[andyr]

As I was sprawled on my couch tonight, listening to a gorgeous ABC-FM broadcast of Bach's "Easter Oratorio" after one too many red wines with dinner, it suddenly occurred to me that manufacturers who carefully step back the mid and tweeter drivers in their 3-way models, to produce a carefully time-aligned setup ... are wasting their time? )   

The reason ... that the mathematics of their time-alignment only work when you have the speakers toed-in so that they are perpendicular to a line pointing towards your ears ... and most owners never have them toed in this much!!??

Am I correct ... or is my reasoning screwed? 

Regards,

Andy

[JLM]

I agree with your logic, but at least it's time aligned along one line (presumably straight out from the front of the speaker).  This assumes that the drivers/crossover are all designed to also be in phase (big assumption).  A few vendors really push all this time/phase alignment stuff (Vandersteen and Green Mountain come immediately to mind).

Some have taken time/phase alignment even more seriously and offer coaxial mid/treble drive units, like KEF or Hawthorne Audio.

And then there are the "extremists".  Come visit the single driver circle, where time/phase coherency, differences in sound between drivers, and crossover issues have all been resolved.   

[andyr]

And then there are the "extremists".  Come visit the single driver circle, where time/phase coherency, differences in sound between drivers, and crossover issues have all been resolved.   

Ahem!    But so-called "single driver" speakers are generally not full-range ... so they need subs to reproduce even 35Hz and super-tweeters to get above 20Khz ... so they are not IMO "single-driver speakers! 

Regards,

Andy

[mcgsxr]

True in many cases andyr (my own included) but in JLM's case, the speakers he uses are darn close to full range, with transmission loaded Fostex.

No comment on the original thread, I lack the math and experience...

[Daryl]

Most crossovers have their own group delay which causes the lower frequencies to be delayed relative to the high frequencies.

Higher order filters cause more group delay than lower order filters.

Normally to compensate Z-axis (travel distance) offset you would use filters of different order for the HP (highpass) and LP (lowpass) sections and optimize their transfer-functions for flat response.

2nd order LP and 4th order HP is common and can compensate 1/4 wavelength or so at the crossover frequency (XO).

In a case where you have more acoustic offset to compensate (if you wanted a higher XO frequency for instance) it becomes more difficult to address (still doable though) and the amount of group delay at XO is increased (relative to one period at the XO frequency) which you may not be willing to accept.

In such a case you may elect to physicly offset/tilt the baffle to compensate some or all of the acoustic offset.

Also a baffle angled back increases excitation of local acoustics and could be the (true) reason for the angled baffle rather than just time alignment.

My MTM (below) uses a somewhat high (3khz) XO for 6.5" woofers and 1" tweeters but the tweeter is mounted in a shallow waveguide (3/4" deep) which means less compensation is necessary in the crossover.



In the case of 1st order crossover speakers like Vandersteen, Theil and Dunlavy it isn't practical to compensate acoustic offset in the crossover due to the wide bandwidth of overlap with 1st order crossovers which would require an outrageous amount of crossover componets for the delay networks.

[andyr]

Most crossovers have their own group delay which causes the lower frequencies to be delayed relative to the high frequencies.

Higher order filters cause more group delay than lower order filters.

...... etc

Daryl, thank you for the erudite discussion on XO phase delay.  But ... so?   

But I think you missed the point of my post ... which is that, only if you listen to time-aligned speakers "straight on" (ie. pointing directly at your ears), will the time-alignment "work".  If you don't toe in the speakers that much then you have destroyed the basis of the time-alignment calcs and you are, in effect, no longer listening to a time-aligned setup! ... and most people, for some reason, don't point their speakers directly at their ears.

Regards,

Andy

[Kevin Haskins]

Time alignment in most cases, is marketing mojo.   As you pointed out, the drivers are at two different vertical locations on a panel.   Euclid as early as 300 BC could tell you why time alignment of two point sources separated in space is impossible for more than one location in space. 

The exception to this is in something like a horn design where you start to get significantly different path lengths due to the length of the horn.   In those cases you need to work on some sort of delay to compensate.   

In normal box speakers though where the relative acoustical off-set is small, I've not found an audible difference as long as it doesn't cause issues in the FR domain.   Of course I use 24dB acoustic filters all the time and don't find that the group delay introduced is really audible either.   Some people in the audio world must think I'm deaf.   

I would like to do a coax design though.   Not so much for time alignment but to provide the advantages in the frequency domain where you don't get off-axis cancellation due to the different path lengths.

[Daryl]

Daryl, thank you for the erudite discussion on XO phase delay.  But ... so?   

But I think you missed the point of my post ... which is that, only if you listen to time-aligned speakers "straight on" (ie. pointing directly at your ears), will the time-alignment "work".  If you don't toe in the speakers that much then you have destroyed the basis of the time-alignment calcs and you are, in effect, no longer listening to a time-aligned setup! ... and most people, for some reason, don't point their speakers directly at their ears.

Regards,

Andy

Hmmm.

The issue with your off-axis speaker is a simple one which I would think you could answer yourself so I thought you were more interested in the bigger picture.

1. If the speaker is intended to be listened to on-axis and you don't toe it in that's your fault (the end user) not the designers.

2. Even if you don't toe them in you won't (shouldn't) be that far off axis and only a small offset should occur.

[*Scotty*]

Daryl,here are a couple of links to some information on Transient perfect crossovers that you might find interesting.
http://www.geocities.com/kreskovs/John1.html
http://www.musicanddesign.com/tech.html
Scotty

[Daryl]

Daryl,here are a couple of links to some information on Transient perfect crossovers that you might find interesting.
http://www.geocities.com/kreskovs/John1.html
http://www.musicanddesign.com/tech.html
Scotty

Hi Scotty,

I talk to John K all the time on the Parts Express board and my post above does address transient perfect considerations.

