Understanding subwoofers

[john61ct]

From my more noobish POV

Why is a DSP considered absolutely necessary for proper subwoofer integration, and what specific room acoustics or phase alignment issues make it indispensable

If phase issues need tweaking beyond basic inversion, or delays required to align arrival times at MLP.

Testing different xover types / slopes / frequencies.

The rest of your questions require the equivalent of university courses to fully address, but really, hands-on learning over years rather than just an academic study approach.

Which use of DSP would accelerate enormously.

 

[DY2]

From my more noobish POV



If phase issues need tweaking beyond basic inversion, or delays required to align arrival times at MLP.

Testing different xover types / slopes / frequencies.

The rest of your questions require the equivalent of university courses to fully address, but really, hands-on learning over years rather than just an academic study aporoach.

Which use of DSP would accelerate enormously.

I completely agree with you and that is exactly why I integrated the DSP into my system. My main goal was to move two of my front subwoofers further back because they were too far forward in the soundstage and it was a mess aesthetically. By using the DSP, I was able to dial in the delays and crossovers to integrate them perfectly with the rest of the stack, which is something I could never achieve manually. As you can see in the screenshots, I managed to set the exact crossover filters and slopes I needed for my subwoofers. I am attaching a few more room measurements along with my current DSP settings so you can see the before and after details.

---

 

[john61ct]

these are contradictory statements as a 4th order bandpass means a degraded transient response (but still a response described by the given frequency response so...)

...

Relevant

AIYIMA A70 Stereo Amplifier Review

This is true - and might be audible for people with very "young" hearing. This particular model (from which the charts are being quoted) though has PFFB and is flat. Right -- thanks to BassTinker pointing out the "A07" typo in that IMD graph, I see that the graph indeed is from the A70 that has...

www.audiosciencereview.com

How do you measure and mitigate that degradation?

Try to go without bandpassing completely?

LR crossovers specifically are to be avoided?

Or do you just mean bandpass ENCLOSURES, as a problematic physical design concept?

Not the case for using electronics to get there, can't be resolved via phase tuning in DSP?

 

[3ll3d00d]

...

Relevant

AIYIMA A70 Stereo Amplifier Review

This is true - and might be audible for people with very "young" hearing. This particular model (from which the charts are being quoted) though has PFFB and is flat. Right -- thanks to BassTinker pointing out the "A07" typo in that IMD graph, I see that the graph indeed is from the A70 that has...

www.audiosciencereview.com

How do you measure and mitigate that degradation?

Try to go without bandpassing completely?

LR crossovers specifically are to be avoided?

Or do you just mean bandpass ENCLOSURES, as a problematic physical design concept?

Not the case for using electronics to get there, can't be resolved via phase tuning in DSP?

a band pass box, by definition, is one where the enclosure design acts as a bandpass acoustic filter, aka has both high and low pass elements (sealed enclosure gives you one side & the PR in the front gives you the other). The transient response is then exactly as you'd expect for such a response.

idk if this is modelled correctly (the schematic is a bit hard to reaD) but it gives an idea

 

[john61ct]

Sorry way above my paygrade there. Yes I know what a bandpass enclosure is

> transient response is then exactly as you'd expect

I've no idea what to expect!

Are those lines good or bad?

Could you please answer my clarifying questions directly?

 

[3ll3d00d]

I've no idea what to expect!

I don't know what your Qs are exactly and/or they don't make so much sense to me (idk what the linked thread has to do with this either)

the point is a bandpass enclosure is a high pass and low pass filter applied physically not via dsp so if you then want to perform some form of excess phase correction using DSP, why bother using a bandpass enclosure in the first place? by definition, if you have the DSP horsepower to perform phase correction at low frequencies then you have DSP horsepower (and tolerance for latency) to create and apply linear phase pass filters filters so just build a simple sealed sub and cross it to a woofer.

 

[john61ct]

if you then want to perform some form of excess phase correction using DSP, why bother using a bandpass enclosure in the first place?

by definition, if you have the DSP horsepower to perform phase correction at low frequencies then you have DSP horsepower (and tolerance for latency) to create and apply linear phase pass filters filters so just build a simple sealed sub and cross it to a woofer.

