hi, my speakers are shown in my avatar.
I use a mini dsp to cross the woofer to the mid and the upper speaker is passive.
I want to calculate two things here and would like to know if those two things needs to be taken into account when trying to compensate with time alignement.
Please let me know if those two points needs to be taken into account, the first one is certain, not sure the second one...
1- The woofer voice coils difference between the midrange voicoils and the woofer voice coil.
2- the distance from my listening position and the woofer and my listening position from my woofer
.as you can see in those graph of my woofer and mid, the approx difference between the voice coil of the woofer and the voice coil of the midrange is around 76 mm.
The Madisound Speaker Store
Beyma Speakers - Beyma 12BR70 speaker - Beyma 12BR70 250 watt 12" woofer for all hi-fi and studio monitor bass speakers. Beyma 12BR70 and other Beyma 12" speakers here.
My listening position put my head at exactly the middle of the midrange driver. the middle of the midrange driver is at around 1 feet away from the middle of the woofer.
Do I need to add 1 feet to this formula?
.
Sounds travels at 343 meters per second, or 1126 feet per second. If you can measure approximately the distance between the tweeter and the woofer cone as indicated above, you can calculate the time delay with one of these formulae:
so should I put the 76mm + the 1 feet distance into account?
that would give me a approx 38cm delay for the mid.
38cm/34.3 = 1.07
Do I get this right?
thanks
I use a mini dsp to cross the woofer to the mid and the upper speaker is passive.
I want to calculate two things here and would like to know if those two things needs to be taken into account when trying to compensate with time alignement.
Please let me know if those two points needs to be taken into account, the first one is certain, not sure the second one...
1- The woofer voice coils difference between the midrange voicoils and the woofer voice coil.
2- the distance from my listening position and the woofer and my listening position from my woofer
.as you can see in those graph of my woofer and mid, the approx difference between the voice coil of the woofer and the voice coil of the midrange is around 76 mm.
The Madisound Speaker Store
Beyma Speakers - Beyma 12BR70 speaker - Beyma 12BR70 250 watt 12" woofer for all hi-fi and studio monitor bass speakers. Beyma 12BR70 and other Beyma 12" speakers here.
My listening position put my head at exactly the middle of the midrange driver. the middle of the midrange driver is at around 1 feet away from the middle of the woofer.
Do I need to add 1 feet to this formula?
.
Sounds travels at 343 meters per second, or 1126 feet per second. If you can measure approximately the distance between the tweeter and the woofer cone as indicated above, you can calculate the time delay with one of these formulae:
- Delay in milliseconds = distance in cm / 34.3
- Delay in milliseconds = distance in inches / 13.5
so should I put the 76mm + the 1 feet distance into account?
that would give me a approx 38cm delay for the mid.
38cm/34.3 = 1.07
Do I get this right?
thanks
Last edited:
I just run a tone at the xo frequency ... invert polarity on one of the drivers and adjust delay for the steepest null on my RS SPL meter.
fine, but want a approximation, a certain ball park.
1 ms or 0.11 ms
If I use point 1, its only around 0.11 but if I Need to take into account the distance between the woofer and the mid to my lsitening position, I add another whole feet, which is not a small difference. It would then be around 1 ms
Should I only use point 1 or also point 2????
thanks
1 ms or 0.11 ms
If I use point 1, its only around 0.11 but if I Need to take into account the distance between the woofer and the mid to my lsitening position, I add another whole feet, which is not a small difference. It would then be around 1 ms
Should I only use point 1 or also point 2????
thanks
Last edited:
Point 1 and 2 aren't important. A drivers acoustic center varies with frequency and isn't governed by the physical dimensions of it's components. Since you're time aligning for the xo frequency, it makes the most sense to use the polarity shift to find the alignment.
With delay all you can do is align the length of flight paths to a specific location.
If you picture a triangle where the woofer is under the tweeter and you draw one line there (point 1 and point 2) and then connect those to point 3, the listening/mic position, say 1M out in from the the baffle, the delay cannot change anything about about the distance between pts 1 and 2. But it can change where 1 to 3 is effectively different than 2 to 3.
I usually have better luck going top down, IE align tweeter and mid first, then align woofer to mid, using the same method puppet describes, or sweeps with flipped phase on one driver looking for the deepest null point. Hope that helps.
If you picture a triangle where the woofer is under the tweeter and you draw one line there (point 1 and point 2) and then connect those to point 3, the listening/mic position, say 1M out in from the the baffle, the delay cannot change anything about about the distance between pts 1 and 2. But it can change where 1 to 3 is effectively different than 2 to 3.
