I have a 3 way speaker sitting on the sidelines and the mid and tweeter are on sleds so that I can individually position fore/aft. I have Omni Mic and DATS measurement tools. Is there a way I can align and make meaningful measurements with the tools I have? Im trying to learn about this and was thinking how does a manufacturer do it? Can I do something (even basic) to test this?
I thought people put in a "square" wave and then try to physically align the drivers (by moving your sleds fore and aft) to make the best looking "leading" edge on the waveform picked up by the mic. Unsure if DATS has an oscilloscope function but REW does.
Remember that "in phase" and "time aligned" is not the same thing. Its that latter you want... this means that the suggestion in post #2 may end up with a nice square wave but still not time aligned....
//
//
For any 2 drivers with centre-centre greater than a quarter wavelength, time arrival only becomes possible on specific axes.
dave
dave
OK so is the juice worth the squeeze? I see products with sloped baffles, Wilson speakers with fancy adjustments etc. Even if it is hard to justify in the final product it would be nice to know the measurement method/process.
So when I read this it seems I may be able to do it for the woofer-mid but not the tweeter-Mid?
Feeling remorse for throwing the "did you search" comment I tried myself and came up with:
- https://www.diyaudio.com/community/...using-excess-group-delay.201381/#post-2798136
- https://www.diyaudio.com/community/threads/help-with-time-aling-speaker-drivers.253133/#post-3860175
- https://www.diyaudio.com/community/threads/what-is-time-alignment.316669/#post-5290259
there is probably much more...
//
- https://www.diyaudio.com/community/...using-excess-group-delay.201381/#post-2798136
- https://www.diyaudio.com/community/threads/help-with-time-aling-speaker-drivers.253133/#post-3860175
- https://www.diyaudio.com/community/threads/what-is-time-alignment.316669/#post-5290259
there is probably much more...
//
Its quite an advance reply and if you are about to learn about phase/time aligning, you did probably not catch the inner detail of that. What it says is that you may in some situations not get time alignment in all directions of the speaker i.e. perhaps only "on axis" (straight in front of the speaker).
But if you did, it will depend on the geometry/dimensions of you speaker. Xover frequencies? Care to share?
//
Yes, it is possible to do this with the tools you have on hand, and yes it is worth doing the measurements if you intend to try it. With Omnimic, you will need to use the '3 measurement method' to determine the Z offset between drivers. You can find a good tutorial here:
https://audiojudgement.com/speaker-acoustic-center/
If you instead wanted to move drivers back and forth to align them, I assume you would first measure all drivers physically in line to determine that 'baseline' flat-baffle z offset in mm. Then you could shift drivers accordingly, remeasure, and iterate from there to dial it in further. Obviously moving the drivers on sleds will introduce reflections, diffraction, and other confounds, but I think this would be the general workflow.
Whether it's worth physically time aligning drivers (as opposed to aligning them via delay/z-offset in crossover) by various methods is a bigger question that I don't know the answer to, in general terms or in your specific case.
Tweeter/mid you will need to choose a listening axis, and then lock your heard there.
dave
^ Vertical axis that is, to expand a little 🙂
Horizontal headlocking is needed if the speaker narrows its response, this happens usually at treble and the worse the bigger the (direct radiating) treble transducer is, could be beaming horn as well.
For vertical "listening window" one can calculate path length difference from both transducers to ear and see how much timing changes. The higher the crossover the more for same positional change. As said about 1/4 wl path length difference would still combine constructively, 90 degree phase difference. If crossover was at 3.4kHz, which is 10cm long so 2.5cm or 1" is 1/4 wavelength. This is quite a lot wiggle room still given the transducers are not too far apart. At 3m listening distance one can move a lot before this is exceeded, so normal seated situation is no problem. Too lazy to calculate it accurately atm 😀 Basically what this is is the vertical "main lobe", easy to see and adjust with "spinorama" measurement set in VituixCAD.
Horizontal headlocking is needed if the speaker narrows its response, this happens usually at treble and the worse the bigger the (direct radiating) treble transducer is, could be beaming horn as well.
For vertical "listening window" one can calculate path length difference from both transducers to ear and see how much timing changes. The higher the crossover the more for same positional change. As said about 1/4 wl path length difference would still combine constructively, 90 degree phase difference. If crossover was at 3.4kHz, which is 10cm long so 2.5cm or 1" is 1/4 wavelength. This is quite a lot wiggle room still given the transducers are not too far apart. At 3m listening distance one can move a lot before this is exceeded, so normal seated situation is no problem. Too lazy to calculate it accurately atm 😀 Basically what this is is the vertical "main lobe", easy to see and adjust with "spinorama" measurement set in VituixCAD.
For what it's worth, in my old 3-way PA the sub/mid time alignment (crossed over at 100Hz) made a considerable improvement to the sound, but I never noticed the same dramatic effect when correcting the (1.8kHz) mid/HF alignment. The system consisted 4 x sealed subs; 10" ported mid; AMT HF.
This gets to the definition of "in-phase".
If it means at one frequency only, in-phase and time alignment are not the same.
If it means at all frequencies throughout xover region, then it includes time-alignment.
If good square waves can be produced across the frequency spectrum, it means not only is the speaker "in-phase" and "time aligned", but also that the speaker has flat phase at zero degrees. (aka linear phase)
That said, nobody in their right mind would use square waves to tune a speaker nowadays, given FFT's ability to show mag and phase separately and directly.
Last edited:
Real simple...just use REW in loopback mode (preferred) or acoustic timing reference mode, and look at delay relative to either.
Keep mic in same spot, and measure each driver. Timing differences will equate to distances between acoustic centers.
The smoother the drivers' passband responses, the better the relative timings.