My preference is to see how far I can go using analog, and then only use DSP as actually needed.

My limit is about 1 cuft per enclosure, up high starting ~200Hz MBM coupler "basstands" down to say 60Hz if possible

if not I might need two sets of trueSubs below that, to get high SPL in each frequency range given the size limitation, I do not expect high bandwidth from each box.

...

> idk what the linked thread has to do with this

After the post I linked to is a discussion on LF distortion, in which I was trying to find out about preventing "muddy" and "boomy" bass transient problems, keeping things crisp and tight.

This is a higher priority to me than SPL, if needed I can add more subs instead.

...

> I don't know what your Qs are exactly

I thought your comments implied that bandpassing could exacerbate these issues.

To the extent there is such a relationship, is that inherent with all the bandpassing methods, or just acoustically?

How do you measure and mitigate that degradation?

Try to go without bandpassing completely?

LR crossovers specifically are to be avoided?

Or do you just mean bandpass ENCLOSURES, as a problematic physical design concept?

Not the case for using electronics to get there, can't be resolved via phase tuning in DSP?

 

[Fahzz]

Question:
If the Subwoofer comes with it's own DSP included (SVS for example), is it unnecessary to employ Dirac/Audyssey or similar software?

 

[Keith_W]

Question:
If the Subwoofer comes with it's own DSP included (SVS for example), is it unnecessary to employ Dirac/Audyssey or similar software?

The subwoofer still needs to be time and phase aligned to the mains. That usually involves delaying the mains. If you don't have DSP upstream of the mains, there is no way you can do that. All that the subwoofer's DSP can do is place a few PEQ's here and there to chop some peaks, and that is it.

 

[Oddball]

The subwoofer still needs to be time and phase aligned to the mains. That usually involves delaying the mains. If you don't have DSP upstream of the mains, there is no way you can do that. All that the subwoofer's DSP can do is place a few PEQ's here and there to chop some peaks, and that is it.

Well not really, especially for the simple 2.x systems that most seem to be interested. Depending on the sub brand, there are delay and phase settings as well - that might just work for a relatively simple phase and delay alignment with 2 mains.

But then, there are many other things that need to fall into place, where a dedicated DSP could potentially do better, or not, depending on how we interpret it.

 

[Keith_W]

Well not really, especially for the simple 2.x systems that most seem to be interested. Depending on the sub brand, there are delay and phase settings as well - that might just work for a relatively simple phase and delay alignment with 2 mains.

I said that the subwoofers are usually delayed with respect to the main speakers. What happens when you delay them even more?

 

[Oddball]

I said that the subwoofers are usually delayed with respect to the main speakers. What happens when you delay them even more?

It would really depend on the setup and of course your goals. But it is a fact that you can delay them or phase adjust them it they have DSP. If not, then there should be no meaningful delay to start with. And then there is the new SVS auto EQ feature - which I know nothing about, but just learned that it is out there.

 

[3ll3d00d]

My preference is to see how far I can go using analog, and then only use DSP as actually needed.

tuning a bandpass is a) not exactly easy, b) requires you to be able to model it accurately, and c) may require you to build multiple test boxes hence (unless you're a woodworking deman) slow. The advantage is that it gives you more output in the pass band. In the end, this is what "going analogue" means (slow, less accurate, can have difficult to correct side effects).

To the extent there is such a relationship, is that inherently with all bandpassing methods, or just acoustically?

generally speaking, any filter, whether acoustic or digital, has an impact but that impact comes from the frequency response you're targeting so if that is your target then it is what it is (unless you get into the excess phase correction game). IMO the issue is that actually tuning a bandpass to hit a given target can be basically impossible once you factor in the physical reality of a realistic build and the difficulty in genuinely accurately modelling it. I suggest you try a few sims on paper (it's really not that hard), try to translate that into a physically practical build and then revisit.

 

[Kal Rubinson]

My preference is to see how far I can go using analog, and then only use DSP as actually needed.

