Multi-Channel, Multi-Amplifier Audio System Using Software Crossover and Multichannel-DAC

[dualazmak]

Hello @LSPhil,

Thank you for your kind attention on my multichannel audio project.
In my post #931, especially in Fig.03 and Fig.12, you can find all the info you are asking.

Spoiler

 

[Sal1950]

Hello, congratulations on such an impressive and complex audio system, as well as on your precise description (122.000 words).

Did you really count them? LOL

 

[LSPhil]

From the diagram, I can identify an active LR2 filter, or in other words, a second-order filter with a Q factor of 0.5. After that, there seems to be a first-order Butterworth filter (though the exact frequency is unclear). I assume the passive crossover from Yamaha is not present, and finally, the driver—such as the midrange—acts as a filter with a resonant frequency, possibly around 333Hz, and a Q factor of about 1.0. Combined, these four filters create an acoustic filter, which is the key element to analyze.

On the other side of the woofer, there's an LR2 filter at 500Hz. We can observe that, despite the use of such different filters, a linear amplitude response seems to be achieved. My concern is that the shallow filter of the woofer, combined with the steeper filter of the midrange, might result in a blending of the woofer’s frequencies, ultimately degrading the sound quality. The woofer is actually the weakest link in the system, primarily due to its age. The suspension is misaligned from wear, and its paper cone isn’t quite on par with that of the midrange. Upgrading this could lead to a significant improvement in quality, perhaps by as much as +10dB.

 

[dualazmak]

Hello @LSPhil,
You can find here (on this thread) and here (remote independent thread post) hyperlink index of this project thread.
I would like to recommend you to firstly read through the Hyperlink Index of this post.

My 30 cm woofer, 8.8 cm Be-midrange dome (still one of the best midrange ever produced), and Be-tweeter are in amazingly designed heavy-rigid YAMAHA NS-1000 (not NS-1000M) sealed cabinet. I eliminated/bypassed all the passive LCR network, and all the SP drivers are directly-dedicatedly driven by each of the four amplifiers (including the one for outer super-tweeter FOSTEX T925A).

Have you carefully read through these posts?

- Where in my multichannel multi-driver (multi-way) multi-amplifier stereo system should I measure/check frequency (Fq) Responses? #393

- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains_Part-1_Fq Responses in EKIO’s digital output level: #394

- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains_Part-2_Fq Responses in DAC8PRO’s analog output level: #396

- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains_Part-3_Fq Responses in amplifiers’ SP output level before protection capacitors: #401

- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains_Part-4_Fq Responses in amplifiers’ SP output level after protection capacitors: #402

- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains_Part-5_Fq Responses in actual SP room sound at listening position using one measurement microphone: #403

- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains_Part-6_Summary, discussions, and a little step forward: #404, #405-#409


- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-1_ Precision pulse wave matching method: #493

- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-2_ Energy peak matching method: #494

- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-3_ Precision single sine wave matching method in 0.1 msec accuracy: #504, #507

- Measurement of transient characteristics of Yamaha 30 cm woofer JA-3058 in sealed cabinet and Yamaha active sub-woofer YST-SW1000: #495, #497, #503, #507

- Identification of sound reflecting plane/wall by strong excitation of SP unit and room acoustics: #498

- Again, "Near ultrasound - ultrasound" ultra-high frequency (UHF) noises in HiRes music tracks, and EKIO's XO-EQ configuration to cut-off such noises: #518

- Perfect (0.1 msec precision) time alignment of all the SP drivers greatly contributes to amazing disappearance of SPs, tightness and cleanliness of the sound, and superior 3D sound stage: #520

- Summary of rationales for "on-the-fly (real-time)" conversion of all music tracks (including 1 bit DSD tracks) into 88.2 kHz or 96 kHz PCM format for DSP (XO/EQ) processing: #532

- Not only the precision (0.1 msec level) time alignment over all the SP drivers but also SP facing directions and sound-deadening space behind the SPs plus behind our listening position would be critically important for effective (perfect?) disappearance of speakers: #687

- A nice smooth-jazz album for bass (low Fq) and higher Fq tonality check and tuning: #910, #63(remote thread)

- Again, summary of my rationales and pros of analog-level relative gain (tonality) controls in addition to gain controls in DSP configuration: #911, #317(remote thread), #313(remote thread)

 

[LSPhil]

Thank you for your swift and precise response, delivered in just 600 words and accompanied by 30 references. Your approach focuses on striving for a linear amplitude response, whereas I, as usual, tend to design systems based on the principles of Linkwitz, Butterworth, or Bessel. Well, everyone follows their own chosen path. I would like to emphasise once again the necessity of replacing that bass speaker, as I believe it compromises the entire system. As for your enthusiasm for the beryllium midrange driver, I have good news for you. An even better beryllium driver has been released by Bliesma – the M74B-6.

