subwoofer and speakers. By also getting the subwoofer and speakers in phase as measured at the listening position, there is less boost from EQ that is needed. Too much EQ boost can cause clipping of the low level signal going into the amplifier. To compensate, the overall signal level may need to be reduced (I believe that is how miniDSP handles it). This can be an issue in some systems, for example if using an amplifier with relatively low gain.

If one does not have phase adjustment available, then use EQ and hope for the best. But, if phase adjustment is available (it is a feature on many popular subwoofers, as well as many audio DSP processors), then adjusting the phase should be the first step. After the phase is optimized, then EQ.

 

[terryforsythe]

Thank you for explaining.
I’m trying to understand phase conflicts. Does this apply to low frequency waves, of a couple of meters in length, for which it is important to connect in the crossover range ? In accordance with the part of the sound waves of other sources ? (subs / monitors). So a top is connected to a top of a different wave, and for instance not a bottom, causing cancellation ?
Cheers !

Yes, the goal is to have the signals from the subwoofer and speakers sum at the crossover frequency, not cancel.

There is a caveat, though. What I have written assumes tuning a subwoofer and speakers for a particular listening position. When tuning multiple subwoofers to cover a larger area, it gets more complicated. In that case, it is best to model the system, e.g. in REW's Room Simulation, to determine the optimal configuration. In REW you can model your room/system, adjust component placements, applied time delays, etc. You can make changes and review the impact of those changes on the overall response. I'm not sure how well REW can handle room shapes that are non-rectangular, though. There may be other software out there that can handle that better, but I don't know.

 

[sigbergaudio]

This is not always practical, especially in terms of aesthetics. In my family room, for example, there only is one location available for the subwoofer without degrading the room's aesthetics.

My apologies, I assumed it was implied that you should measure what position measures best among the locations available. I should have been more explicit.

 

[ppataki]

Yes, the goal is to have the signals from the subwoofer and speakers sum at the crossover frequency, not cancel.

There is a caveat, though. What I have written assumes tuning a subwoofer and speakers for a particular listening position. When tuning multiple subwoofers to cover a larger area, it gets more complicated. In that case, it is best to model the system, e.g. in REW's Room Simulation, to determine the optimal configuration. In REW you can model your room/system, adjust component placements, applied time delays, etc. You can make changes and review the impact of those changes on the overall response. I'm not sure how well REW can handle room shapes that are non-rectangular, though. There may be other software out there that can handle that better, but I don't know.

That's exactly what I do as well, if the room is rectangular I use REW's Room Sim to simulate if and where it makes sense to place one or more subs
Then I confirm that by performing measurements in those positions

However, if the room is non-rectangular, for example in this project, then what I do is that I put the sub in all the physically/aesthetically possible positions and do a measurement
If there is an option to use two or more subs then you can combine the measured responses from various positions in REW (Trace Arithmetic --> A+B) to see approx. how the final combined response would look like

 

[mdsimon2]

I've been meaning to explore the IIR subtractive delay filter concepts described in this thread ->https://www.diyaudio.com/community/...s-a-global-phase-linearization-by-fir.393435/, but haven't had a chance to do so.

Seems like a good way to implement linear phase crossovers if you do not have FIR capability but have the ability to implement relatively long delays. As the Flex HT can do 30 ms of delay this could be a good option for you.

Michael

Follow-up on this, I had some time to play around with IIR subtractive delay filters and now understand them more. Unfortunately it requires you have the ability to subtract a Bessel LPF from a delayed full range response, and because all of the mixers in the miniDSP are upstream of the delay/x-over/EQ blocks this really isn't possible in the miniDSP.

You can do it in more advanced platforms like CamillaDSP or even the Okto XMOS DSP. I modeled a subtractive filter system based on a Bessel 6th order filter that is -6 dB at 80 Hz and it can give you flat frequency response and linear phase using only IIR and 7.4 ms of delay.

Michael

 

[terryforsythe]

Follow-up on this, I had some time to play around with IIR subtractive delay filters and now understand them more. Unfortunately it requires you have the ability to subtract a Bessel LPF from a delayed full range response, and because all of the mixers in the miniDSP are upstream of the delay/x-over/EQ blocks this really isn't possible in the miniDSP.

You can do it in more advanced platforms like CamillaDSP or even the Okto XMOS DSP. I modeled a subtractive filter system based on a Bessel 6th order filter that is -6 dB at 80 Hz and it can give you flat frequency response and linear phase using only IIR and 7.4 ms of delay.

View attachment 404196
View attachment 404197
Michael

Thank you for investigating this!