Home Theater Forum and Systems banner

1 - 20 of 22 Posts

·
Registered
Joined
·
66 Posts
Discussion Starter #1
Hi,

I have seen that there are numerous questions on time aligning speakers and there seem to be several approaches to this. Some of these approaches aren't really clear to me and some don't usually work well for the subwoofers.

Can someone clarify if the acoustic timing signal can be used for time aligning speakers, and how to read it? And is this data reliable for all speakers and how is it calculated. Is the calculation done based upon how much time the timing signal needs in travelling from output to input, and this time is deducted from the following measurement? But on which signal/input is the following measurement timed?

I have been using the center speaker as my acoustic timing reference and my front speakers are -0.15ms in time from the acoustic signal, which makes sense, since Center and Front speakers are placed at nearly identical distance from LP. The acoustic timing reference for my front surround speakers are either 3.15ms or 5.1ms which equals closely the delay settings set for distance. The subwoofer however always shows a huge delay that doesn't make sense, but when I include the subwoofer (x-over) setting for all the other speakers the acoustic timing reference is the same for those speakers as when measured without the subwoofer.
Does this weird subwoofer delay has something to do with the second approach to time align loudspeakers.

I have read that you can use the impulse response graph to time align loudspeakers, but that this doesn't work very well for subwoofers as their frequency response is smeared over time. (Is this what the acoustic timing signal is based upon?) I assume that this approach is based measuring the timing difference between the first impulse peak and zero time as I could not find any examples.

I third approach that I have read about is using the excess group delay graph, but again I could not find any examples in how to use this information for time aligning speakers. Since it is related to phase, it is a tough subject to understand.

A fourth approach is also related to phase, and again I fail to comprehend the big picture. I found an example in this thread http://www.hometheatershack.com/forums/rew-forum/171530-measure-time-delay-rew-2.html#post1596666 which show the steps in working out the timing, but I don't really know what I should be looking at when phase is involved.

Do I need to align the phase angle of each speaker to have the exact same degree at x-over point, or should I align 1st,2nd,3rd phase cycle of each speaker, or should I look at a similar phase respond for each speaker and align these up at the x-over frequency. Since each speaker show a different phase respond to me that is all guess work for me.
 

·
Registered
Joined
·
139 Posts
Best Method for time aligning speakers.

I also have been researching this area and have similar questions and confusion, although I am more interested in the wired loopback measurement method. I am attempting to correctly integrate 2 different subwoofers into a 7.2 Home Theater System using a miniDSP 2X4 DSP box. The conclusion that I have reached at this point is that one goal is to maximize the frequency response handoff of the main L or R speakers to the subwoofer(s). In order to successfully do that the phase relationship should match between the two speakers or all the speakers at or around the XO frequency. For instance this could be a Front Left speaker and 2 or more subwoofers. However, how to efficiently and accurately accomplish this objective for a pair of speakers let alone a whole HT System is still eluding me.

My feeling is that there are a number of parameters that need to be optimized simultaneously - like phase, impulse response, step response and frequency response magnitude for each pair of speakers individually and then combined in the total system. In other words it is not just a simple 1, 2, 3 process! I think member: jTalden has in the past done the most post explanations on this topic. He has said that it takes a significant amount of time (several hours) to analyze and come up with possible optimized settings. BTW he knows what he is doing, I am really just shooting from the hip and hacking around!

Maybe a single general purpose procedure could be clearly defined in this post for the use of all of us who are in the dark woods with no flash light on this subject!
 

·
Registered
Joined
·
139 Posts
Simplifying the Problem of time aligning speakers.

I have been thinking about this speaker phase alignment quest! One thing that is confusing to anyone who has never done this before is getting a good understanding of the required steps and the reasoning behind them to handle the many decisions and tradeoffs that have to be made based on the many parameters that may be affecting the outcome. I think breaking this into several different cases with functional steps that anyone with some REW measurement experience could accomplish would be very helpful. Member jTalden has written more than a few posts on this topic. I have read some (probably not all) of his posts. I have gained some insight for each one studied however the clarity of the objectives is sometimes unclear at least to me.