 

[Holmz]

Thank you for your swift and precise response, delivered in just 600 words and accompanied by 30 references. Your approach focuses on striving for a linear amplitude response, whereas I, as usual, tend to design systems based on the principles of Linkwitz, Butterworth, or Bessel. Well, everyone follows their own chosen path. I would like to emphasise once again the necessity of replacing that bass speaker, as I believe it compromises the entire system. As for your enthusiasm for the beryllium midrange driver, I have good news for you. An even better beryllium driver has been released by Bliesma – the M74B-6.

Is it possible to design a system based upon the output SPL better matching the electrical signal?

 

[dualazmak]

Well, everyone follows their own chosen path. I would like to emphasise once again the necessity of replacing that bass speaker, as I believe it compromises the entire system. As for your enthusiasm for the beryllium midrange driver, I have good news for you. An even better beryllium driver has been released by Bliesma – the M74B-6.

I know very well and understand what you are emphasizing and recommending!

Throughout my present project, however, I strictly dare to limit myself using rather vintage but still wonderful YAMAHA NS-1000 and its SP drivers plus L&R subwoofer YAMAHA YST-SW1000 and supe-tweeters FOSTEX T925A.

I once intensively considered addition and/or replacement of woofer as you can find in my posts;
- Just thinking about possibility and feasibility of adding a pair of woofer (or a pair of completed SP as woofer) in present multi-channel multi-driver (multi-way) multi-amplifier stereo setup: #451, #167(remote thread), #170 (remote thread), #183(remote thread), #455, #459-#461, #462-#466, #469, #31(remote thread), #470, #471, #473, #474, #476, #477, #495

My search for such new woofer(s) (and its cabinet) is still in progress, but if I would change the SP drivers and/or cabinet, I would be better to start a new project thread, I believe.

This project thread is mostly dedicated to my present treasure SP drivers and cabinets in my present listening room acoustic environments.

 

[dualazmak]

Furthermore, my listening room and acoustic environment is of course not anechoic, and not in geometrically symmetrical; I have many random/odd furniture alignments as you can see the photos of my listening room!

My efforts in my multichannel project are for optimization of my music listening enjoyments, and very much strict SOTA precision tunings are not always needed and feasible...

 

[LSPhil]

Is it possible to design a system based upon the output SPL better matching the electrical signal?

Of course you can. Many companies such as Canton, Isophon, KEF, JBL build great constructions that are not only based on linear amplitude characteristics, but also on the concept of frequency division and impulse response of the entire speaker.

 

[dualazmak]

My DSP EKIO's delay (group delay) settings have 0.01 msec granularity/precision, but I am suspicious about whether we need of 0.01 msec tuning in our room acoustic environments with reflections/absorptions by many planes furniture floor ceiling, etc.

I believe room mode treatments would be much more critical factors than 0.01 msec time alignment tuning.

We hear music by our ears and brain, not by our measurement microphone...

 

[Sal1950]

My DSP EKIO's delay (group delay) settings have 0.01 msec granularity/precision, but I am suspicious about whether we need of 0.01 msec tuning in our room acoustic environments with reflections/absorptions by many planes furniture floor ceiling, etc.

But if you can measure and correct to that fine a precision, why not use it and suffer any worry?

 

[Keith_W]

But if you can measure and correct to that fine a precision, why not use it and suffer any worry?

He is right, there is no point time aligning the sub to 0.01ms precision. It is far more important to align the phase of the sub to the woofer - if you fail to do this, you get nasty peaks and dips across the overlap frequencies. That is far more audible than if your sub is 0.01ms out of whack.