I usually have better luck going top down, IE align tweeter and mid first, then align woofer to mid, using the same method puppet describes, or sweeps with flipped phase on one driver looking for the deepest null point. Hope that helps.
hi everyone
yesterday ive done a blind test.
Iv set up different delays: 0.1ms, 2.5ms, 5 ms, 7.5, ms, 1ms
The 1 ms sounded like the singer came from the back of a tunnel, not good
the 0.5 placed the bass a bit too fast, hard to explain
0.25 was the perfect setting, everythign is there, with a focus Ive never had before.
0.1 was almost okay, but I prefered the 0.25, the sound was a bit too forward
So now I know its in the ballpark of 0.25ms
The sound os my speaker is now MUCH better. I used to never be able to have a nice bass bass, always too loud or too loiw. Now, the bass is perfectly in focus and I can, if I want add more bass and the sound will not be boomy or lacking focus!
I am stunt by the difference. Going back to no delay is really noticable. God damn we learn something everyday. Makes me wonder if I should go completely active to be able to do the same with the tweeter/mid, but I like the sound of my upperspeaker and the FR is ruler flat.
I will continue exploring this, but the changes the delay made is stunning, WOW. I'm getting closer and closser to the coherence of a fullrange im used to, with all the 3 way advantage!
thanks to anyone
yesterday ive done a blind test.
Iv set up different delays: 0.1ms, 2.5ms, 5 ms, 7.5, ms, 1ms
The 1 ms sounded like the singer came from the back of a tunnel, not good
the 0.5 placed the bass a bit too fast, hard to explain
0.25 was the perfect setting, everythign is there, with a focus Ive never had before.
0.1 was almost okay, but I prefered the 0.25, the sound was a bit too forward
So now I know its in the ballpark of 0.25ms
The sound os my speaker is now MUCH better. I used to never be able to have a nice bass bass, always too loud or too loiw. Now, the bass is perfectly in focus and I can, if I want add more bass and the sound will not be boomy or lacking focus!
I am stunt by the difference. Going back to no delay is really noticable. God damn we learn something everyday. Makes me wonder if I should go completely active to be able to do the same with the tweeter/mid, but I like the sound of my upperspeaker and the FR is ruler flat.
I will continue exploring this, but the changes the delay made is stunning, WOW. I'm getting closer and closser to the coherence of a fullrange im used to, with all the 3 way advantage!
thanks to anyone
"Time-Aligning" the drivers (acoustic center) and then designing the crossover is to forget that the crossover itself introduces different delays for the HP and LP filters (passive or active). You really need to design everything together as a system.
This is why many DIYers now use the method that has been popularized by Jeff Bagby (he was not the first to do it) in which you take measurements of each driver separately and then measurements of pairs of drivers (e.g. W+M, M+T) operating together without any filters or components in the signal chain. Using programs like Jeff's Passive Crossover Designer or my Active Crossover Designer you can then discover the exact acoustic offset for each driver as seen from the microphone position. This information can then be used during the crossover design process to make sure that you know the phase and phase alignment well.
The above does not eliminate your concerns about time alignment, but you can use the same procedure to determine the offset for each driver and then construct a stepped or sloped baffle that will largely zero out the time offsets. Then you can re-measure the acoustic centers and see how close you are to "time-aligned". But, again, as soon as you add in the crossover you will find that the group delay of each filter will introduce new delays and screw up your time alignment.
You probably know that time alignment is really only valid at one position in space, and it is not the end all of loudspeaker design, but it you want to bark up that tree I wish you good hunting.
-Charlie
This is why many DIYers now use the method that has been popularized by Jeff Bagby (he was not the first to do it) in which you take measurements of each driver separately and then measurements of pairs of drivers (e.g. W+M, M+T) operating together without any filters or components in the signal chain. Using programs like Jeff's Passive Crossover Designer or my Active Crossover Designer you can then discover the exact acoustic offset for each driver as seen from the microphone position. This information can then be used during the crossover design process to make sure that you know the phase and phase alignment well.
The above does not eliminate your concerns about time alignment, but you can use the same procedure to determine the offset for each driver and then construct a stepped or sloped baffle that will largely zero out the time offsets. Then you can re-measure the acoustic centers and see how close you are to "time-aligned". But, again, as soon as you add in the crossover you will find that the group delay of each filter will introduce new delays and screw up your time alignment.
You probably know that time alignment is really only valid at one position in space, and it is not the end all of loudspeaker design, but it you want to bark up that tree I wish you good hunting.
-Charlie
Any time your mess with "time alignment" via delay you will also be changing the phase for that driver. What you likely heard is constructive and destructive interference in the frequency response around the crossover points. That is MUCH more audible compared to waveform time distortion.
Did you measure the system frequency response each time your changed the delay?
