Time-Alignment for Folded Horns (Patent)

[D-MAN]

This is actually a reasonable idea for those of you concerned with the time (phase) alignment issues of high frequencies vs. low frequencies in folded bass horns. Also pretty easily verified and tested by experimentation without permanent harm.

The principle is that non-expanding columns do not change the waveform. PWK used this approach in the La Scala. Valid in theory. Assumes full radial turns.

Patent cover page follows...

DM

[D-MAN]

Here's the patent cover page.

DM

[sunnysal]

since pwk thought of horns as "air columns" it stands to reason that the ultimate expression of this would be a non-expanding tube...the air pulse at one end would instantly cause the same pulse at the other end of the air column...there is no wave...the column moves as one...it is the interface between the end of the air column and the surrounding air the causes a wave to be launched into the open space...this interface can be tapered long or short and tapered in a variety of ways all afecting the efficiency, response, etc. of this passing of the wave into open space...

time alignment issues between two different driver/horn combinations is a very real and easily measured phenomenon...I am not sure how it relates to these patents though...

tony

[PrestonTom]

I am curious on what is actually going on here.

Transmitting sound through a tube has big consequences. If it is flared, then you get the benefit of a loaded design (horn). On the other hand even if it non-expanding, a tube will introduce resonances (nodes & anti-nodes). This is not acoustically neutral. Additionally when the wavefront leaves the mouth of the tube, all sorts of things can happen. I can not beleive that even a "simple tube" will not impact the sound

-Tom

[dragonfyr]

I would agree with Tom.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

In a horn, when a wave front encounters a discontinuity in the rate of flare expansion, a reflection occurs, resulting in resonance. This is exactly the same concept behind the Helmholtz resonator. And all horns exhibit these discontinuities to lesser or greater extents.

This would potentially contribute significantly to the coloration of the horn corresponding to the increased throat length

These discontinuities result in a portion of the initial acoustic energy being reflected back to the compression driver. Here this energy is both partially absorbed and partially reflected yet again with a phase delay ranging upwards of several milliseconds. Increasing the length of the throat will have the affect of increasing the delay perhaps significantly. These subsequent reflections will then establish a tuned resonance as well as become the source of expressed acoustic impulses that are delayed in time from the initial signal.

The lowest octaves of the horn's usable range would be most affected due to the lower frequencies being more strongly reinforced and reflected (as opposed to the higher frequencies). Thus, the path length and the resulting reflections will cause additional horn frequency response coloration due to the resonance and superposition of wave forms that are directly related to the signal path and horn throat length.

[DrWho]

Couldn't the flare on a HF horn just be continued long enough so as to put the driver itself far enough back to match the MF and perhaps LF drivers? (ignoring size of course)

[sfogg]

Tony,

" it stands to reason that the ultimate expression of this would be a non-expanding tube...the air pulse at one end would instantly cause the same pulse at the other end of the air column...there is no wave...the column moves as one.."

Air is compressible and it is a pressure wave moving through the air that causes sound... not the air moving itself. Besides, if the pulse instantly came out the other end of the column then this wouldn't make much of a delay would it?

Shawn

[sfogg]

Way back 30 years or so there was an article about building a DIY delay line in Audio Amature. It was basically putting a tiny driver in one end of tubing and a mic. in the other end. For every foot of length of the tube you gained about 1ms of delay.

Digital makes some things so much easier....

Shawn

[WMcD]

I recall that in a Dope From Hope PWK described an experiment where he adjusted time delay by the use of a tube. Midrange probably.

It was necessary to use an equalizer to get the response back to that of a normal set up. Once that was done no one could hear a difference.

Gil

[DrWho]

----------------

On 8/20/2005 3:38:42 PM sfogg wrote:

Digital makes some things so much easier....

----------------

so true...now to find a processor that supports every decoding fromat and tri-amping with digital delay filters...

[Al Klappenberger]

Guys,

I hate to be a pest, but extreme-slope crossover networks make time

alignment unnecessary! If only one driver is playing each sound there

is nothing to align! All that's left is phase errors between the

different sounds and you ear / brain is totally deaf to that!

Al K.

[DrWho]

Guys,

I hate to be a pest, but extreme-slope crossover networks make time alignment unnecessary! If only one driver is playing each sound there is nothing to align! All that's left is phase errors between the different sounds and you ear / brain is totally deaf to that!

I don't think a steep slope crossover is going to have much affect on the time delay, especially with khorns. If you had a single sound where all frequencies were supposed to be simultaneous (like a kick drum). The 1kHz snap would arrive first followed by a short silence and then the boom from the bass horn comes way late. The khorn is a great speaker, but I cannot get over that aspect of the sound and feel that time correction would make a world of difference. Making the slopes steeper is only going to affect the crossover region...the effect that I feel is so blatantly apparent is well outside the crossover region and therefore a crossover mod isn't going to help. I have talked with others on the forum that have noticed the same thing, but they just never post about it because the khorn seems to be a "perfect" speaker to many on the forum.

