This may interest anyone integrating subwoofers with main speakers. My system consists of twin subs and quite large floor standing Open Baffle main speakers, digital crossovers, all time & phase aligned. The larger wavelengths, green/red in the images below are two subs, the other plots are the main speakers, from measurements taken at the listening position
1. The most common method (?)....
I often read suggestions that time alignment should be made against the first impulse peaks of both mains and subs. To me, this produces a rather unnatural bass response, slightly hollow sounding in the listening room. Very subtle, but 'not quite right'
2. An alternative solution....
Instead, I prefer to align against the second impulse peak as shown in the second image below, with the first rise in the sub impulses closer to the first main peak. This seems to produce a more natural bass within a very smooth overall frequency response, more 'lifelike' for want of a better word. Individual bass notes, drum strikes etc are heard as identifiably separate items within the soundstage and transient attack is sharp and punchy. Bass guitar also has the correct 'growl' heard at a live performance, upright bass sounds 'in the room' etc
I had also tried aligning with the third and 'highest' peak but this robbed bass of some of its transient impact as well as another subtle change in the way bass was presented, lower register instruments becoming slightly less well defined
The speakers and second sub are time aligned against the furthest from the chair (i.e. slowest) sub, all three driver sets align to the second impulse peak of the main speakers. By using alignment as I have done, music at the listening chair is extremely lifelike, with sharp imaging and an almost holographic 3D soundstage when the original recording is mastered that way
The DEQX processor I use includes the facility for time and phase alignment as well as for room eq which only needs to be very subtle - much of what you adjust can also be done in real time with music playing so changes can be heard instantly and tweaked ‘on-the-fly’
The room has solid walls and floor and only required 3 adjustments of - 1.75dB @33hz, -0.95dB @ 37 hz (from the two subs) and + 0.95dB at 90hz hz. Having two subs in different positions in the room also makes the job of taming and cancelling standing waves much easier. In other places in the room the bass is not quite perfect but as it is a dedicated space, music at the single chair is all that matters to me
As well as being a pretty good sounding room in the first place, it has a lot of vinyl storage along the rear wall behind the chair and other varied surfaces, CD/SACD shelving and materials which also assist in treating the sound. When a piece of furniture was added/removed/moved or anything else substantially changed, I was able to re-run the measurement at the listening chair and by comparing the resulting plots, see the impact this had and at what frequency. That means I didn't have to guess about what was happening as it was visible on the graphs and I was able to treat everything logically
Having said all that , the final measure is listening to music and that is why I had to make that final change in alignment
Why might this have worked?....
Aligning differently from the first impulse peak is not unusual in the pro world - true alignment is when each impulse starts, not just the peak which I believe is a common misconception and may only address phase but not timing. Lower frequencies are slower and therefore start with a gentler and longer wavelength as is evident in both plots at the top of this article
Here are a couple of quotes from a well respected US speaker brand - Green Mountain - where time coherence is explained:
"When a speaker is not time-coherent, it often allows the highs to come out too soon. This usually makes the aggressive sounds of rock far too aggressive. However, the common remark heard is, "These speakers are very revealing of what must be a bad recording.""
"When each one's beginning portion of the pulse arrives at the microphone at the same instant, the total pulse should appear quickly and die away smoothly"
...and a further quote illustrating where people often mix up phase coherence (my first illustration) with phase AND time coherence (the second illustration)
"What about speakers that are only phase-coherent? Phase coherence means simply that the twin peaks and valleys of the same pure test tone coming from two drivers (such as from a woofer and tweeter) line up at your ear. Make those two waves also start and stop at the same time and you have a speaker that is both phase and time coherent."
Why measurement may not always provide the optimum result....
I feel I should add one further clarification to the above. Although the DEQX algorithms time and phase align all frequencies within a given measured driver, speaker or sub; when manually time aligning mains to subs as I have shown, the impulse plots are a sum of all the frequencies from the particular speaker set being measured
Therefore manual time alignment is always going to be something of a compromise because on both sides of the crossover the measured impulse response includes ALL frequencies from that speaker arriving at once but only one curve is visible. As lower frequencies have longer (slower) wavelengths, it may be safe to assume that the first rise of the subwoofer is made by the highest (fastest) frequency it handles, usually from a crossover somewhere between 80-100hz but the impulse of the main speaker will probably start with its highest frequency, often in excess of 30khz. Add to this the slope of the crossover (I am using very steep @ 72dB) and you can see that any human interpretation of time alignment relying on a graph alone is likely to be less than optimal
The trick is to find the point at which the impulse of both sets of drivers is aligned. That is why it is important to listen carefully and not just rely on a graph. It may sound unimportant to someone without the means to try it but changes in timing of just a few milliseconds can make quite a dramatic difference to the way bass sounds. Of course anyone reading this may feel free to remain cynical and I only make this post in case others may wish to experiment. The results have certainly been wholly positive for me....
