Mr Kirchner from Kirchner Electronic germany described once in a pdf documentation how he time aligns his loudspeakers. My measurement system I bought from him.
He applies to tweeter and midrange driver some pink or white noise at low level and listens to them. He describes its clearly audible to find out like this the best position of the drivers in the depth to each other.
Maybe you protect the tweeter with some bigger cap. He showed also a system with time aligned drivers and 12 and 12 db crossover showing good square waves after time alignment
He applies to tweeter and midrange driver some pink or white noise at low level and listens to them. He describes its clearly audible to find out like this the best position of the drivers in the depth to each other.
Maybe you protect the tweeter with some bigger cap. He showed also a system with time aligned drivers and 12 and 12 db crossover showing good square waves after time alignment
found it 3.3 describes it
https://kirchner-elektronik.de/wp-content/uploads/2019/04/15508573-Photo-Story-OK.pdf
https://kirchner-elektronik.de/wp-content/uploads/2019/04/15508573-Photo-Story-OK.pdf
in this pdf story he shows time alignment by ear
page 10
https://kirchner-elektronik.de/wp-content/uploads/2019/04/15508560-Lautsprecherentwicklung-1.pdf
page 10
https://kirchner-elektronik.de/wp-content/uploads/2019/04/15508560-Lautsprecherentwicklung-1.pdf
Just a quick comment as I don't follow these threads much anymore. Form my point of view, time alignment is the physical alignment of the driver, without crossover, such that the impulse response of each driver arrives at the design point at the same time (or with the same delay). Phase alignment is alignment of the drivers with crossover in place such that the phase difference between drivers, matches the matches the text book phase difference between them for the crossover chosen. For example 90 degrees for a Butterwoth crossover, 0 for a Linkwitz/Riely. Neither has any particular significance with respect to transient accuracy unless the crossover design is one which eliminated phase induced distortion.
By simply wiring a DPDT switch to a 4 conductor 'extension lead', you can sit in your listening position and 'toggle' between phase orientation.
This way you're listening to a whole band of audio.
PS.
I can really see how DSP with Bi or Tri amping is the true solution for so many things.
Yes, you can switch polarity that way but it doesn't change the slope of phase. DSP has nothing to do with it.
Here I show the measured phase (of sound pressure) of a two way crossover in three configurations. Each of these will give you a null at the crossover frequency of 500Hz. They each have text book slopes, each accounts for a different amount of delay and one of them has reversed polarity to the other two.
Here I show the measured phase (of sound pressure) of a two way crossover in three configurations. Each of these will give you a null at the crossover frequency of 500Hz. They each have text book slopes, each accounts for a different amount of delay and one of them has reversed polarity to the other two.
Yes, certain;y they may sound different. Which is better is subjective. But ponder this. Take a pair of drivers, woofer and tweeter. Choose your design axis. Let the drivers be vertically aligned, no left or right offset. Now to align them, time align or acoustic center aligned, whatever you choose to call it, you have basically two methods: 1) offset one driver front to back; 2) add delay using DSP or an all pass filter. On the design axis 1 and 2 are equivalent. Next, consider what happens at 90 degree off axis, left or right. Is the acoustic center of each driver in the same place? One thing is for sure, if the drivers are aligned vertically with no left/right offset when view from the front, there is no physical offset when viewed from 90 degress as it would require left or right offset when viewed from the front. Yet the system using electronic delay would still have the same time delay applied. Thus, regardless of how you align the drivers with reference to the design axis, the alignment will change off axis and the response of the system will be different even when identical on the design axis. Consequently they may sound different.
Well, any filter will add additional (group) delay as well.
So depending how far the two drivers are apart, one could use and higher order filter for the driver that is more close by.
For tweeters one could use a 3rd or 4th order filter. In combination of choosing the right crossover frequency, this could give quite a decent alignment.
This is obviously not as flexible as a delay. But can sometimes work or at least make the issue a lot smaller and therefore less noticeable.
The critical area here is around the crossover point (+/- an octave roughly) . When one of the two is already attenuated too much (because of the slope of the filter) , there won't be any (big) problems anymore when it comes down to relative distance. Or this problem is a whole lot less significant, since there is no destructive interference anymore.
Last edited:
Funny how I thought I was asking a basic question and so many views emerged.
Thanks @GM for attaching the Altec paper! As @Pano said, clear and to the point. And in line with what I was trying to do, but will first run Altec's test to find out the proper alignment (or difference in depth between mounting flanges and hence thickness of the baffle).
Including flush mounting of the tweeter should provide a little improvement and even if it doesn't, it should look better anyways. Will do.
I do digital crossovers, correction with Acourate which uses FIR, so seems I'm headed in a good direction. Aligning physically is meant to minimize the need of digital time correction.
Maybe physical alignment isn't needed if time alignment through FIR filters is going to be implemented. I wonder if the back-wave in my open baffles would also be aligned if the drivers weren't physically aligned.
thanks for the inputs!
Belts and suspenders. (And you also took a long time to add that rather important little tidbit of information.)
The only difference in physical alignment of drivers if you are using FIR filters to time align is the time alignment at extreme off-axis positions (a schematic is shown in the bottom figure to give you a notion as to how that works):
If using IIR filters (i.e., normal crossover filters used in active and passive crossovers without direct phase control), then for each order of the electrical filters used (first order, second order, etc.), there will be an impressed 90 degrees of phase lag per number of filter order on the lower frequency drivers relative to the higher frequency drivers--for each crossover filter "way" used. If using fourth order IIR filters, you will need to insert an additional full period of delay on the upper frequency drivers to achieve time alignment with the lower frequency drivers. So physical alignment using IIR filters isn't very useful.