Surely, this presumes you are using only analog sources.
Logically, those of us with only digital sources are unconstrained.

 

[john61ct]

Surely, this presumes you are using only analog sources.
Logically, those of us with only digital sources are unconstrained.

Of course it's a mix.

However that factor has no bearing that I can see on how the speakers are built. Of course if you use a PC as hub and player I suppose you may as well use DSP there from the get go.

I don't.

 

[Fahzz]

It would really depend on the setup and of course your goals. But it is a fact that you can delay them or phase adjust them it they have DSP. If not, then there should be no meaningful delay to start with. And then there is the new SVS auto EQ feature - which I know nothing about, but just learned that it is out there.

So I looked it up: https://www.svsound.com/pages/svs-auto-eq-faq
Answers my question - use SVS auto EQ then your DSP.

 

[john61ct]

Yes if you see difficulty as a negative, physical implementation of bandpassing is ofc more challenging,

but that's what I like about it. Journey rather than destination all that. Electronic EQ and crossovers as well, but there cost is a huge incentive with 15+ channels.

Since the HxW dimensions are fixed, I plan to use modular "tunnel sections" in 3/4" increments to prototype the depths of the chambers between end wall and baffles, total depth can be up to 15.5"

Of course using modeling software to get started.

See @Wolf 's 'Overdrive10' project, aka Kilauea as my source for inspo.

He used Unibox iirc in separate segments, but I've been told VituixCAD has direct support for enclosure modelling a band-pass with passive radiators

others suggest WinISD...

tuning a bandpass is a) not exactly easy, b) requires you to be able to model it accurately, and c) may require you to build multiple test boxes hence (unless you're a woodworking deman) slow. The advantage is that it gives you more output in the pass band. In the end, this is what "going analogue" means (slow, less accurate, can have difficult to correct side effects).

generally speaking, any filter, whether acoustic or digital, has an impact but that impact comes from the frequency response you're targeting so if that is your target then it is what it is (unless you get into the excess phase correction game). IMO the issue is that actually tuning a bandpass to hit a given target can be basically impossible once you factor in the physical reality of a realistic build and the difficulty in genuinely accurately modelling it. I suggest you try a few sims on paper (it's really not that hard), try to translate that into a physically practical build and then revisit.

 

[3ll3d00d]

The difficulty is not the negative, as that is entirely subjective and depends heavily on time/aptitude/desire, the lack of precision (partially driven by the difficulty in accurately modelling such a device) is the objective negative.

See @Wolf 's 'Overdrive10' project

I've already shown you a model of the same (or rather, an approximation as best could be done with the rough schematic available) using hornresp. You can make your own judgement on that

others suggest WinISD...

hornresp is most flexible, vituixcad can do this particular alignment, winisd is a toy

 

[john61ct]

Thanks much!

wrt imprecision

Say my stack per L/R side is

A - bookshelf "main" HPF only ~200Hz

B - MBM coupler "basstands" bandpass 80Hz up

C - trueSub #1 bandpass 50Hz up

D - trueSub #2 bandpass 15Hz up

If I can only afford say four always-on DSP output channels

I'm thinking use that for B&C or B&D only

or maybe B, then for "upstream" of C+D just for timing, use analog for that xover between them

To get good use out of MSO, overlaps are in fact required, so maybe NOT crossing over is better, imprecision becomes an advantage in the LFs, objectively :-D

I've already shown you a model of the same (or rather, an approximation as best could be done with the rough schematic available) using hornresp.

Not that this noob can necessarily grok that, but could you link to that? I musta lost track and can't find it searching.

 

[john61ct]

Question:
If the Subwoofer comes with it's own DSP included (SVS for example), is it unnecessary to employ Dirac/Audyssey or similar software?

Some subwoofers apply their DSP to the main pair as well, e.g. a basic delay and HPF.

That might be enough for you and your room.

Those two systems offer a lot more features and options

and include a lot of user-friendly automation.

There are dozens of other software projects and approaches, many have very steep learning curves some years of hands-on

Just sorting out which software is good for what, various filter types, hardware required etc is in itself a college level course's worth of self-study