Time aligned speakers - do they make sense?

[Multicore]

One problem you get into from a practical viewpoint of speaker design is that phase alignment can only work at the crossover point. Away from there, the phase is no longer correct, and you get response ripples. You can hear those. The only mathematically perfect way to correct for time of flight difference is time delay.

Right! It's important to take into account that in the context of loudspeaker design, memoryless signal processing has been the norm. It can be a slow and difficult task, to shift a paradigm.

 

[fpitas]

Right! It's important to take into account that in the context of loudspeaker design, memoryless signal processing has been the norm. It can be a slow and difficult task, to shift a paradigm.

That's one big reason I'd never go analog. The digital time delay is pretty much a necessity in my design.

 

[gnarly]

^I beg your pardon^

^Like a good 'passive' design with 'coherence' in both time/phase domains?^

^So! If I hear you correctly: A 'decent acoustic design' in a passive speaker with low-order filtering does or does not make good sense.^
I think that was the original question about time alignment?

Hi, hey I hope i didn't offend with my take on passives.
I unabashedly think passives leave sound quality on the table for all but the simplest of well designed two-ways, but that's surely not grounds to knowingly offend.

(I'm trying to leave the whole passive vs active thing alone, simply because many folks seen to have an investment in passive that often turns emotional, or akin to religion/politics.)

And yes, i do strongly think a good acoustic design, that is both time and phase aligned makes for highest sound quality.
Like said, for simple 2-ways I think such is reasonably achievable passively. Maybe even simple 3-ways, with considerable xover skill.
(By simple I mean using drivers that have smooth rolloffs and considerable bandwidth overlaps with each other...and SPL is not a priority.)

 

[gnarly]

Like usual, Rod has some interesting things to say on the subject:

Phase, Time and Distortion in Loudspeakers

sound-au.com

Full disclosure: I time align with DSP digital delay. Despite the digital glare, I think it's worth pursuing.

Very good article, I've read it several times before.
Rod is such a great (and generous) resource imo.

As humans, we of course tend to pull out what we agree with....and I'm gonna do that here...
...since his conclusion/opinion directly hits at this thread titles question...Does Time Aligned make sense?


"Conclusions
For what it's worth, I originally started this article not to praise, but to debunk the theory that time alignment is the only way a speaker should ever be designed. Having done the research, run tests, and written the article, I confess that I must agree with many (perhaps even most) of the points made by the time alignment proponents. Mind you, there is still a lot that you will hear and read that is either gross exaggeration or a downright lie, and it can be very difficult to tell the difference unless you know exactly what the real story is.

My overall opinion, based on the research for this article (primarily tests and simulations), is that time alignment is a very good thing, and perhaps all speakers should be designed this way. On the negative side, the offset required to achieve time alignment can lead to diffraction effects that may damage the sound quality far more than the misalignment. A sloped baffle means that you are always listening off axis from the drivers - not by a great deal perhaps, but off axis nonetheless. This conundrum can be resolved, and it has been by several manufacturers, each in their own way."

 

[dualazmak]

You hear every sound all day with the same decline so...

Please allow me responding you as my final message on our discussion on your point.

All of audio systems including SP drivers have more or less some kinds of "tone control" features. If we would strictly accept and implement your policy/philosophy, any audio playback system cannot be existing, I believe.

(And please understand again that my hearing decline in high Fq zone is not so much, but very slightly; about only 1.5 to 2 dB down from the average of people in 35-year old.)

 

[NTK]

You hear every sound all day with the same decline so...

Please allow me responding you as my final message on our discussion on your point.

All of audio systems including SP drivers have more or less some kinds of "tone control" features. If we would strictly accept and implement your policy/philosophy, any audio playback system cannot be existing, I believe.

(And please understand again that my hearing decline in high Fq zone is not so much, but very slightly; about only 1.5 to 2 dB down from the average of people in 35-year old.)

I think you have misunderstood. Let say when you compare the loudspeaker reproduction of a cymbal hit to a real one, both sounds you heard are affected by the same hear loss. If the loudspeaker used to reproduce the cymbal hit is a neutral one, then you'll hear a reproduction that closely matches the timbre of the real cymbal when you heard it, and you don't find the reproduction "colored".