After investigating this topic for several years and listening to multiple variations of phase and timing, always with the same selection of music tracks, I now know with confidence that my system is as well aligned as it can be given the technology available today. With music playing, it certainly sounds like it !
1. The most common method (?)....
I often read suggestions that time alignment should be made against the first impulse peaks of both mains and subs. To me, this produces a rather unnatural bass response, slightly hollow sounding in the listening room. Very subtle, but 'not quite right'
2. An alternative solution....
Instead, I prefer to align against the second impulse peak as shown in the second image below, with the first rise in the sub impulses closer to the first main peak. This seems to produce a more natural bass within a very smooth overall frequency response, more 'lifelike' for want of a better word. Individual bass notes, drum strikes etc are heard as identifiably separate items within the soundstage and transient attack is sharp and punchy. Bass guitar also has the correct 'growl' heard at a live performance, upright bass sounds 'in the room' etc
I had also tried aligning with the third and 'highest' peak but this robbed bass of some of its transient impact as well as another subtle change in the way bass was presented, lower register instruments becoming slightly less well defined
The speakers and second sub are time aligned against the furthest from the chair (i.e. slowest) sub, all three driver sets align to the second impulse peak of the main speakers. By using alignment as I have done, music at the listening chair is extremely lifelike, with sharp imaging and an almost holographic 3D soundstage when the original recording is mastered that way
The DEQX processor I use includes the facility for time and phase alignment as well as for room eq which only needs to be very subtle - much of what you adjust can also be done in real time with music playing so changes can be heard instantly and tweaked ‘on-the-fly’
The room has solid walls and floor and only required 3 adjustments of - 1.75dB @33hz, -0.95dB @ 37 hz (from the two subs) and + 0.95dB at 90hz hz. Having two subs in different positions in the room also makes the job of taming and cancelling standing waves much easier. In other places in the room the bass is not quite perfect but as it is a dedicated space, music at the single chair is all that matters to me
As well as being a pretty good sounding room in the first place, it has a lot of vinyl storage along the rear wall behind the chair and other varied surfaces, CD/SACD shelving and materials which also assist in treating the sound. When a piece of furniture was added/removed/moved or anything else substantially changed, I was able to re-run the measurement at the listening chair and by comparing the resulting plots, see the impact this had and at what frequency. That means I didn't have to guess about what was happening as it was visible on the graphs and I was able to treat everything logically
Having said all that , the final measure is listening to music and that is why I had to make that final change in alignment
Why might this have worked?....
Aligning differently from the first impulse peak is not unusual in the pro world - true alignment is when each impulse starts, not just the peak which I believe is a common misconception and may only address phase but not timing. Lower frequencies are slower and therefore start with a gentler and longer wavelength as is evident in both plots at the top of this article
Here are a couple of quotes from a well respected US speaker brand - Green Mountain - where time coherence is explained:
"When a speaker is not time-coherent, it often allows the highs to come out too soon. This usually makes the aggressive sounds of rock far too aggressive. However, the common remark heard is, "These speakers are very revealing of what must be a bad recording.""
"When each one's beginning portion of the pulse arrives at the microphone at the same instant, the total pulse should appear quickly and die away smoothly"
...and a further quote illustrating where people often mix up phase coherence (my first illustration) with phase AND time coherence (the second illustration)
"What about speakers that are only phase-coherent? Phase coherence means simply that the twin peaks and valleys of the same pure test tone coming from two drivers (such as from a woofer and tweeter) line up at your ear. Make those two waves also start and stop at the same time and you have a speaker that is both phase and time coherent."
Why measurement may not always provide the optimum result....
