Setting the delay of the subwoofer.

Thread Starter
Joined
Nov 24, 2019
Messages
14
Hello! I apologize for my English. I am interested in the possibility of using a subwoofer in a wide frequency band (30-250 Hz). Studying the pulse characteristics taken in the car with different crossover cutoff frequency, it turns out that at the cutoff frequency of 80 Hz and 250 Hz, different delay values are obtained. This is elementary for professionals, but I was surprised. On what principle should I set the subwoofer delay, if I want 30-250 Hz, so that 50 Hz and 200 Hz come without lagging from each other?
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
3,039
The delay setting has no effect on the relative phase of signals coming from the subwoofer, only on how they align with the mains through the crossover region. For alignment of sub to mains look at the threads on the Alignment tool.
 
Thread Starter
Joined
Nov 24, 2019
Messages
14
The delay setting has no effect on the relative phase of signals coming from the subwoofer, only on how they align with the mains through the crossover region. For alignment of sub to mains look at the threads on the Alignment tool.
Thanks for the answer! But with all due respect, we're talking about different things. I'll try to explain it in more detail. Let's imagine that I have two identical low-frequency speakers. By means of crossovers, one has 30-70 Hz, the other 70-250 Hz. The first one has a measured pulse response delay of 15 milliseconds. And the second one has 10 milliseconds. With two speakers, it is easy to subtract the delay of 15 and 10 milliseconds. But what if there is only one speaker that is used in the 30-250 Hz range? How do I set the delay for it: 15 or 10 milliseconds? We are considering a hypothetical case if You agree that the delay at low frequencies increases.
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
3,039
It makes no a sense to take the behaviour of drivers with two different sets of crossover settings applied to them and try to imply that has a bearing on a third set of crossover settings. Choose delay to get good integration between sub and mains, hypothetical musings on unrelated measurements will not get you there any quicker.
 

AudioCAT

Member
Joined
May 4, 2018
Messages
15
Thanks for the answer! But with all due respect, we're talking about different things. I'll try to explain it in more detail. Let's imagine that I have two identical low-frequency speakers. By means of crossovers, one has 30-70 Hz, the other 70-250 Hz. The first one has a measured pulse response delay of 15 milliseconds. And the second one has 10 milliseconds. With two speakers, it is easy to subtract the delay of 15 and 10 milliseconds. But what if there is only one speaker that is used in the 30-250 Hz range? How do I set the delay for it: 15 or 10 milliseconds? We are considering a hypothetical case if You agree that the delay at low frequencies increases.
Is there a difference of 5 ms between the two speaker units? This is equivalent to a distance difference of 1740mm, is it possible?
It is recommended that you check the impulse curve to see where Ref is. It is likely that Ref is not at the peak position of the first impulse.
Maybe you should manually move the Ref to the starting point of the impulse.
 
Thread Starter
Joined
Nov 24, 2019
Messages
14
It makes no a sense to take the behaviour of drivers with two different sets of crossover settings applied to them and try to imply that has a bearing on a third set of crossover settings. Choose delay to get good integration between sub and mains, hypothetical musings on unrelated measurements will not get you there any quicker.
I still believe that google translate is distorting the meaning of my question. I'll try again. As far as I understand, the peak of the pulse response has a period corresponding to the highest measured frequency. Because of this, the time from the reference point to the peak of the pulse will be different for the maximum frequencies of 80 and 250 Hz. A quarter wave of 80 Hz is 3 msec. And for 250 Hz is 1 msec. Therefore, the difference in the delay will be 2 m sec. This delay is provided not only by the physical distance, but also by a quarter of the maximum frequency wavelength. I don't know if I can use a subwoofer up to 250 Hz, or if it's better to use a mid-bass for better timing. Please give me literature to fill in the blank spots in my head. I want to learn the basics. Thank you for your patience!
 
Last edited:
Thread Starter
Joined
Nov 24, 2019
Messages
14
Is there a difference of 5 ms between the two speaker units? This is equivalent to a distance difference of 1740mm, is it possible?
It is recommended that you check the impulse curve to see where Ref is. It is likely that Ref is not at the peak position of the first impulse.
Maybe you should manually move the Ref to the starting point of the impulse.
I gave an example of 5 m sec. But I don't see anything surprising here. The average lag time of the subwoofer in the car interior is 15 m sec, which corresponds to 5 meters. Although the length of the cabin is only 1.5 meters.
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
3,039
The t=0 time for a bandwidth limited driver such as a subwoofer is not at the peak, it is at the start of the response. Aside from that, a subwoofer response does not have flat group delay, there is no time delay figure which will align a subwoofer to a main speaker across the whole subwoofer response. Fortunately there is absolutely no need to try and align the sub to the mains everywhere, only in the region their responses overlap, which is generally about where the subwoofer crossover low pass meets the main crossover high pass. For good integration between the two the subwoofer and main speaker phase responses should align through that region. That is achieved by delay adjustment (usually by delaying the main speaker). The alignment tool helps that process and there are threads on it, such as this one.
 
