How to manually time align multiple subs to acoustic timing reference speaker?

[Spennis]

Hi,

I'm following OCA's tutorial on manual Audyssey calibration, and I've stumbled upon an issue regarding how to time align multiple subs to the acoustic timing reference speaker.

I have two subwoofers (SW1 & SW2), which I have time aligned to each other using the alignment tool. SW2 needs a time delay of 3,6 ms (=1080 mm) to give a smooth response together with SW1. Now, I am not sure how to proceed with the time alignment to the acoustic timing reference speaker. Following OCA's tutorial, I should now cross-correlation align all my speakers, including the subs, to the acoustic timing reference speaker (the center speaker in my case), in order to find the correct AVR distance setting for each speaker.

Question 1) Should I time align only SW1 to the center speaker, enter the resulting distance into my AVR, and then add the delay/distance for SW2 that the alignment tool gave me to get the proper alignment between both subs and the center speaker?

Now, If I cross-correlation align SW1 to the center speaker, the resulting distance (Cumulative IR shift) is 10806 mm. This is too much for Audyssey, which is capped at a maximum of 6 meters. To my understanding, a fix for this is to manually adjust the timing of the subs to the center speaker by looking at the impulse peaks. But how do I do that in the best way? Looking at the resulting impulse response match between SW1 and the center speaker after cross-correlation align, I can see that there is one small dip and one large peak before the impulse peak of the center speaker, so I suppose I can match to those? But again, how?

Thanks in advance.
/Isak

 

[JStewart]

Question 1) Should I time align only SW1 to the center speaker, enter the resulting distance into my AVR, and then add the delay/distance for SW2 that the alignment tool gave me to get the proper alignment between both subs and the center speaker?

Because you have aligned the subwoofers as a group I expect OCA’s method to align the subwoofer group with other speakers. I’ve not seen this video though.

Perhaps OCA will see this and confirm. On AVNirvana he is @serko70

 

[Spennis]

Because you have aligned the subwoofers as a group I expect OCA’s method to align the subwoofer group with other speakers. I’ve not seen this video though.

Perhaps OCA will see this and confirm. On AVNirvana he is @serko70

Thank you. The AVR (Denon X4500H) has two independent subwoofer outputs, so both distance and volume must be set individually for each sub. If I cross-correlation align the "Aligned sum" of SW1 and SW2 to the center speaker, the distance is 12206 mm.

 

[sm52]

the distance is 12206 mm

I tried to enter a distance of 11m for Denon 4700H. Then Denon told me, quite seriously, that this was too much. That he will not allow more than 8 m. I am disappointed. Why should anyone decide what is good for me, I thought. But we have to live with it. Or buy another receiver.
As for your task, forget for a moment that you have already done some work and do it again, but this time the center speaker will be the reference for all measurements. Just measure all speakers this way. You can post the mdat file here. To make it easier to discuss what happened.

 

[JStewart]

Thank you. The AVR (Denon X4500H) has two independent subwoofer outputs, so both distance and volume must be set individually for each sub. If I cross-correlation align the "Aligned sum" of SW1 and SW2 to the center speaker, the distance is 12206 mm.

All measurements taken with a timing reference and if acoustic timing reference then all measurements used the same? Including the measurement for the subs as a group?

If that’s the case then without seeing measurements it’s impossible to know why. Cross correlation does align to the impulse response peak. Possible that the subs have strong late arriving reflections? Just guessing.

Why not use cross correlation for all speakers with high frequency drivers and use the alignment tool to align the sub group to the mains with phase instead of IR peak? Again, I haven’t seen the video referenced in the OP.

@serko70 ?

 

[Spennis]

I tried to enter a distance of 11m for Denon 4700H. Then Denon told me, quite seriously, that this was too much. That he will not allow more than 8 m. I am disappointed. Why should anyone decide what is good for me, I thought. But we have to live with it. Or buy another receiver.
As for your task, forget for a moment that you have already done some work and do it again, but this time the center speaker will be the reference for all measurements. Just measure all speakers this way. You can post the mdat file here. To make it easier to discuss what happened.

Here's the .mdat file attached. The center speaker (C) is the acoustic timing reference. In this file, I have just aligned SW1 and SW2 with the alignment tool and generated the "Aligned sum" (where SW2 is delayed by 3,6 ms). I have not time aligned the subs to the center speaker, so you can try that for yourselves.

Here's how it looks when I cross-correlation align SW1 to C (Cumulative IR shift = 10806 mm, with speed of sound set to 300 m/s).

Here's how it looks when I cross-correlation align the "Aligned sum" of SW1 and SW2 to C (Cumulative IR shift = 12206 mm, with speed of sound set to 300 m/s).

Bottom line, I need to find an acceptable distance setting for the subs to C, and I need to know how to make this adjustment in REW

 

[sm52]

Here's the .mdat file attached

He is not here.

 

[Spennis]

He is not here.

