Principles: Time of Arrival Delays, vs Phase tuning ?

[3ll3d00d]

Then trueSub(s) below also active electronics, if the whole stack seems well blended, sum magnitudes reasonably

Then look at the phase side of things, also if any time of arrival delay issues.

Fwiw this (saying that looking at phase/timing comes after things have summed nicely implies wave superposition is not understood) indicates fundamental knowledge gap despite all these posts, running the most simple simulation will illustrate this (one practical test beats 1000 forum posts hands down).

 

[melomane13]

Then active LR 24dB/octave but non-DSP to bandpass the MBMs, do the blending best as possible that way.
....

Then look at the phase side of things, also if any time of arrival delay issues.
...
If that stuff actually needs fixing that's where I start with DSP. "Linearizing" doing phase / time correcting the existing xovers IIR or FIR
...

"To achieve a correct LR (Linkwitz-Riley) crossover, the phases must be equal at and around the frequency in question—for a 24dB/oct slope, at least 1/2 octave on both sides.
To obtain an identical phase, as explained by Linkwitz for his 'Pluto' speaker, a Linkwitz Transform must be used on the smaller woofer to bring it to the same transfer function as the larger woofer.
Possibly, if the two drivers are not coincident, or at least within a quarter of the wavelength (for example, 1m if the crossover is at 100Hz), it will also be necessary to introduce a delay to compensate."

 

[melomane13]

Of course, the Linkwitz Transform and LR24 crossover can be achieved without a DSP, in the analog domain, but it requires op-amps and a lot of calculations—see Linkwitz's Pluto.
As far as I know, analog transform circuits don't really exist outside of the DIY."
However, this is not the case for the delay."

 

[john61ct]

Sorry I do not understand. Are these explanations of theoretical factors? Or supposed to be practical suggestions using actual devices?

 

[thcdru2k]

melomane13 is mainly concerned with getting the acoustic LR crossover to sum correctly, not just applying textbook electrical LR24 filters. Since your project involves multiple LF sections and several handoffs, that matters a lot.

Each handoff has to sum correctly in the room. Your project is doable, but it will be a serious measurement and troubleshooting project. There will absolutely be phase rotation issues, but it will not just be simple A-to-B rotation issues. With several LF sections, it becomes multi-source / multi-handoff interaction.

I imagine all-pass filters could become a useful last-mile phase tool

 

[john61ct]

I want to do the crossovers independently of room effects, rather than having to redo everything whenever the space changes.

I guess that means working outdoors, or perhaps the 2-mic tech I read about designed to simulate that?

If the relative distances of all the drivers can't remain the same, can those delays (as opposed to linearisation / phase tuning) be handled at the same time as DRC?

 

[john61ct]

Since your project involves multiple LF sections and several handoffs

Does "sections" = speakers in this context?

 

[levimax]

A couple of comments. I agree with the concerns about acoustic vs electronic crossovers. Part of it is the order you have things written down. One method of getting acoustic crossovers is to "flatten" the response 1/2 to one octave above and below the crossover point on both drivers, which will also flattens the phase, and then apply the 24 dB electronic crossover filters. You can also measure each driver with REW and create the crossover slope in REW and then use manual or automatic filters to "match" the crossover slope (it will NOT be a 24 dB filter slope). Both methods accomplish the same thing, creating acoustic rather than electronic crossovers. Nothing to do with the room at all.

One other comment I have is that I don't think you need MBM and Subs. Every additional driver added to the stack creates a lot of issues and oftentimes make things worse rather than better. A "big" sub with a "pro" driver can easily cover 20 Hz to 500 Hz and will be especially strong over 100 Hz (Pro drivers need to have a high shelf to "shave off" response above 100 Hz to balance their 20 Hz output, pro-drivers extremely high efficiency makes this work i.e. both low extension and powerful midbass so it eliminates the need for a MBM and having all the bass and some mid-bass on one driver eliminates MANY phase issues. If you crossed around 180 Hz you could get the drivers closes enough ( 1/4 wavelength) to use Linear Phase crossovers which really improves time domain performance and also cancels pre-ringing. I have a similar set up and it work very well.

 

[john61ct]

Great response, tx

acoustic vs electronic crossovers

Those aren't different toolsets right ?

I see the acoustics as the results you (try to) get from tools. Passive vs electronics vs DSP are how I classify the tools.

> measure each driver with REW

Even within the multi-way passives? I'm hoping I can just treat each of those as a single unit, like a "virtual driver"?

> and create the crossover slope in REW

rePhase as well, right?

Or alternatively REW just for measuring, then Acourate or Audiolense for filter creation?

...

> A "big" sub with a "pro" driver can easily cover 20 Hz to 500 Hz

All that sounds GREAT as a goal

But you think that can be done within a 1.x cuft box?

I am limiting the form factor to around a 12x12" face, 1-1.5 cu ft max internally - can be over 15" deep if needed, more internal volume means deeper bass, right?

 

[john61ct]

Fwiw this (saying that looking at phase/timing comes after things have summed nicely implies wave superposition is not understood) indicates fundamental knowledge gap despite all these posts, running the most simple simulation will illustrate this (one practical test beats 1000 forum posts hands down).

Yes I realise.

Note this is not a description of proper workflow, I know I have no clue about that currently

But my stages if learning once my situation allows starting with the hands on.

As that curve progresses I will eventually get better results, but early days it's just experimenting, observing cause and effect.