Ask the question another way, do you perceive real cymbals hits becoming duller and duller as you grow older and older? The live sound you heard everyday are similarly affected by high frequency hearing loss that progresses with age.

 

[dualazmak]

So, do you mean all of the tone (relative gain) control functionalities in many "pre-amplifiers" and "integrated amplifiers" should be abandoned??

If your answer is "no", then what would be your justifications if people apply a little bit of tone control (relative gain control) based on preference (or slight hearing decline in low FQ and/or high Fq) in his/her audio system?

Should we always invite cymbal player with his/her cymbal to our listening room for our comparative listening to live performance vs. recorded sound of "such" live performance in our audio listening room?

Then, we will be highly possibly going into a kind of endless spiral discussions.... Please let me retire from this discussion, therefore, at least at present on this thread.

If needed, please start new thread on this issue.

 

[fpitas]

Very good article, I've read it several times before.
Rod is such a great (and generous) resource imo.

As humans, we of course tend to pull out what we agree with....and I'm gonna do that here...
...since his conclusion/opinion directly hits at this thread titles question...Does Time Aligned make sense?


"Conclusions
For what it's worth, I originally started this article not to praise, but to debunk the theory that time alignment is the only way a speaker should ever be designed. Having done the research, run tests, and written the article, I confess that I must agree with many (perhaps even most) of the points made by the time alignment proponents. Mind you, there is still a lot that you will hear and read that is either gross exaggeration or a downright lie, and it can be very difficult to tell the difference unless you know exactly what the real story is.

My overall opinion, based on the research for this article (primarily tests and simulations), is that time alignment is a very good thing, and perhaps all speakers should be designed this way. On the negative side, the offset required to achieve time alignment can lead to diffraction effects that may damage the sound quality far more than the misalignment. A sloped baffle means that you are always listening off axis from the drivers - not by a great deal perhaps, but off axis nonetheless. This conundrum can be resolved, and it has been by several manufacturers, each in their own way."

Obviously, I agree with him, too. Although I realize I have the luxury of digital delay, which I have already paid for in my DSP unit.

 

[RayDunzl]

I have HF loss, but my personal testing says the cutoff is more like "brick wall" than declining perception, so there is no use to try to hear the highs via EQ.

I do tune the system, to "flat", and allow the musical sources to come across with the speakers producing the levels of the source accurately (as I see it), verified by comparing the peak RTA of the raw source to the RTA of the speaker output at the listening position.

Nobody complains, so it must be "good enough".

As for "time alignment", it's a two-way, crossed at 180Hz, the panels throw a flat phase to the listening position. If I read the phase correctly on REW, the woofers lag a little even with the DSP phase correction applied.

Did a little test, with DSP applied, to try to see the alignment of the highs with the lows. Oddly, it seemed to have a 2 millisecond delay of the highs relative to the lows in this little experiment.

 

[Dumdum]

Heh, I can say I did precisely as you suggest…except I failed in any attempt to use FIR…2x4 hd has limited available taps, and I was trying to align under 500 hz between twin 15” woofers and a fullrange. as I recall there are only 1200 taps available per channel…and I remember convincing myself I can’t get there from here.

A quick one, you can add and remove taps from each individual channel within a total number of taps and get very low if you want, you dont need as many taps to correct the top half of a two way setup vs the bottom half… I seem to recall being able to correct to below 150hz and still have enough taps for the upper half of the crossover

 

[gnarly]

Yep, it's cool the 2x4HD lets its 4096 taps per channel be assigned.
It's unfortunate that there is a 2048 max per single channel, though.
And for low frequency work, it's doubly unfortunate its FIR must run at 96kHz, (which is only 1/2 as effective as 48kHz)

If using the max 2048 taps for linear phase work, where impulse is centered, frequency resolution is only about 94 Hz. (1024 effective taps @ 96k)
A rule of thumb is a FIR filter begins to be decently effective at 3X its frequency resolution. So the 2x4HD's FIR really has only a very smoothed over effect below about 280Hz.

If using the max 2048 taps for IIR replication, where impulse is at the start, freq resolution would double to about 47Hz.
But doing that makes no sense compared to simply using IIR filters to begin with, which have near infinite freq resolution.

Imo, the 2x4HD FIR is designed for 2-ways using a xover say anywhere from 500Hz up. It's not worth a dang for sub integration.