 

[LSPhil]

My DSP EKIO's delay (group delay) settings have 0.01 msec granularity/precision, but I am suspicious about whether we need of 0.01 msec tuning in our room acoustic environments with reflections/absorptions by many planes furniture floor ceiling, etc.

I believe room mode treatments would be much more critical factors than 0.01 msec time alignment tuning.

We hear music by our ears and brain, not by our measurement microphone...

So, from what I can see, based on your diagram, you've simply defined all the filters as second-order Linkwitz-Riley without taking into account the amplitude response. You’ve also added a first-order Butterworth filter and come up with something. If this is how it’s done, you'd likely get the lowest passing mark at school, just before failing. Your approach lacks any real concept, apart from the idea that you're doing it because you can, and the amplitude response seems acceptable. I'm quite certain no one designs three-way systems like this.

Whether you correct the time alignment of the speakers or not is another odd matter. One moment, you talk about precise tuning with accuracy down to 0.1ms (+/- 3.5cm), and in another post, you say it doesn't matter because music happens in the mind, only to later mention improving the stereo image. I admire the courage of writing and presenting this system as something noteworthy, but all suggestions are politely ignored and dismissed.

The errors in the crossover are irrelevant, time alignment compensation doesn't matter, and an old, off-centre 40-year-old driver is perfectly fine because beryllium drivers are supposedly so wonderful. I'm impressed.

 

[Sal1950]

He is right, there is no point time aligning the sub to 0.01ms precision. It is far more important to align the phase of the sub to the woofer - if you fail to do this, you get nasty peaks and dips across the overlap frequencies. That is far more audible than if your sub is 0.01ms out of whack.

Sure, and much of this is beyond me, but is it an either or situation?

 

[Keith_W]

Sure, and much of this is beyond me, but is it an either or situation?

It usually is. There are two different approaches to subwoofer time alignment:

- Align the impulse peak of the sub to the impulse peak of the woofer (if you are using linear phase filters) or the start of the sub impulse to the start of the woofer impulse (if using minimum phase).
- Align the phase of the sub with the phase of the woofer.

Both require calculating the delay of the sub relative to the woofer, then delaying one driver (usually the woofer) so that there is alignment. Here is the rub - aligning subwoofer and woofer impulse peaks in time also changes the phase relationship between both drivers. Or to put it another way - you have to choose between a single impulse having some time distortion, vs. a continuous bass tone having misaligned phase and producing peaks and dips in the frequency response.

Lately I have been reading some studies about the audibility of group delay in subwoofers - what happens when the GD of the sub is a lot more than the GD of the woofer (as in the case when subs and woofers are not time aligned)? Obviously, if we are talking about 10 minutes of GD it is audible. But what is the audible threshold? It turns out there are no good studies that show what the threshold is, but it is thought to be about 20-30ms or even more. DSP allows us to time align down to 0.021ms and I could go even lower if I wanted to. But there is little point of doing so if the audibility threshold is 20-30ms. So in my view, it is more important to align the phase and sacrifice alignment of the impulse.

There are of course far more things to consider than this. The room itself distorts phase, and as you know, produces its own peaks and dips. You could design a series of all pass filters to compensate for the phase distortion, but all pass filters are difficult to design. At the moment my approach is to simply lop off the peaks and try to deal with the dips with all pass filters.

Of course I am open to discussion about this because it's my opinion only. And you know what they say about opinions I am happy to change my position if I am proven wrong.

 

[dualazmak]

Whether you correct the time alignment of the speakers or not is another odd matter. One moment, you talk about precise tuning with accuracy down to 0.1ms (+/- 3.5cm), and in another post, you say it doesn't matter because music happens in the mind, only to later mention improving the stereo image.

I believe you still did not yet carefully read and understand my relevant posts!

I have established 1 msec precision time alignment between my L&R subwoofers and woofers (at my LP), and I also established 0.1 msec time alignment between my woofer and midrange (plus tweeter and super-tweeter) at my LP; and I said all of these time alignment tunings , together with room acoustic tunings, do matter for optimization of my hearing/listening sensations/experiences.

You strongly insisted that 0.01 msec precision (at LP?) is indispensable and critical, but for which myself and @Keith_W do not agree at all.