One case that I believe is probably the simplest is trying to integrate a single subwoofer into a stereo setup. In other words, a 2.1 (L+R+Sub) speaker system driven by an AVR amp for a single listening position. This system has the basic high to low frequency XO crossover integration problem. Even for this simple case, there are a number of operations that need to be done to optimize the setup. Speaker placement. Listening position. Crossover frequency. Type of crossovers. Gain settings for both main speakers and the subwoofer. Delay times (distance settings) for the sub and/or main speakers. If you have already noticed this example case does not include using Audyssey or similar auto EQ systems nor does it include Sub equalization. Those would be extensions or additions to the base case.

I think if this case was well understood it could be the stepping stone for more advanced setups like Home Theater 7.2 system with multiple subs and external equalization of each subwoofer or more. This would enable a firm grasp of the fundamentals that need to be applied.
 

·
Registered
Joined
·
66 Posts
Discussion Starter #4
What I discovered so far doing many measurement is as following:

The time reference delay; You would like this to be as close to zero as possible. This is a rather good starting point before for making any tweaks.

Excess Group Delay (EGD) and SPL graph both have their uses. I think, ideally you want both of them to be as smooth and flat as possible. Your subs and all other speakers will have a specific EGD signature based upon speaker and mic position due to nearby room bounderies. When crossing over, one to the other, you don't want to introduce any new spikes. Ideally you want to eliminate spikes and end up with the flattest and smoothest EGD. However, the flattest and smoothest EGD doesn't always produce the best SPL graph. I think that you want to try as little as these EGD spikes to be line up with dips in the SPL graph, as these would be difficult to correct using EQ.

If you have the ability to tweak your cross over, you can use it to create a visible hump in the EGD. If this hump is in the same frequency range as the cross over point your speakers will be roughly time aligned. You can then start to tweak by making small distance/time adjustments to your subwoofer to see if you can eliminate the hump, and or start tweaking your cross-over settings to try and eliminate the hump. The hump itself can be measured in time, and should give a rough indication about the timing difference.

The cross over for my Left Surround Back speaker is set to 100Hz. Inside JRiver I can set the individual high pas and low pass for this cross over point creating different results. The hump around 100Hz is still not ideal, but I might have to live with it as all other speakers are fine, and the SPL graph shows up fine to.
The EGD spike at 70Hz is the typical spike from my subwoofers. The other spikes, except the 100Hz one are from my LSB speaker.

I hope someone else will join the conversation, as I still don't know if what I am doing is correct, or that I am just spitting out a lot of Bull manure.
 

Attachments

·
Registered
Joined
·
19 Posts
What I discovered so far doing many measurement is as following:

The time reference delay; You would like this to be as close to zero as possible. This is a rather good starting point before for making any tweaks.

Excess Group Delay (EGD) and SPL graph both have their uses. I think, ideally you want both of them to be as smooth and flat as possible. Your subs and all other speakers will have a specific EGD signature based upon speaker and mic position due to nearby room bounderies. When crossing over, one to the other, you don't want to introduce any new spikes. Ideally you want to eliminate spikes and end up with the flattest and smoothest EGD. However, the flattest and smoothest EGD doesn't always produce the best SPL graph. I think that you want to try as little as these EGD spikes to be line up with dips in the SPL graph, as these would be difficult to correct using EQ.

If you have the ability to tweak your cross over, you can use it to create a visible hump in the EGD. If this hump is in the same frequency range as the cross over point your speakers will be roughly time aligned. You can then start to tweak by making small distance/time adjustments to your subwoofer to see if you can eliminate the hump, and or start tweaking your cross-over settings to try and eliminate the hump. The hump itself can be measured in time, and should give a rough indication about the timing difference.

The cross over for my Left Surround Back speaker is set to 100Hz. Inside JRiver I can set the individual high pas and low pass for this cross over point creating different results. The hump around 100Hz is still not ideal, but I might have to live with it as all other speakers are fine, and the SPL graph shows up fine to.
The EGD spike at 70Hz is the typical spike from my subwoofers. The other spikes, except the 100Hz one are from my LSB speaker.

I hope someone else will join the conversation, as I still don't know if what I am doing is correct, or that I am just spitting out a lot of Bull manure.
Yes, I used REW with my Front Left as a reference. I used it to adjust my 3 subs plus all of the surround speakers.
 

·
Registered
Joined
·
333 Posts
I read (in the subwoofer white pages I think) the the best place to start is..... aligning the subwoofers with the room / seating location to find the BEST SOUNDING bass, NOT necessarily the smoothest graph ect. Then adding in your mains/center. The only way I could get perfect bass is by placing my subs near field which almost completely eliminated room issues thus giving me complete control to create the bass I like. Good Luck

Assuming your seating and mains are in the correct position for your room of course.
 