I feel I should add one further clarification to the above. Although the DEQX algorithms time and phase align all frequencies within a given measured driver, speaker or sub; when manually time aligning mains to subs as I have shown, the impulse plots are a sum of all the frequencies from the particular speaker set being measured
Therefore manual time alignment is always going to be something of a compromise because on both sides of the crossover the measured impulse response includes ALL frequencies from that speaker arriving at once but only one curve is visible. As lower frequencies have longer (slower) wavelengths, it may be safe to assume that the first rise of the subwoofer is made by the highest (fastest) frequency it handles, usually from a crossover somewhere between 80-100hz but the impulse of the main speaker will probably start with its highest frequency, often in excess of 30khz. Add to this the slope of the crossover (I am using very steep @ 72dB) and you can see that any human interpretation of time alignment relying on a graph alone is likely to be less than optimal
The trick is to find the point at which the impulse of both sets of drivers is aligned. That is why it is important to listen carefully and not just rely on a graph. It may sound unimportant to someone without the means to try it but changes in timing of just a few milliseconds can make quite a dramatic difference to the way bass sounds. Of course anyone reading this may feel free to remain cynical and I only make this post in case others may wish to experiment. The results have certainly been wholly positive for me....
After investigating this topic for several years and listening to multiple variations of phase and timing, always with the same selection of music tracks, I now know with confidence that my system is as well aligned as it can be given the technology available today. With music playing, it certainly sounds like it !
Maybe this is a newbie question, but why not just use the alignment that is loudest for steady-state notes in the crossover point? While the wiggles at the start of a sound are impressive, I doubt our brains are doing much with the first cycle or two when our brains are next getting a hundred more cycles in the next fraction of a second.
BTW, I finally went DSP about a year ago. I forever dispaired of aligning my diverse group of speakers. But the results proved to be detectably good. Of course, every system is a bunch of compromises and the extra punch from time alignment is just one desirable feature which might be accessed only at the cost/complexity of other features.
Ben
BTW, I finally went DSP about a year ago. I forever dispaired of aligning my diverse group of speakers. But the results proved to be detectably good. Of course, every system is a bunch of compromises and the extra punch from time alignment is just one desirable feature which might be accessed only at the cost/complexity of other features.
Ben
Last edited:
I know that I'm a little late to this thread, but I'm using a DEQX Express II to attempt to align my matching stereo subs but I find that have a bit of a dilemma. By aligning on the the first impulse-peak, but I will try your second-peak method soon, the right sub has a 9.5ms delay and left has an 11.5ms delay. Since you can't very well have a 2ms difference between the two main speakers during the quest to achieve perfect sub alignment, I know it's a compromise, but I have found that splitting the difference and loading in a 10.5ms delay on both main speaker seems to produce the best results so far. Also, the two subs currently measure to be about same distance to the listening seat.
How do you or anyone else think would be the best way to deal with this issue? Start moving subs around? Align the subs in mono . . . or?
How do you or anyone else think would be the best way to deal with this issue? Start moving subs around? Align the subs in mono . . . or?
As the OP on this thread, I received an email informing me of your question:
Your explanation seems to imply that you are delaying the main speakers differently which is not correct if you are seated centrally to these. The subs should be aligned together as a pair and then this pair aligned to the main speakers.
The most common situation is that the subs are slower than the main speakers so what you do is slow one sub to match the other (you mention 11.5ms and 9.5ms ie it should be a 2ms delay to the 'fastest' sub appearing to the left on the step response view) and then align the peaks of both subs (which should now be aligned together as in my images) to either the first or second impulse of the main speakers.
As you are using DEQX, it is easiest to use the step response plots in a 'viewer' window (found under 'Tools'). Apologies if I am telling you something you already use !
Andrew
Andrew, when I read your reply, a light-bulb lit inside of my head and it finally dawned on me that there is in-fact a time delay tab for each subwoofer in the configuration window. Sheesh! I've been playing with the DEQX calibration software for two or three years now and when you think that you have it all figured out you find that you have missed a basic fundamental. In this case the fundamental of aligning L+R subwoofers prior to aligning them with the main speakers.
Anyway, I did exactly what you said and added a 2ms delay to my fastest sub bringing them into perfect alignment with each other. I then created two configurations where P1 sets the delay on my main speakers to align the subs with the main speaker's first impulse peak (L+R both at 11.5ms) and the second configuration where P2 sets the delay to align on the second peak (L+R both at 10.5ms) as your original post suggested.
Upon first listen, and it didn't matter if the DEQX was on P1 or P2, I was totally struck by how much more coherent the system sounded. Not only it sounded better in bottom-end, is sounded overall much better through the entire mid-range and up. Imaging was much better than ever before and I thought that it was actually really good before I created this new DEQX configuration.
Now I think all of this goes to show that if your subs are out of phase, even by 2ms, which doesn't sound like much but believe me it is, and not in perfect alignment with your main speakers, you are not even close to getting the full potential out of your system. This exercise made a believer out of me!