When using FIR filters, you have more-or-less complete control of the phase separate from amplitude, so physical alignment doesn't mean very much unless there is a large difference in path lengths between the two ways (i.e., a long midrange horn/compression driver, crossing to a direct radiating woofer, etc.). In that case, when using FIR filters, aligning the mouth of the horn to the front edge of the direct radiating woofer(s) will yield the best off-axis time alignment performance,
More on how to do time alignment here for others that may not use Acourate (which is a bit expensive for the task): Using REW to Determine Time Delays Between Drivers
That tutorial makes use of spectrograms to see the time misalignments. You can also use excess group delay plots to read the time misalignments directly.
Chris
Last edited:
Glad you mentioned that. For an open baffle system, you MUST physically align the drivers first. This is because applying delay (via FIR, IIR, etc) will cause the reverse effect behind the speaker compared to in front of it, so while you might improve the alignment as seen from the front you would be making it worse as seen from the rear. Only by moving the acoustic center physically can you balance front and rear alignment, and then you manipulate phase via the crossover and it will have the same effect at front and rear. Note I did not say "align the acoustic centers". For a dipole the plane where the 90 degree off axis position intersects the driver is at the mounting flange, not back at the voice coil or top plate. When I was first getting into dipole systems I did a lot of measurements of this but because one is using the radiation from both the front and rear sides of the cone, the edge of the mounting flange is where the acoustic outputs meet and cancel. This is a nice result, because you can just mount the drivers in the baffle as usual or line up the mounting flanges, and you are done. Easy! But keep in mind that this is not making the relative group delay between drivers zero, only making it the same as seen from the front and rear. This is because the raw driver response itself is "mechanically filtered" by the driver's free air response and this causes the driver to have a phase response that is non flat. The phase response translates to a group delay response, and that will be different for each driver because the phase response is different.
@Cask05 I was trying to make an example of a hypothetical dipole system that does not have the acoustic centers lined up in such a way as to appear the same from the front and rear - if you use delay to try and "fix" that, you help one side and hurt the other. Do you understand the point I am making?
I can say that when you use the proper setup for a dipole system the front wall reflection should "disappear". How do you do that? Like this:
1. The front wall should be a smooth planar, hard surface such as concrete or stone/brick. Nothing else. No TV, no equipment, etc. above about 16" tall.
2. The front and rear frequency response and radiation pattern should be identical across as much of the audio band as possible, at least 100Hz-7kHz.
3. The speakers should be placed towards the listener from the front wall by at least 1m. Farther is OK.
4. Side walls should also be at least 1m farther to the side than the loudspeakers.
Hearing does not work in the same way that a microphone measures pressure. The interference of a reflection that you measure with a microphone is not conveying how you will "hear it" (or not). Due to how the brain interprets signals from your ears, sounds that arrive later by at least 4.5 msec that are identical in tonal balance are perceived as part of with the first arrival - they are NOT perceived as separate sounds by the brain and are not causing any "interference". You can think of this as either suppression (you don't hear the front wall reflection) or that it is reinforcement (the brain combines it with the direct sound and you perceive a single louder sound). This is supported by research into hearing and acoustic perception, etc.
Also, designing and setting up the loudspeaker in this way has the interesting effect that no matter where you are positioned in the listening space (as long as you are at least 1m away from boundaries) the tonal balance remains almost exactly the same. I have this effect in my own systems and it is remarkable - I can walk around to the side of the loudspeakers past 90 degrees until I approach the front wall and the tonal balance remains almost constant. Only the apparent location of the sound source moves around as I move through the null at 90 degrees and am getting mostly rear radiated output reflecting off of the front wall.
I can say that when you use the proper setup for a dipole system the front wall reflection should "disappear". How do you do that? Like this:
1. The front wall should be a smooth planar, hard surface such as concrete or stone/brick. Nothing else. No TV, no equipment, etc. above about 16" tall.
2. The front and rear frequency response and radiation pattern should be identical across as much of the audio band as possible, at least 100Hz-7kHz.
3. The speakers should be placed towards the listener from the front wall by at least 1m. Farther is OK.
4. Side walls should also be at least 1m farther to the side than the loudspeakers.
Hearing does not work in the same way that a microphone measures pressure. The interference of a reflection that you measure with a microphone is not conveying how you will "hear it" (or not). Due to how the brain interprets signals from your ears, sounds that arrive later by at least 4.5 msec that are identical in tonal balance are perceived as part of with the first arrival - they are NOT perceived as separate sounds by the brain and are not causing any "interference". You can think of this as either suppression (you don't hear the front wall reflection) or that it is reinforcement (the brain combines it with the direct sound and you perceive a single louder sound). This is supported by research into hearing and acoustic perception, etc.
Also, designing and setting up the loudspeaker in this way has the interesting effect that no matter where you are positioned in the listening space (as long as you are at least 1m away from boundaries) the tonal balance remains almost exactly the same. I have this effect in my own systems and it is remarkable - I can walk around to the side of the loudspeakers past 90 degrees until I approach the front wall and the tonal balance remains almost constant. Only the apparent location of the sound source moves around as I move through the null at 90 degrees and am getting mostly rear radiated output reflecting off of the front wall.