Ever wonder how "big" is a Klipschorn?
I forever despaired of ever time-aligning my diverse electrostatic speaker system... esp with a folded corner horn sitting two feet behind the right ESLs. The front of the Klipschorn is about of 93 inches from my cranium when I sit in my music chair.
Turns out, good old REW measures distance with impulse displays, simple as could be (OK, not so simple).
DATA POINT
Turns out, REW says the Klipschorn sound originates 28.0 feet from my head location (AKA 24.91 milliseconds or 8.54 m). Subtract 93 inches, and that makes the Klipschorn 20.25 feet long. A bit longer than Klipsch and his carpenters figure.
I wouldn't take any of these assertions too seriously since I am new to impulses on REW, but provides some ideas.
BTW, I aligned the mids and the tweeters to that 28.0 feet and listened, as always first to percussion music (Paul Lansky). My impression is that there was a lot more visceral feeling. You could feel the impacts on the percussion instruments. A very nice improvement to reproduction. Otherwise sound was good, but I am very cautious about posting vague impressions here (I must be the only one to behave that way).
Ben
I forever despaired of ever time-aligning my diverse electrostatic speaker system... esp with a folded corner horn sitting two feet behind the right ESLs. The front of the Klipschorn is about of 93 inches from my cranium when I sit in my music chair.
Turns out, good old REW measures distance with impulse displays, simple as could be (OK, not so simple).
DATA POINT
Turns out, REW says the Klipschorn sound originates 28.0 feet from my head location (AKA 24.91 milliseconds or 8.54 m). Subtract 93 inches, and that makes the Klipschorn 20.25 feet long. A bit longer than Klipsch and his carpenters figure.
I wouldn't take any of these assertions too seriously since I am new to impulses on REW, but provides some ideas.
BTW, I aligned the mids and the tweeters to that 28.0 feet and listened, as always first to percussion music (Paul Lansky). My impression is that there was a lot more visceral feeling. You could feel the impacts on the percussion instruments. A very nice improvement to reproduction. Otherwise sound was good, but I am very cautious about posting vague impressions here (I must be the only one to behave that way).
Ben
Last edited:
Ben,
The folded horn leading the mids and the tweeters by about 20ms will give a more visceral feeling due to the Haas effect.
Have you tried measuring distance with the Klipschorn using no filters?
Art
"Filters".... what's a filter?
Not sure how the Haas Effect, as usually understood, relates here (110 Hz, sharp crossover)?
Appreciate hearing from someone with your large experience in this topic. Thanks.
Ben
Filters are either electronic or passive, as in coils and capacitors, or electronic crossovers which can use RC networks or digital equivalents. A long horn is also a low pass filter. If the REW delay finder is like the one in Smaart (the program I use) it can be "confused" by in line filters, as it reads off the highest frequencies, which normally arrive first in a loudspeaker. The room modal reflection from the Khorn may be louder where your mic is positioned than the initial wave, so REW reads that as the first arrival, so it includes the horn path length plus a room reflection added in. If you moved the mic to the horn mouth, (and inches from the mid/high speakers) it would probably read the correct delay times, which would be the path length from the acoustic center plus the latency from any processing or filters in line. Then you can use simple math to figure path length differences between the components to figure the correct delays required for the various components to time align at the listening position.
That said, each component of your system has a unique phase response with frequency, the tricky part is getting the phase to match through the crossover region. A 24 dB per octave filter has a much smaller crossover region than a 6 dB per octave filter, so phase problems are less apparent.
Normally, the low frequency range of a loudspeaker lags behind the transient, you have reversed that by delaying the HF more than the actual "time of flight" from the Khorn.
The Haas Effect, simply put, is what we hear first, sounds loudest. Below 110 Hz is where visceral impact "lives", if that range leads the HF transient, it sounds more "gutsy".
Art
Last edited:
You might try this. Use basic time of flight in the horn to your ear vs time of flight from the electrostatics to your ear as an initial delay guess. In REW set tone generator to the XO freq and flip the phase of the HF top. Play the tone and slowly adjust delay until the sound hits a minimum, usually 10dB less than max. There may be many minimums but the strongest one closest to your initial delay guess I probably the correct one. When the sound is at minimum the time alignment is overlapped. Now flip phase of HF top back to normal and it will be loud. This should get you close. If you are off by an entire cycle you may think it is time aligned but it will be audibly off when listening to percussion.
REW has an impulse display assessment that times from the loop-back signal to arrival of the speaker sound. Neato. You can also eyeball the situation graphed to make sure REW agrees with your own assessment. Moreover, the figures I presented are coherent in simple tape-measured terms. So I don't think there are any artifacts arising from reflections arriving first.
The arrival times were synchronized among the Klipschorn, Dayton-Wright mids, and Dennesen tweets. Sounds good. Thanks to the miracle of DSP from Behringer DCX2496.
Like EQ in a small room, getting phase spot-on is hopeless and, as you say, needless at 24 and more dB/8ave.
I don't share your views of Haas or impact. Haas Effect relates to where your ear sees the sound originate in horizontal position (that is, which speaker box seems to be the origin because it speakers first), not where you "see" it in frequency.