Having trouble uploading. Does this link work? --> REW measurements (shared)

 

[serko70]

All measurements taken with a timing reference and if acoustic timing reference then all measurements used the same? Including the measurement for the subs as a group?

If that’s the case then without seeing measurements it’s impossible to know why. Cross correlation does align to the impulse response peak. Possible that the subs have strong late arriving reflections? Just guessing.

Why not use cross correlation for all speakers with high frequency drivers and use the alignment tool to align the sub group to the mains with phase instead of IR peak? Again, I haven’t seen the video referenced in the OP.

@serko70 ?

6m limit is the difference with the sub and closest speaker i.e. if your left surround is at 1m, sub limit is 7m. Quite busy today but I'll have a look soon.

 

[JStewart]

of sound set to 300 m/s).

Should be 343 m/s?

Bottom line, I need to find an acceptable distance setting for the subs to C, and I need to know how to make this adjustment in REW

OCA is in the house

The .mdat was not attached in the previous post. Highly likely it will be needed.

 

[Spennis]

The center speaker is the closest speaker at 2,32 meters. Given that SW2 has to be delayed by 3,6 ms (1080 mm) to SW1, SW1 can't have a distance more than 7,24 meters (2,32 + 6 - 1,080), right?

 

[Spennis]

Should be 343 m/s?


OCA is in the house

The .mdat was not attached in the previous post. Highly likely it will be needed.

My Denon X4500H uses 300 m/s, like many older Denon/Marantz AVRs.

Did the link work?

 

[JStewart]

Yes, link in post 8. Missed it.

 

[JStewart]

My Denon X4500H uses 300 m/s, like many older Denon/Marantz AVRs.

I did not know that. Must have been pretty cold where they where.

 

[sm52]

Relative to the central channel, SW1 needs to be done earlier by 2.18 ms (-2.18 ms) = minus 0.75 meters (increase the distance in the receiver by 0.75 meters). And then make both subs earlier by 32 ms (-32 ms) = minus 11 meters (increase the distance in the receiver by 11 meters). My speed of sound is 344 m/s.
This is if your measurements were correct. Because I have a lot of questions. Maybe there's a double application of delays going on somewhere? In the receiver and somewhere else? I have never seen 11 meters for the first sub peak. The first peak is usually from 3 to 6 meters. For example, does the center channel play down to 23 hertz? Also the front left and front right play down to 23 hertz? Also FHL, RHR. I also wouldn't point the mic up to measure the front speakers and subs. And I did not apply any averaging before determining the delay time for each of them.

 

[serko70]

My Denon X4500H uses 300 m/s, like many older Denon/Marantz AVRs

Actually, that's a bit of an advantage because you can dial in longer sub distances. I believe that was the reason Sound United initially went with that absurd choice.

 

[serko70]

If I cross-correlation align the "Aligned sum" of SW1 and SW2 to the center speaker, the distance is 12206 mm.

This sounds a bit too much unless you have a MiniDSP or some other processing in-between. You might need to sync to an earlier impulse peak. 1/1 smoothed Group delay graph helps with finding the correct peak with subwoofers:

Also, correctly dialing in two subs in the ".ady" is a bit tricky. You need to set "trim" and "delay" correctly for the second sub. Trim is relative volume of the second sub compared to SW1 in dB and delay is relative delay of the second sub in meters. Sometimes they will show up under SW1, sometimes in SW2, sometimes in both depending on the AVR model and subwooer locations. The latest scripts in Audyssey ART video will recognize and handle (sort of) most dual sub configurations but you will need to create a dummy SW2.txt file and later delete it manually in the ady.

 

[serko70]

Having trouble uploading. Does this link work? --> REW measurements (shared)

I hadn't seen that you've shared your mdat...

You're right, the subs have extreme delays. The better sub combination (red line below) is when one of them is inverted and SW2 is delayed an extra 3.135m) but that combination cannot even be aligned to speakers' for "IR starts". The less optimal combination (purple) can "just" do that (IR starts alignment):

Is there some internal processing going on in them like PEQ, etc? If so, try to get rid of them and align the inverted combination for IR peaks. You should also measure subs 0-24kHz to make use of GD graphs.

Here are the final settings:

Sub volume could need -8dB instead of -2dB, I am not sure how combined sub volumes are handled by Audyssey. Your speakers can be cross-overed much lower but I chose to fix it at THX standard 80hz. With the better sub alignment, I would XO them individually.

I produced correction filters and included the mdat files for your reference. The technique is a bit different to video tutorials and should sound better if you opt to use them.



PS For ultimate timing accuracy, you should use Atmos sweeps for LFE (and go for the newer TrueHD lossless ones). There's a 120Hz lpf in Dolby LFE channel which adds a little bit of delay and it's always applied in the AVR when listening to Atmos content.

 

[Spennis]

Because I have a lot of questions. Maybe there's a double application of delays going on somewhere? In the receiver and somewhere else? I have never seen 11 meters for the first sub peak. The first peak is usually from 3 to 6 meters. For example, does the center channel play down to 23 hertz? Also the front left and front right play down to 23 hertz? Also FHL, RHR. I also wouldn't point the mic up to measure the front speakers and subs. And I did not apply any averaging before determining the delay time for each of them.