·
Registered
Joined
·
2 Posts
I use REW for some things, but for precise driver phasing I've found the following more direct (if primitive) method better. I have a couple of big 4-ways (incl. sub). First I measure the physical offsets of the drivers and make a ballpark guess at the delays. For the sub/woof point (70hz) it's not a huge deal as the half-wavelength is 8'. For the other xover points, I do the following to fine-tune my delays:

1. Create a tone burst using Audacity and its editor. It's easy. 3 or 4 cycles with leading and trailing silence intervals of perhaps 4x the burst's time interval.

2. Set up a mic a distance away from the drivers and scope the output. If you aren't using a powered (condenser) mic, you'll need a mic preamp.

3. Play the tone burst in loop mode. You may want earplugs.

4. Vary the volume of one driver so you can tell which is which. When they are perfectly aligned you should simply see the burst waveform increase/decrease its amplitude. In most instances (esp. at higher frequencies) you'll see clearly-defined bursts, albeit with some ringing, probably from the filters rather than the drivers. Try to locate the burst itself within ringing, if any, and recall that FIR filters have a half-cycle pre/post ring typically.

It is truly the only foolproof method IMO. Use of nonlinear (e.g. minimum phase) filters makes the process far less than perfect even with alignment, though it's still a good idea to align at the xover frequency.
 

·
HTS Moderator , Reviewer
Joined
·
3,335 Posts
The subwoofer however always shows a huge delay that doesn't make sense, but when I include the subwoofer (x-over) setting for all the other speakers the acoustic timing reference is the same for those speakers as when measured without the subwoofer.
Does this weird subwoofer delay has something to do with the second approach to time align loudspeakers.
That may not be as bizarre as it initially seems. Frequently the extra distance added by room EQ is due to the DSP programming in the amplifier. With the advent of 'intelligent' processing in most subwoofer amps, manufacturers will often tailor the response curve and protection mechanisms to their specific needs. That signal processing introduces some latency, and as a result the delay (distance) will increase to ensure alignment in the time domain. After 50+ subwoofer reviews I have seen distance increases as low as 6 extra feet, relative to physical distance, but some as high as 15 additional feet. It all depends upon how much manipulation is happening.
 

·
Registered
Joined
·
7 Posts
Regarding the above linked post back from 2010. As is often the case when mentioning the use of REW, looping back outputs on soundcards is mentioned as a requirement.

What about those of us who don't have soundcards in our system but are instead outputting audio via HDMI (usually from the HDMI out on the GPU card in the PC) to the HDMI input on the receiver. There's no 'loop back' option there so how do we go about time aligning etc ?

It would be helpful as well if someone could simplify down the essence of that linked post, the procedure to take, as it was rather lengthy
 

·
Registered
Joined
·
66 Posts
Discussion Starter #12
I read (in the subwoofer white pages I think) the the best place to start is..... aligning the subwoofers with the room / seating location to find the BEST SOUNDING bass, NOT necessarily the smoothest graph ect. Then adding in your mains/center. The only way I could get perfect bass is by placing my subs near field which almost completely eliminated room issues thus giving me complete control to create the bass I like. Good Luck

Assuming your seating and mains are in the correct position for your room of course.
My mains are in the best position for deep and wide soundstage. They are not in an ideal position for the low bass. The subs are in the best place when it comes to frequency response and ear, along the halfway length of the room creating the smoothest response and the least amount of room node boosts. There is still the trouble with decay time of low frequency sound energy.

I have 3 solid brick walls, and a length room node at 35Hz that is giving a huge wide peak of +15dB followed by a rather large dip around the 50Hz room width node. I do find the position half way the length of the room the least interfering with the rest of the frequency spectrum and produces a rather flat respond from 30Hz to 80Hz, but I do find myself having the idea that there is a lack of impact with big explosions. I am kinda hoping that time aligning the subs with the mains will improve things there.
 