Next I will evaluate the differences between the P1 (first peak alignment) and P2 (second peak alignment) configurations. Upon the initial listening session, I could hear that there was a difference, but at this point I haven't have a chance to make a serious evaluation.
On a side note, it would be great if our DEQX community would compile a document that could be posted somewhere on DEQX tips and other undocumented configurations and features that are not found in the DEQX software manual. I think that the information found in this thread is what many users may find useful in refining their DEQX configuration for the best sound possible. It sure helped me!
Andrew, thank you for the great advise.
Very glad to assist Forrest. I consider myself an advanced DEQX user and have helped quite a number of others to get the best out of their setups.
As for aligning to the main speakers, there is no hard and fast rule - what you are trying to do is align the first rise rather than the peaks so the delay (to the main speaker pair) that sounds best is the one to go for. It is always going to be a slight compromise & may be somewhere between the first and second peaks. Your ears will tell you when it is correct!
Good luck
Andrew
As for aligning to the main speakers, there is no hard and fast rule - what you are trying to do is align the first rise rather than the peaks so the delay (to the main speaker pair) that sounds best is the one to go for. It is always going to be a slight compromise & may be somewhere between the first and second peaks. Your ears will tell you when it is correct!
Good luck
Andrew
I wonder if you are missing my point? It is great when people rave about changes they make but that's not the same as trustworthy evidence arrived at by blind testing. (... and that's as tactful a way of saying it as I can state it.)
B.
Hi Ben, You know, it really doesn't matter what sounds best to your ears or for that matter to my ears either. The way I read Andrew's first post is that he is presenting an alternative DEQX alignment method that worked well for him in his system and gave some reasons why. I don't think that I read anywhere where he stated that second-peak sub alignment is head-and-shoulders better than any other method and you damn sure ought to use it. He is just saying here is how it's done and if you are using a DEQX to give it a try. You may just like the results. Does one method produce more accurate results than the other? Hell, I don't know, but the product of one method may produce slightly better results to my ears, yours ears, his ears, or someone else's ears than another.
If everyone agreed on what sounds best, we would all just put together essentially the same system using the same parameters and be done with it. But we all know that's not the case. Looking at your current projects line in your post footer, though I'm not sure what "motional feedback woofing" is, but I do know what large ESL panels are, and I can deduce from that line you are building or have built one heck of a fantastic system that I for one would really like to hear about. Moreover, like Andrew did in his first post, and even if you are not using a DEQX, I would really like to hear what you think of the SQ of your system as it relates to some of its technical aspects. For instance, how did you go about aligning the super-fast ESL panels to the MFW? Are you using some sort of DSP? Active crossovers? What electronics brings it all together. [ Post Edit: I now see that you are a frequent forum poster here and I'm sure that somewhere in your many posts you have probably described your system/ESL/motional feedback woofing in great detail in at least some of of those 2000+ postings. I try to read through some of them to get caught up.]
Anyway, this forum is "DIY Audio" and that's why I'm here - to learn. I think Andrew is a knowledgeable straight shooter and is not trying to BS anyone. In a couple of days, as if it really matters to anyone, I will post my opinion of my experience with first vs second peak sub alignment.
If everyone agreed on what sounds best, we would all just put together essentially the same system using the same parameters and be done with it. But we all know that's not the case. Looking at your current projects line in your post footer, though I'm not sure what "motional feedback woofing" is, but I do know what large ESL panels are, and I can deduce from that line you are building or have built one heck of a fantastic system that I for one would really like to hear about. Moreover, like Andrew did in his first post, and even if you are not using a DEQX, I would really like to hear what you think of the SQ of your system as it relates to some of its technical aspects. For instance, how did you go about aligning the super-fast ESL panels to the MFW? Are you using some sort of DSP? Active crossovers? What electronics brings it all together. [ Post Edit: I now see that you are a frequent forum poster here and I'm sure that somewhere in your many posts you have probably described your system/ESL/motional feedback woofing in great detail in at least some of of those 2000+ postings. I try to read through some of them to get caught up.]
Anyway, this forum is "DIY Audio" and that's why I'm here - to learn. I think Andrew is a knowledgeable straight shooter and is not trying to BS anyone. In a couple of days, as if it really matters to anyone, I will post my opinion of my experience with first vs second peak sub alignment.
Last edited:
Depending on the DSP (or analog) implementation of the low pass filter it will generally delay the sub by 90 degrees per pole, (90 degrees for 6 dB per octave, 180 for 12 dB per octave, etc.) so eliminating the LP for alignment would not be advisable.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.