I would very much like to see plots of that data...
Time alignment is nothing but changing the phase of the driver. Your blind testing with various time delays is showing you what effect it has on the frequency response. As the phase changes, the vector sum between the two drivers changes and that results in the different sound you are hearing.
What you need to do is measure using a software like HolmImpulse and pay attention to the phase of one driver with the next. When the phase of the two drivers overlap through the crossover region, you have achieved the correct delay. Of course, remember that crossovers also introduce phase changes, i.e., delays. The LP will move the woofer back and HP will move the tweeter towards the listener.
Here's a great thread explaining how all this works:
Crossover mods for the AR4x - Mods, Tweaks, and Upgrades to the Classics - The Classic Speaker Pages Discussion Forums
Don't worry if you don't get it the first time. Keep reading it again and again until it sinks in.
What you need to do is measure using a software like HolmImpulse and pay attention to the phase of one driver with the next. When the phase of the two drivers overlap through the crossover region, you have achieved the correct delay. Of course, remember that crossovers also introduce phase changes, i.e., delays. The LP will move the woofer back and HP will move the tweeter towards the listener.
Here's a great thread explaining how all this works:
Crossover mods for the AR4x - Mods, Tweaks, and Upgrades to the Classics - The Classic Speaker Pages Discussion Forums
Don't worry if you don't get it the first time. Keep reading it again and again until it sinks in.
no I didnt, should I?
What is obvious is which setting I prefered. I did the blind test with me, then my girlfriend also, and we both prefered the 0.26ms, in a blind setting. EAch time I chose the 0.26ms.
ABSOLUTELY!
I am sure that you are hearing real differences. But this is not because of time alignment per se. As ra7 re-iterated above, when you change the delay you are introducing phase lag by 2pi*F*dt, that is the phase change resulting from the delay increases with (is not constant with) frequency.
My guess is that if you measure the system frequency response for some of the delays you used, you will see changes in the frequency response somewhere in the 300-1kHz range, which has the effect of changing the timbre of voice and apparent soundstage depth, presentation, etc. Very small changes across this entire band are audible. It's probably where the woofer and midrange crossover is located.
What is the W-M crossover frequency and what order crossover are you using for that?
Yes, you should measure, on the axis halfway between the centers of the two drivers in question (or at least that's going to be very close to what you want). If you have ARTA, an easy way to get the right neighborhood is to use the RTA function with PN(pink) signal, using the shortest signal length that covers your crossover range. With polarity inverted on one driver, you can then adjust the delay until the null at xover point is deepest. Then do regular measurements to refine it. Then put the polarity back. If your xover is pretty low, you may have trouble measuring the null indoors.
However, you only get textbook-like results and delay that matches the real acoustic offset if you have first measured each driver and EQ'd the individual driver responses to be flat well past the xover point. If you don't or can't do that, the best settings will generally be different delay and asymmetrical crossover filters - much like passive xovers and just as tricky to get right.
However, you only get textbook-like results and delay that matches the real acoustic offset if you have first measured each driver and EQ'd the individual driver responses to be flat well past the xover point. If you don't or can't do that, the best settings will generally be different delay and asymmetrical crossover filters - much like passive xovers and just as tricky to get right.
I use BW/18db for the woofer and BW 6db for the midrange. but the mid crossover makes it more a 2th order acoustic
the xo is around 375hz
okay thanks everyone, more work for me to do!
the xo is around 375hz
okay thanks everyone, more work for me to do!
I have noticed that it is very tricky to get delay right when xo is high, around 3kHz LR4 in my case. The wave is so short that phase easily rolls over! Lower crossovers are easier to match. Minidsp plugin shows delay time and respective distance, it makes things easier. A measurement system is needed but the microphone quality is not critical.
You should also check phase after rotating the sepaker 30 and 60 degrees. If the phase and response get messed, you might have have rolled over the phase!
I have 0.30ms delay for 900Hz xo and 0.056 for 3,5kHz xo. The baffle is is not straight and the tweeter has horn - not a typical case.
You should also check phase after rotating the sepaker 30 and 60 degrees. If the phase and response get messed, you might have have rolled over the phase!
I have 0.30ms delay for 900Hz xo and 0.056 for 3,5kHz xo. The baffle is is not straight and the tweeter has horn - not a typical case.
OK, the 0.26 ms offset preference makes sense [to me anyway] as your ~375 Hz XO calcs a ~104 deg mean phase rotation [if I did the math right] or ~0.577 ms offset if in same polarity and ~0.288 ms reversed.
If it's not too much hassle, you may want to audition a 0.52-0.58 ms offset with them wired same polarity to make them closer to time/phase aligned through the XO BW like a synergy horn is.
GM