 

[MAB]

Maybe, your point would be related to our personal preferences which may vary individual to individual.

I still would like to clearly hear the "sharp triangle sound" in 4th movement of Brahms Symphony number 4, and also would like hear the extremely high energy high Fq clear sound of Bimmel Bolle Antique Orgel (ref. here and here; Track-20).

I periodically check my hearing ability by using nice headphone and audiometer software, and I know my ability is still much better than the average of my age group. I set a little bit of upward slope in around 8 kHz to 20 kHz (ref. here), however, just for my better listening sensation (maybe simulating my memorized sensation of 20 years ago).

That is it, and I essentially do not care any negative theoretical discussion at least on my "slightly upward Fq response curve in 8 kHz to 20 kHz".

It sounds like you like ‘bright’.
That’s not a preference, that’s what you like personally. Most people don’t like that, I know I don’t like that frequency response. Just saying that what we like is way different than a preference.

 

[norman bates]

 

[NIN]

Please allow me responding you as my final message on our discussion on your point.

All of audio systems including SP drivers have more or less some kinds of "tone control" features. If we would strictly accept and implement your policy/philosophy, any audio playback system cannot be existing, I believe.

(And please understand again that my hearing decline in high Fq zone is not so much, but very slightly; about only 1.5 to 2 dB down from the average of people in 35-year old.)

No, it have nothing to do with that.
If you hear your wife everyday with the same decline (lets say 2dB down at 12Khz). If you would EQ up that decline your wife voice would not sound "more right", because it will sound different from what you hear every day.
Same with music. If you hear instruments with 2dB decline at lets say 12Khz, that is your reference. To reproduce the music with a boost at where your decline are it will not sound like the reference you have.

Off course you can prefer it, but to reproduce something like you hear it IRL, it will not be closer.

 

[pseudoid]

Same with music. If you hear instruments with 2dB decline at lets say 12Khz, that is your reference.

Is it possible to directly compare the conclusion of your argument to those afflicted with cataracts and the surgery-correction there upon?
[Replace music (and/or 'wife') w/cataracts and replace decline in hearing, w/decline in vision.]
Which, then, is the correct reference: Pre-surgery decline or the post-correction (+2db/12kHz) improvement?

 

[dualazmak]

It sounds like you like ‘bright’.
That’s not a preference, that’s what you like personally. Most people don’t like that, I know I don’t like that frequency response. Just saying that what we like is way different than a preference.

No! not so much bright as you are guessing, I believe, since my present Fq response "house curve" is shown in this repeatedly shared diagram;

More than 95 % of about 25,000 tracks of my music library can be well covered by 20 Hz to 7 kHz Fq response where my highly efficient metal horn super-tweeter (Fostex T925A) works only very very faintly. And as you can see above "house curve", I have rather high gains in sub-woofer and woofer coverage compared to midrange-squawker and tweeter.

Consequently, the total sound (and the house curve thereof) is rather warm and gentle side in my setup for daily ordinary music listening sessions; my friends visiting my listening room (many of them are professional or semi-professional classical and jazz musicians) also agree with me on "rather warm side total sound" of my audio setup.

Then, only when I play the high Fq "reference/sampler" track (such as ref. here and here with video clip), all of the guests were/are much impressed by the excellent stunning reproduction (high S/N, very low distortion) of the sound of Bimmel Bolle Antique Orgel.

As shared here, I usually offer them "flexible control of the gain for super-tweeter by themselves on-the-fly" while listening to the quiet but really high S/N violin "reference/sampler track" which contains various harmonics tones up to almost 22 kHz. It is always my interest what Fq slope(s) they would prefer in 7 kHz to 20 kHz (it always varies depending on person to person, but usually within plus/minus 3 dB range). (Or more specifically, plus/minus 3dB at 20 kHz from the my standard linear slope of ST Gain-B in above diagram.)

They also can have similar stunning(?) experience of low Fq sound reproduction with my present audio setup (ref. here), as I recently shared here;
- Reproduction and listening/hearing/feeling sensations to 16 Hz (organ) sound with my DSP-based multichannel multi-SP-driver multi-amplifier fully active stereo audio system having big-heavy active L&R sub-woofers: #782