·
Registered
Joined
·
66 Posts
Discussion Starter #13
That may not be as bizarre as it initially seems. Frequently the extra distance added by room EQ is due to the DSP programming in the amplifier. With the advent of 'intelligent' processing in most subwoofer amps, manufacturers will often tailor the response curve and protection mechanisms to their specific needs. That signal processing introduces some latency, and as a result the delay (distance) will increase to ensure alignment in the time domain. After 50+ subwoofer reviews I have seen distance increases as low as 6 extra feet, relative to physical distance, but some as high as 15 additional feet. It all depends upon how much manipulation is happening.
I might need to take another look at aligning my subs then. They are "delayed" with 27 foot. 7 foot for distance and 20 foot for processing. I have set all DSP settings on the Pre Pro to 0 and speakers to large with should bypass bass management and using JRIver for all DSP based settings.

The one thing that I just found out and can't figure out, is that one of the 2 identical subs at identical distance is producing a bigger delay when I measure them both solo based on the acoustic timing signal. They are being fed from a single LFE output via a y-splitter. One of the subs is using a phase angle of 90 degree, as this produced the biggest SPL output when used together, but changing this never produces any different result when measured solo.

I would kind off expect to see a difference in delay or perhaps in the phase response, although as I understand it, this doesn't effect the phase response of the sub.
 

·
Premium Member
Joined
·
1,562 Posts
Primare Knob,
I scanned this thread quickly as I am currently very busy with relocation activities.
  • Similar SWs can be timed together just by adjusting the delay to account for any difference in distance to the LP. If you want confirmation of the timing, post an mdat containing each SW measurement using acoustic or loopback timing based on one of the mains.
  • If your SW's are well timed and you want confirmation of timing of the SWs to the mains you need to provide an mdat containing 3 full range measurements of; L, R, SWs (again all 3 with timing relative to one of the mains).
I will look at the data and confirm or recommend timing changes. I do not have the time right now to go into much detail beyond that review.
 

·
Registered
Joined
·
66 Posts
Discussion Starter #15
Primare Knob,
I scanned this thread quickly as I am currently very busy with relocation activities.
  • Similar SWs can be timed together just by adjusting the delay to account for any difference in distance to the LP. If you want confirmation of the timing, post an mdat containing each SW measurement using acoustic or loopback timing based on one of the mains.
  • If your SW's are well timed and you want confirmation of timing of the SWs to the mains you need to provide an mdat containing 3 full range measurements of; L, R, SWs (again all 3 with timing relative to one of the mains).
I will look at the data and confirm or recommend timing changes. I do not have the time right now to go into much detail beyond that review.
Thanks, that would be highly appreciated. At the least it will also provide an example, of what is right, which might give me a slight insight on what to look for.


I have uploaded 2 group measurments. One group is made earlier when the sub distance was still set to 7ft according to distance measurements. This group doesn't have an individual measurement for each sub. The other group is with the sub distance set to 27ft, and when I tried to compensate for delays as suggested by the acoustic timing signal. This group does include an individual sub measurement.
https://drive.google.com/open?id=1st7wItQJLNdJ4atIRaLASaPxci7YcsIX

Can't thank you enough.
 

·
Registered
Joined
·
7 Posts
Nowadays REW provides an acoustic timing reference as an alternative to a loopback connection.
I have worked that out and now hopefully have a basic understanding of the impulse alignment method. Of course, I now find out there is an opinion that's the wrong method to use and instead you should phase align/track.

Any simple guides on that as so far from my search, I haven't found anything particularity informative on how to approach that, the links I have found (with attached images etc) seem to assume you already know how to proceed.

Example, I have my SPL and phase trace in REW for both the sub only and sub plus main but well, what then ? The phase trace wraps round several times for the sub (each wrap being a 180 phase change if I'm correct) and many times for the sub plus mains but what about that am I looking to track with the SPL trace? And then what should I expect to see when its tracking correctly as regards those traces. Still wrap arounds ?

I'm using REW, along with EQ APO for all bass management and crossover settings (With the AVR management completely disabled) and it has been suggested Rephase is a good program to use to correct phase issues.

So yes, basically when people say the phase trace in REW should be tracking the SPL, especially around the XO frequency for correct phase setup, what exactly do they mean by that and what should it look like in REW ?
 

·
Premium Member
Joined
·
1,562 Posts
I have uploaded 2 group measurments...
I can see that the 2 SWs are well timed to each other so that is not an issue. I can also see in the IR overlay chart that the SWs are not timed well to the mains. It also appears that the XO was not active for either the L, R measurements or the SWs measurement.


This set of data then does not allow analysis using the phase tracking method. The data needed is very specific for that process.