I would love to take some pulse signal measurements to show this effect better (perhaps 3 different simultaneous test tones centered in the middle of each driver's frequency range). If nothing else, to prove to myself that I'm not crazy (or perhaps that I am) []

[Al Klappenberger]

DrWho,

You are correct that a filter has virtually no effect on the time dealy. It's virtuall all propogation time difference from the driver to you ear. The delay difference is not as noticable as you would think. What does seem to be noticable is when you hear a single sound twice. The classice example is the clicks from a tap dancers heels. You hear it first for one driver then from the other. If each component of a complex wavefore is only poduced by a single driver, you hear everything only once. The only thing left is that the components will arrive a slightly different times from each driver. This does not seem to be a problem unless the difference is large enough to be greater than the brain can process two sounds. That will start to happen only when the errors is on the order of many feet. PWK showed that the difference in propogation time between the Khorn woofer and squawker is below this limit. What I noticed is that replacing a conventional squawker / tweeter crossover with an extreme-slope network removed a smearing sound that I had been living with for years. I believe it was becasue I was hearing everyting twice, first from the tweeter and then from the squawker. The actual phase errors casue by the different path lengths of the waveform components is not audable becasue they are only phase errors between DIFFERENT things. When two drives produce the same components you have phase errors between two of the SAME components. That causes trouble.

Al K.

[Deang]

Doesn't hearing everything twice mean it's twice as good?

[DrWho]

DrWho,

You are correct that a filter has virtually no effect on the time dealy. It's virtuall all propogation time difference from the driver to you ear. The delay difference is not as noticable as you would think. What does seem to be noticable is when you hear a single sound twice. The classice example is the clicks from a tap dancers heels. You hear it first for one driver then from the other. If each component of a complex wavefore is only poduced by a single driver, you hear everything only once. The only thing left is that the components will arrive a slightly different times from each driver. This does not seem to be a problem unless the difference is large enough to be greater than the brain can process two sounds. That will start to happen only when the errors is on the order of many feet. PWK showed that the difference in propogation time between the Khorn woofer and squawker is below this limit. What I noticed is that replacing a conventional squawker / tweeter crossover with an extreme-slope network removed a smearing sound that I had been living with for years. I believe it was becasue I was hearing everyting twice, first from the tweeter and then from the squawker. The actual phase errors casue by the different path lengths of the waveform components is not audable becasue they are only phase errors between DIFFERENT things. When two drives produce the same components you have phase errors between two of the SAME components. That causes trouble.

Al K.

the localizability of frequencies below 80Hz is also supposed to be inaudible, yet many can easily distinguish the origin of the sound. Just because it is supposed to be unhearable doesn't mean that it is...

We deal with time aligning subwoofers all the time with PA systems (and I usually crossover steeply around 70Hz) so perhaps I'm a bit more sensitive to time delay with the bass. I seriously think you would be able to hear the difference if we could do an AB with a processor implementing and not implementing time delay between the LF and MHF sections. Once you've heard it you can't go back.

[sfogg]

"the localizability of frequencies below 80Hz is also supposed to be inaudible, yet many can easily distinguish the origin of the sound."

Many of those times though it is because the woofer is acting up outside of its pass band. A woofer distorting a 60hz tone now suddenly has output at 120hz or even 180hz (2 and 3 harmonic perhaps higher) for example. A chuffing port is a dead give away too.

But you are correct that even considering that there are perception differences between mono and stereo bass below 80hz but they are different then just localization.

"and I usually crossover steeply around 70Hz)"

How steep? Al is talking about crossovers with initial slopes of around 150db/octave... 25th order crossover. That is far steeper then most crossovers out there.

" I seriously think you would be able to hear the difference if we could do an AB with a processor implementing and not implementing time delay between the LF and MHF sections."

ABX is better. At some point I will probably be trying something like this though for my own curiosity (and cause I'm messing with my system). I don't think I can do the test at as steep of a slope as Al's ES networks though.

I have played with delay WRT subwoofers but the main difference I notice there is simply in the crossover blend. But again that is with much shallower slope crossovers... only fourth order.

Shawn

[D-MAN]

Stereo pairs of low frequency speakers on a full-range system certainly can produce directional and depth queues for low bass frequencies.

The idea of low frequencies being non-directional (or omni-directional)comes from the subwoofer manufacturers and sales staff selling them to unsuspecting consumers.

It is true for ALL MONO channel speakers, regardless of frequency...

DM

[sfogg]

D-Man,

" Stereo pairs of low frequency speakers on a full-range system certainly can produce directional and depth queues for low bass frequencies. "

I agree, I have been running stereo subs for many years because of this and use processing to take what is often mono bass in the mix out to stereo.

It isn't exactly localization though. It is more a sense of spaciousness, and bass externalization (out of your head) and envelopment that stereo bass response gives compared to mono bass response.

There is a lot of information about this here:

http://world.std.com/~griesngr/

Check out the presentations at the very bottom of that page for example though that information is in multiple papers on that site.

Shawn

[DrWho]

I know I made a loose argument for the localization of low frequencies and I wasn't intending to defend it either...the point I was trying to make is just because someone claims a certain trait is inaudible doesn't mean that it actually is. There are all sorts of "facts" floating around which are true for general listening, but not so much for critical listening. PWK had every incentive to claim that the time delay made no difference, just like subwoofer manufacturers have incentive to claim low bass is unlocalizable. That said, I can see why PWK would claim that it doesn't make a big difference, but that's because his time domain sacrifice isn't as huge as the benefit gained by the horn loading. Like my sig, it's all about compromise.

Btw, we were generally using 24dB/octave filters in the PA applications....always as steep as we could go depending on the equipment placed in my finger tips. I believe there was a dbx digital processor that had up to 96dB/octave slopes? I forget exactly, but we still had to dial in the time delay. A PA system is hardly ever a compromise free environment either...

[D-MAN]

Shawn, cool site.

DM