Analog crossovers, only using DSP to treat issues the analog system cannot fix

If I have Main L/R comprised of a mix 2- 3-way speakers + midbass couplers + stereo trueSubs all co-located Those subs easily switched to mono if that turns out to test better, there are also other trueSubs already in mono thus freely placed, can be treated as a single sub-group in the DSP...

www.audiosciencereview.com

 

[levimax]

Great response, tx

Those aren't different toolsets right ?

I see the acoustics as the results you (try to) get from tools. Passive vs electronics vs DSP are how I classify the tools.

> measure each driver with REW

Even within the multi-way passives? I'm hoping I can just treat each of those as a single unit, like a "virtual driver"?

> and create the crossover slope in REW

rePhase as well, right?

Or alternatively REW just for measuring, then Acourate or Audiolense for filter creation?

...

> A "big" sub with a "pro" driver can easily cover 20 Hz to 500 Hz

All that sounds GREAT as a goal

But you think that can be done within a 1.x cuft box?

Acoustic vs electronic: Acoustically correct crossovers are the result, nothing to do with the tools or techniques. In a perfect world with perfect drivers you could just use stock electronic crossover filters to drive your speakers and your crossovers would sum perfectly. In the real world the speakers are not linear so if we want the acoustic output of the speakers to follow the LR 4 curve we have to make allowances for these non linearities in the electronic filters so the speakers outputs the curve we want. There are many was to do this. I would start by doing as much as possible with REW including creating the filters. REW can pretty much do anything just not always the easiest. It is still easier and more reliable to learn one tool well than trying to get multiple tools to work together.

Re sub. For that size you might want to go dual opposed and you aren't going to get to 20 Hz but for sure you don't need or want 3 boxes in my opinion.

 

[john61ct]

front firing only I'm afraid

I guess once I start building them I'll see. Modeling will give some idea even before that.

Tools for building better speakers

I don't know why this point keeps getting made. Of course NFS, or anechoic chamber don't design speakers by themselves. And having one as tools, doesn't assure a good design. I as a reviewer could have done my job without these as well. But with massive amount of work and possibility of...

www.audiosciencereview.com

 

[Sigberg Audio]

@john61ct It is the deep bass that requires a big box. Giving that same box midbass duties (instead of having a separate box) as well doesn't increase the size requirement.

 

[melomane13]

Sorry I do not understand. Are these explanations of theoretical factors? Or supposed to be practical suggestions using actual devices?

"Both, really! The theory directly connects to practical use with real devices."
I applied these principles to my own system.

In-depth explanations can be found here:

Pluto + subwoofer

 

[Bland_username42069]

@john61ct

I'm so confused that I get FOMO about figuring out what you're really asking here, so I'm diving back in.

I suggest you ask a single question at a time, and then pursue that to completion (until you fully understand the answer), before asking another.

I will start with one of my own. It appears you are planning on designing a multiway speaker, where not all the ways (hm) are in the same enclosure, and potentially not even colocated with the additional enclosure(s). Is that correct?

From what I can ascertain he is basically trying to make an avr but just for stereo upmixing. The twist is he is refusing to use prologic let alone atmos and instead only a 70s style quad upmixer and it must have dsp but he’s not using active speakers.

This could all be done with a $100 used AVR with prologic and audessy.

 

[Waveform Fidelity]

Exactly why I say I prefer not to do that.

There are literally thousands that simply accept DC power input, designed to run off an external PSU. Mostly 12V with internal boost conversion, but many at 24/36/48Vdc

3e A7 series is OK with even higher, but even 51Vdc is pushing up to the TPA3255 limit, not a great idea, when every last dB is needed best to buy bigger in the first place.


> For someone that thinks that DAC oscillators drift enough to mess up the sound

Simple conversion is required, but not sufficient for the goal here, did you read the use case details here?

Are you asserting that with maybe a dozen independent DSP units, each running crossover and delay filters on their downstream channels, of varying types and tap counts, clock drift will not have audible consequences?

How about when MSO gets involved on the overlapping mono subs?

Or when they are also servicing DRC eventually, not just bass management?

If you know how to manage phase timing issues reliably with a bunch of rPi and/or Pico mcu please link to relevant resources.

Nearly every member I respect thinks it's a fool's errand to even try without using one central DSP, going through on big A/D converter, so all the myriad I/O analog ports are served by a single clock.

If you take 2 identical DACs, one outputting exactly 1000.000Hz (say f1) and the other running slightly faster, lets say its drifted due to aging, temp change etc by 100ppm its frequency will be 1000.100 Hz (say f2) and pass these through a speaker, you will now have 3 signals f1,f2, and a new signal called the beat frequency, f3. F3 is the difference between f1 and f2, so that 0.1Hz.
So nothing to do with timing delay, since both signals are leaving the source at the same time and traveling the same distance.

Your speaker, let alone your hearing, will not be able to detect that.

You may also want to read about “wow and flutter” which causes detectable changes in pitch (from tape recorders) and how the arrival of digital audio (mainly the CD) completely dispensed with that problem.

 

[john61ct]

Thanks but I really don't understand the relevance.

 

[john61ct]

From what I can ascertain he is basically trying to make an avr but just for stereo upmixing.

I am using the SYN for that, not trying to "make" anything related to that.

AVR means AC powered amps right? Yes I don't want that.

I can always slot an AVP in instead if I want later.

 

[john61ct]

Pluto + subwoofer

Excellent! Put me onto what looks like a great driver for the MBM

Peerless by Tymphany SLS 10" 830668

I see @Wolf has used them together with the SB SB15SFCR-00 5x8" PR

 

[Waveform Fidelity]

Thanks but I really don't understand the relevance.

Your posts^^ were claiming how bad the sound is if DAC’s are not synchronized because of delays, which the above example shows you that is non-sense