Setup:
  • Acoustic timing active on a chosen main speaker for all measurements (Right main for this example).
  • The XO must be active for all measurements.
  • Use a full range sweep (20-20k Hz) for all measurements (including the SWs).
  • Record/save SW distance used for the measurements as the recommended setting will be a change to the SW distance, not an absolute setting. Using the actual measured distance is a good choice, but any SWs distance can be used.
  • The main speaker distance settings should be at their actual measured value.
Measure:
  1. Left alone (SWs off)
  2. Right alone (SWs off)
  3. SWs alone [To do that with the XO active it is necessary to measure using the Left main channel (channel 1) as the measurement channel with the Left main speaker disconnected. The full range sweep will thus be redirected to the SWs and left main will be silent. The Right main is still the reference channel in this example.]
The XO settings change the ideal timing/distance and thus the XO needs to be active for all 3 measurements.
 

·
Registered
Joined
·
66 Posts
Discussion Starter #18
I can see that the 2 SWs are well timed to each other so that is not an issue. I can also see in the IR overlay chart that the SWs are not timed well to the mains. It also appears that the XO was not active for either the L, R measurements or the SWs measurement.


This set of data then does not allow analysis using the phase tracking method. The data needed is very specific for that process.

Setup:
  • Acoustic timing active on a chosen main speaker for all measurements (Right main for this example).
  • The XO must be active for all measurements.
  • Use a full range sweep (20-20k Hz) for all measurements (including the SWs).
  • Record/save SW distance used for the measurements as the recommended setting will be a change to the SW distance, not an absolute setting. Using the actual measured distance is a good choice, but any SWs distance can be used.
  • The main speaker distance settings should be at their actual measured value.
Measure:
  1. Left alone (SWs off)
  2. Right alone (SWs off)
  3. SWs alone [To do that with the XO active it is necessary to measure using the Left main channel (channel 1) as the measurement channel with the Left main speaker disconnected. The full range sweep will thus be redirected to the SWs and left main will be silent. The Right main is still the reference channel in this example.]
The XO settings change the ideal timing/distance and thus the XO needs to be active for all 3 measurements.
Thank you for your time and patience. I have added the new data to the same online folder.

It's nice to know that the subs are time aligned. It looks like the REW auto timing data isn't very reliable for the subwoofers.
Would it be possible to program/automate your process? As it might be a nice future feature for REW.

The process that you are using, is this something that can be read or found in another forum thread. I would like to get a better understanding, but can understand it will take a lot of your time to explain it to people like me.
 

·
Premium Member
Joined
·
1,562 Posts
The timing Look very good with the current SW distance setting.
Some charts to show that;

XO range (about 60-160Hz):
SPL XO Range.jpg

Left and Right Phase tracking in that range range.
Phase tracking.jpg

Left Calculated SPL support through the XO range
Left SPL.jpg

Right Calculated SPL support through the XO range
Right SPL.jpg

The phase tracking is very good. Above it is shown with a 10 cycle/octave FDW. It looks even better when I apply a more normal 6 cycle/octave FDW to eliminate more of the room influence on the direct sound. However you arrived at the 27 foot setting worked out extremely well.
 

·
Registered
Joined
·
66 Posts
Discussion Starter #20
The timing Look very good with the current SW distance setting.
Some charts to show that;

XO range (about 60-160Hz):
View attachment 147354

Left and Right Phase tracking in that range range.
View attachment 147362

The phase tracking is very good. Above it is shown with a 10 cycle/octave FDW. It looks even better when I apply a more normal 6 cycle/octave FDW to eliminate more of the room influence on the direct sound. However you arrived at the 27 foot setting worked out extremely well.
Thank you for taking your time with this. It's always nice to have someone more knowledgeable clear up the confusion. When I try and read the phase graph I am getting confused about the phase cycles not being in sync. For example at 70Hz, where the Left and Right phase cycle is starting to sync, but around 58Hz they seem to be out of sync, the same with the sub. Am I on the right track, in assuming that every speaker at a different location will produce an unique phase response within the same frequency range because of position in space and interaction with the room? The only real difference between the left and right speaker is that the right wall has a glass door/window in it halfway.

I was under the impression that at 100Hz (X-over point) the phase angles should closely match. Left and Right have a phase angle of 0 degree while the subs have a phase angle of (+/-) -60 degree, but is this not a real problem?
 
1 - 20 of 22 Posts
Top