That is the comments about first order crossovers.

John has (or used to have) also information about TP designs that use non-zero acoustic offsets and others with higher order filters which might make delay networks practical (maybe if the overlap is small enough).

[PaulHilgeman]

*Plug* (Not shameless)

To add to that, keep in mind that much of what we hear is not just direct sound, but reflected sound.  With non-coax speakers, the path length differences are great, as are the cancellations created in the frequency domain.

[Kevin Haskins]

*Plug* (Not shameless)

To add to that, keep in mind that much of what we hear is not just direct sound, but reflected sound.  With non-coax speakers, the path length differences are great, as are the cancellations created in the frequency domain.

Certainly not shameless and easy to verify with measurements.   

[PaulHilgeman]

Check the ONE technical article on my site, the above the tweeter time/frequency responses for a pretty generic monitor sized speaker with good on-axis response.  All real measurements, performed by yours truly.

http://www.nomad-audio.com/article_1.htm

-Paul

[Brian Cheney]

Paul:

Please comment on digital delay lines for each section of the speaker, time compensated for the listening distance.  This is what I do, along with the first order filters and in phase drivers.

[JLM]

Good point Andy.  That's why I'm careful to use the term "extended range driver" not "full range driver".  (It also saves a good many "unfruitful debates".)

As Mark indicated, my single driver speakers are about as close to "full range" that you're likely to find.  And most designers recognize that below 200 Hz, the ear is less sensitive to delay/phase errors anyway.

One of the primary places that audiophiles have "lost their way" is believing that response below 35 Hz or above 20,000 Hz or that spls above 110 dB are essential (or at least more important than the six octaves in the middle reproduced well at more "real world" spls.  Much grief (and marketing successes) have come from such a pursuit.

[*Scotty*]

Daryl, have had any success at applying his transient perfect crossover information to your own DIY designs. I was looking over his Generalized Filler Driver crossover spreadsheet and it certainly looks like you could produce a phase coherent 3way with this approach.
Scotty

[nrenter]

The reason ... that the mathematics of their time-alignment only work when you have the speakers toed-in so that they are perpendicular to a line pointing towards your ears ... and most owners never have them toed in this much!!??

I reject the very premise of this thread. How else can this "logic" be applied? Mp3's are good enough since most owners don't critically anyways? C'mon now...

My entire system is designed around time and phase coherence, and yes, my GMA Callistos are set up such that the drivers point just over my shoulders. Is the sound time and phase coherence outside of this window? No it is not.  However, the sound is never time and phase coherence when listening to 2nd, 3rd or 4th order crossovers. Maybe this matters...maybe it doesn't. To me, it does. However, to say it's not an important design characteristic because (allegedly) most users of time and phase coherence speakers do not properly set up their system is....silly.

[andyr]

Good point Andy.  That's why I'm careful to use the term "extended range driver" not "full range driver".  (It also saves a good many "unfruitful debates".)

As Mark indicated, my single driver speakers are about as close to "full range" that you're likely to find.  And most designers recognize that below 200 Hz, the ear is less sensitive to delay/phase errors anyway.

One of the primary places that audiophiles have "lost their way" is believing that response below 35 Hz or above 20,000 Hz or that spls above 110 dB are essential (or at least more important than the six octaves in the middle reproduced well at more "real world" spls.  Much grief (and marketing successes) have come from such a pursuit.

I agree with you that "extended range driver" is a more appropriate term.  However, I totally disagree with your contention that below 35Hz is not important - nor above 20Khz.  You obviously don't listen much to Bach's organ music if you think that below 35Hz is unimportant.  

Regards,

Andy

[markC]

Good point Andy.  That's why I'm careful to use the term "extended range driver" not "full range driver".  (It also saves a good many "unfruitful debates".)

As Mark indicated, my single driver speakers are about as close to "full range" that you're likely to find.  And most designers recognize that below 200 Hz, the ear is less sensitive to delay/phase errors anyway.

One of the primary places that audiophiles have "lost their way" is believing that response below 35 Hz or above 20,000 Hz or that spls above 110 dB are essential (or at least more important than the six octaves in the middle reproduced well at more "real world" spls.  Much grief (and marketing successes) have come from such a pursuit.

I agree with you that "extended range driver" is a more appropriate term.  However, I totally disagree with your contention that below 35Hz is not important - nor above 20Khz.  You obviously don't listen much to Bach's organ music if you think that below 35Hz is unimportant.  

Regards,

Andy

I don't think that JLM meant that it wasn't important, (don't want to speak for him or anyone else), I just think that the amount of audio information down there is far less than in higher frequencies in most music.

[Daygloworange]

I agree with you that "extended range driver" is a more appropriate term.  However, I totally disagree with your contention that below 35Hz is not important - nor above 20Khz.  You obviously don't listen much to Bach's organ music if you think that below 35Hz is unimportant.  

Regards,

Andy

I think it depends on what you typically listen to. A lot of older existing pop music doesn't typically have a lot of valueable information below 40 hz, with exception of pop music within the last few years, as sub-sonics have become more prevalent.

If you're a hardcore classical buff and listen to solo organ works and/or large orchestras, then sub 40 hz playback becomes increasingly important.

I think for the most part, a system capable of playing solidly in the mid to low 30's is sufficient for the vast majority of listeners.

As far as the importance of having a speaker capable of extending beyond 20 Khz, I'm not entirely sure that the theories of how it add to realism actually play out in reality. I do find it a fascinating area of study.

Cheers