Thread Starter
Joined
Nov 24, 2019
Messages
14
The t=0 time for a bandwidth limited driver such as a subwoofer is not at the peak, it is at the start of the response. Aside from that, a subwoofer response does not have flat group delay, there is no time delay figure which will align a subwoofer to a main speaker across the whole subwoofer response. Fortunately there is absolutely no need to try and align the sub to the mains everywhere, only in the region their responses overlap, which is generally about where the subwoofer crossover low pass meets the main crossover high pass. For good integration between the two the subwoofer and main speaker phase responses should align through that region. That is achieved by delay adjustment (usually by delaying the main speaker). The alignment tool helps that process and there are threads on it, such as this one.
Thank you for the clarification! I've learned something. But forgive my meticulousness... As I understand it, the pulse response is superimposed on the original signal at each moment of time and adds up to the previous overlays with a shift for the same infinitesimal period of time. If I'm not mistaken, it's called a convolution. And that's how a real speaker system works. Does this mean that the maximum SPL will not be at the peak of the original signal, but shifted away from it by the delay of the peak of the impulse response? I don't understand why we need to focus on the beginning of the response rather than the peak... You write: "Aside from that, a subwoofer response does not have flat group delay, there is no time delay figure which will align a subwoofer to a main speaker across the whole subwoofer response." does this Mean that it is futile to search for the correct high pass delay relative to the subwoofer?
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
3,039
OK, let's try it your way. Here are two idealised driver responses, with an LR2 low pass for the sub at 200 Hz and an LR2 high pass for the main at 200 Hz.

31874


Their impulse responses are exactly aligned at their peaks.

31875


Let's see what their summed response looks like.

31876


Not pretty, a big dip at the crossover frequency and a bit of a bump afterwards. Looking at the phase traces we can see they are a long way apart at the crossover frequency, so let's invert the main and see how that helps.

31877


Better, but still not flat. Let's try something different, how about aligning the impulse responses so they start at the same time.

31878


Here is how the sum looks now, again with the main inverted.

31879


Which would you prefer?
 
Thread Starter
Joined
Nov 24, 2019
Messages
14
OK, let's try it your way. Here are two idealised driver responses, with an LR2 low pass for the sub at 200 Hz and an LR2 high pass for the main at 200 Hz.

View attachment 31874

Their impulse responses are exactly aligned at their peaks.

View attachment 31875

Let's see what their summed response looks like.

View attachment 31876

Not pretty, a big dip at the crossover frequency and a bit of a bump afterwards. Looking at the phase traces we can see they are a long way apart at the crossover frequency, so let's invert the main and see how that helps.

View attachment 31877

Better, but still not flat. Let's try something different, how about aligning the impulse responses so they start at the same time.

View attachment 31878

Here is how the sum looks now, again with the main inverted.

View attachment 31879

Which would you prefer?
These charts are very valuable to me! Thank you very much for them! You have convinced me that the fastest way to get a flat frequency response is to combine the beginning of the peaks. But will not the situation with the frequency of 60 Hz be felt by ear? We will get a flat frequency response, but the bass will lag by a quarter of the subwoofer's wavelength (4.2 m sec). Isn't it worth sacrificing 1-2dB of frequency response at the cutoff frequency for more accurate synchronization?
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
3,039
But will not the situation with the frequency of 60 Hz be felt by ear? We will get a flat frequency response, but the bass will lag by a quarter of the subwoofer's wavelength (4.2 m sec). Isn't it worth sacrificing 1-2dB of frequency response at the cutoff frequency for more accurate synchronization?
No, it won't be heard, and it isn't worth sacrificing the frequency response.
 
Thread Starter
Joined
Nov 24, 2019
Messages
14
No, it won't be heard, and it isn't worth sacrificing the frequency response.
The last question about setting Twitter and midrange delays. Is the program correctly oriented to the peak of the impulse response to calculate the delay? Do I need to manually move t=0 to the maximum of the first pulse?
 

Attachments

Harrycr

Member
Joined
Apr 22, 2020
Messages
13
OK, let's try it your way. Here are two idealised driver responses, with an LR2 low pass for the sub at 200 Hz and an LR2 high pass for the main at 200 Hz.

View attachment 31874

Their impulse responses are exactly aligned at their peaks.

View attachment 31875

Let's see what their summed response looks like.

View attachment 31876

Not pretty, a big dip at the crossover frequency and a bit of a bump afterwards. Looking at the phase traces we can see they are a long way apart at the crossover frequency, so let's invert the main and see how that helps.

View attachment 31877

Better, but still not flat. Let's try something different, how about aligning the impulse responses so they start at the same time.

View attachment 31878

Here is how the sum looks now, again with the main inverted.

View attachment 31879

Which would you prefer?
So we could align subs and mains by aligning the impulses at the start.
I always thought it was the first peak.
 
Top Bottom