You discussed a lot, but I (we) have not yet seen none of your objective data/diagram nor diagram of your own system, even no SLP house curve measured at your LP. I have shared, on the other hand, a lot of my methods and objective data/diagrams thereof so far.

Just for example, let me share again my SPL house curves measured at my LP while all the SP drivers sing together as follows (I do not like too much smoothing such as psychoacoustic smoothing);
In my post #410 you can find:

In my post #931 (the latest system setup), you can find;

I am very much looking forward to seeing your total audio system setup (including room acoustic environment) and total signal path (hopefully on my independent thread here);
"Let's share diagrams (and photos) of our total physical audio system and the whole signal path, with a few words and/or links"

and, I am also looking forward to seeing your preferred systematic tuning methods in detail and objective data/diagrams measured at your listening position on your own new thread.

 

[dualazmak]

I admire the courage of writing and presenting this system as something noteworthy, but all suggestions are politely ignored and dismissed.

It would be totally "up to you, and your freedom" to ignore and dismiss other people's efforts and info sharing.

I myself still believe, however, it would be worthwhile sharing my efforts and experiences, like just for recent example of "the wide-3D reflective dispersion of super-tweeters' high Fq sound".

Spoiler

- A new series of audio experiments on reflective wide-3D dispersion of super-tweeter sound using random-surface hard-heavy material:
Part-1_ Background, experimental settings, initial preliminary listening tests: #912
Part-2_ Comparison of catalogue specifications of metal horn super-tweeter (ST) FOSTEX T925A and YAMAHA Beryllium dome tweeter (TW) JA-0513; start of intensive listening sessions with wide-3D reflective dispersion of ST sound: #921
Part-3_ Listening evaluation of sound stage (sound image) using excellent-recording-quality lute duet tracks: #926
Part-3.1_ Listening evaluation of sound stage (sound image) using excellent-recording-quality jazz trio album: #927
Part-4_Provisional conclusion to use Case-2 reverse reflective dispersion setting in default daily music listening: #929

 

[Sal1950]

Of course I am open to discussion about this because it's my opinion only. And you know what they say about opinions I am happy to change my position if I am proven wrong.

Thanks Keith_W. Being completely honest I can only begin to clearly grasp the alignment approaches presented.
But then thinking of attempting the phase alignment method with dual or more subs could lead to madness ???

 

[Keith_W]

Thanks Keith_W. Being completely honest I can only begin to clearly grasp the alignment approaches presented.
But then thinking of attempting the phase alignment method with dual or more subs could lead to madness ???

DSP for subwoofers IS a topic that can send you to the asylum for sure. This is why dualazmak and myself are certifiably nuts and even then I wouldn't consider myself an expert on it. There are multiple approaches to getting the smoothest sounding bass in a room, and I have only described ONE way. Here are a few others:

- Geddes multi-sub approach. Requires a minimum of three subwoofers. One goes into a corner, and the other two into non-corner positions. One of these needs to be elevated. The output of the subs are then adjusted to produce the smoothest bass possible. Disadvantages: no time alignment (and therefore potentially audible if the GD of the subs is too high), and the subwoofer may become directional if XO'ed too high or if it produces audible harmonic distortion.

- MSO. Measurements are taken at multiple positions, and the program calculates appropriate phase delay and amplitude adjustment over the listening area (MIMO approach). Advantages: much easier to optimise over a large area than a manual approach since it relies on brute force computation. It's free. Disadvantages: correcting over a large listening area means many subs are required, otherwise each position is compromised.

- DBA and VBA. You know what these are, plenty of threads on ASR from StigErik and others including myself. Disadvantage of DBA: expensive, needs a special room. Disadvantage of VBA: does not work as well as you would hope.

- Room treatment only. Advantages: reduces bass ringing more effectively than DSP (but can not compare to a DBA though). Disadvantages: intrusive.

There may be other schemes I am unaware of. Plenty of clever people out there.

 

[3ll3d00d]

I would think the question being asked is more about the choice of crossovers as designing a speaker involves more than applying textbook filters and then further eq on top. For example, one may consider directivity through the crossover, the driver capabilities as it goes out of band, power handling, filter complexity (if doing anything passively) and probably some other stuff I am not thinking of off hand. Not saying you blindly applied textbook filters (I have no idea how you arrived at that layout) but I think that is the point being made.