Aside from resonance of my sternum or internal organs per se which are low-freq matters, the visceral impact arises from also having high power capacity in tweeters and mids, not from a "fast bass." The mids AND tweets provide a lot of that newly enhanced visceral impact from my newly synchronized speakers, just as percussion instruments do in live performance. (I think a lot of systems are deficient in tweeter power capacity.)
Previously the sound initiation of the Klipschorn was 22 feet behind the rest of the system!!!
Ben
Last edited:
Ben,
You could test your hypothesis easily by simply checking the impulse display assessment at the horn mouth compared to the listening position and compare it to the tape measure distance difference.
Obviously, the Klipschorn path length is not 20.25 feet long, so something is off.
The DSP I use has an inherent latency of just under 2ms, what is the latency of the Behringer DCX2496?
Art
Ben's set up uses a Klipschorn, a corner horn with about 6.88 foot path length, around a 6 ms time of flight from driver to horn mouth. In addition, it is 2 feet behind one of the ESLs, so a minimum of about 8ms upper delay would be expected.
Attachments
Are you referring to generic "bass bins" or to a folded corner horn made by Klipsch?
While your remark that "2. Someone has..." is wonderfully oracular, could you be more specific about who misunderstood what thing in your opinion?
Thank you xrk971. Your procedure is clearly a good next operational move and smart measurement technique. Granted we are talking of a few ms of error improving the synchronizations (1) with 10 foot waves at crossover and (2) one-meter square curved ESL panels with wide dispersion synchronizing with a refrigerator-sized double-mouth folded horn at an odd angle to the ESL panels... as compared to previously sounding great with 20 feet of time error.
But why use phase inversion if using a mic to assess?
Ben
Ben,
We are talking about around 12ms error, equivalent to more than a wavelength at crossover.
xrk971's suggested polarity (not phase) swap "null test" would be unlikely to work at your listening position.
Art
It depends...
Polarity switches used to be called "phase reverse" switches, but phase and polarity are different, so there has been a move towards calling them the correct thing.
This article explains the differences:
http://www.theatresupply.com/Data/Documents/POL_PHASE_TECH.pdf
At any rate, Ben's room, speaker locations and type would make a polarity swap "null test" difficult to interpret.
Tried it and concluded I can't do it because across the crossover band, the phases or each driver are constant changing (and the loudness of each driver too) due to the crossover filter as well as the way phase spins with distance from the speakers according to frequency, not to mention a bunch of other acoustic factors. Cancellation depends on absolute synchrony of absolute phases. So, unlike in those nice textbook pictures, unlikely to find a freq that does the trick.
Although I don't have access to 1/3 octave filtered noise these days, I'm just guessing that that may be a way to do this type of test using a band centred on the crossover frequency. That gives you a stochastic solution which is as suitable as a definitive solution, if there were one in practice.
Ben
Last edited:
You can also do this stochastically by using the pink noise source option on the generator and use the RTA screen to view the spectrum at the crossover. There may be problems as Weltersys mentioned having to do with the fact that the speakers are far apart and you have room effects. Normally this is done to align drivers from same speaker in near field where phase is more predictable and not going through wierdness.
Thanks for that suggestion about pink noise.
For sure, my large curved dipole speakers and folded corner horn are a complicated geometry to address. Certainly as compared to the usual kind of speaker box. Perhaps this is an example of how audiophiles pursuing one set of virtues in their systems have to compromise on other virtues.
My previous post was a critique of alignment by working at the crossover region. Or more specifically, at the crossover "point." Anybody who has eyeballed the spectrum of a square wave or percussion strike, knows the range of frequencies involved. Good alignment means good alignment across the drivers' whole bands, with the crossover region just a bad jumble of competing waves.
Ben
Hi Ben, here's a pretty well regarded guys take on alignment...
Phase Alignment of Subs – Why I don’t use the impulse response | Bob McCarthy's Blog
Mark
PS I've never had any luck measuring delay indoors, unless I did as Art suggested...mic right at the mouth / speaker.
Phase Alignment of Subs – Why I don’t use the impulse response | Bob McCarthy's Blog
Mark
PS I've never had any luck measuring delay indoors, unless I did as Art suggested...mic right at the mouth / speaker.
Thanks for interesting link; will read shortly.
About "mouth" - which mouth of my Klipschorn do you mean? The one 9 feet away or 7 feet away? They'd only be equal if I sat 45-degrees to the corner. Not to mention there is also 3 feet of height to either mouth resulting in another variation in distance to my head.
As a method, it involves adding delay at the mouth and kind-of acoustic inches from wherever I think the mouth is to my chair and doing so for each of the drivers. Maybe that's the best way, but something of a house of cards.
Maybe the simplest method is just a two-beam oscilloscope showing the electrical signal and the acoustic output and just measuring the time difference between.
Now where did I put my scope?
To answer an earlier question from Art, the Behringer DCX2496 DSP device has pretty negligible delay, The worst of the speaker circuits I looked at was under 1/4 millisecond, in the unlikely event my REW technique is correct. So not a material factor in measuring distances.
Ben
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.