I use two XTZ Sub 10.17 Edge subwoofers. There are three EQ modes on these and no "Off" option (I use the "REF" mode). There's also a parametric EQ, but I have that one set to Off (bypass). Regarding my measurements, I've measured using a UMIK-1 and @serko70 1M lossless Atmos sweeps. The Dali Opticon MK2 center channel does not play down to 23 Hertz clean, it's all distortion at that point as you can see in the distortion graphs.

This sounds a bit too much unless you have a MiniDSP or some other processing in-between.

Also, correctly dialing in two subs in the ".ady" is a bit tricky. You need to set "trim" and "delay" correctly for the second sub. Trim is relative volume of the second sub compared to SW1 in dB and delay is relative delay of the second sub in meters. Sometimes they will show up under SW1, sometimes in SW2, sometimes in both depending on the AVR model and subwooer locations. The latest scripts in Audyssey ART video will recognize and handle (sort of) most dual sub configurations but you will need to create a dummy SW2.txt file and later delete it manually in the ady.

I do not have a miniDSP in the chain (per your recommendation), but I do have a miniDSP 2x4 HD laying around if that would help setting the distances?

OK. Thank you for that info. I am not sure how mine works yet. Here's my baseline .ady file and it's dummy (link).

You're right, the subs have extreme delays. The better sub combination (red line below) is when one of them is inverted and SW2 is delayed an extra 3.135m) but that combination cannot even be aligned to speakers' for "IR starts". The less optimal combination (purple) can "just" do that (IR starts alignment):

View attachment 68056
Is there some internal processing going on in them like PEQ, etc? If so, try to get rid of them and align the inverted combination for IR peaks. You should also measure subs 0-24kHz to make use of GD graphs.

I'm not sure I can get rid of the internal DSP in the subs. There are three EQ modes on them and no "Off" option (I use the "REF" mode). OK. I was recommended to use 0-200 Hz in another thread. Should I redo the sub measurements with 0-24000 Hz?

Here are the final settings:

View attachment 68057

Sub volume could need -8dB instead of -2dB, I am not sure how combined sub volumes are handled by Audyssey. Your speakers can be cross-overed much lower but I chose to fix it at THX standard 80hz. With the better sub alignment, I would XO them individually.

I produced correction filters and included the mdat files for your reference. The technique is a bit different to video tutorials and should sound better if you opt to use them.

PS For ultimate timing accuracy, you should use Atmos sweeps for LFE (and go for the newer TrueHD lossless ones). There's a 120Hz lpf in Dolby LFE channel which adds a little bit of delay and it's always applied in the AVR when listening to Atmos content.

Thank you very much!! The speaker settings I had look different from yours though (see below).

My reference speaker distance is 2,32 meters (measured with laser) and my SLA and SBL distances are very different (and true to life, as SLA and SLB are equidistant to the MLP, and SBL and SBR as well). The SPL offset values are different as well. The ones I have were recorded prior to any EQ and the offsets were added to the measurements in the .mdat file I shared. Are these differences due to mistakes, or am I missing something?

 

[serko70]

My reference speaker distance is 2,32 meters (measured with laser)

But your unit uses 300 m/s for speed of sound, right?

 

[serko70]

Are these differences due to mistakes

These are the cross corr align results performed by REW with SOS set as 300 m/s. I am quite confident of their accuracy however, I aligned them all to center channel and if it's having problems with the very HF, this could have caused issues. The highest peaks you see in the IR window belong to the highest frequency measured i.e. 24000Hz. Volume levels are results of REW "Align SPL" for 2.5 octaves each way around 100kHz which gave the least variation between your 1/1 smoothed speaker responses.

I was recommended to use 0-200 Hz in another thread

You certainly can but this will decimate the IR and cause its sampling rate to drop (unless you have unticked "decimate IR" in REW default settings). IR peak shape will look more handsome but will be less accurate.

 

[serko70]

These are the correct settings derived from original Audyssey measurements in your ady file:

Many of them are identical to REW measurements but there are some differences. Is your "adjust clock with acoustic reference" option ticked in REW? That's a must with USB mics otherwise they are prone to clock deviations from measurmeent to measurement. I'd trust more on Audyssey mic's distance settings.

Audyssey has inverted both your subs and has decided to apply a 0.78m relative delay to the second sub (2.6ms in 300m/s sos terms) and moved the combined response to 9.12m which is not applicable. The correct setting is 12.41m which is even more inapplicable. Your extreme subwoofer delay remains a mystery unless it's caused by the center speaker (which also strangely seems to need around -6-7dB volume alignment). I'd try using another speaker as acoustic timing reference and remeasure the sub to opt out some possible causes. Also, if there's a phase dial on the subs, move it to 0 degrees.

 

[Spennis]

But your unit uses 300 m/s for speed of sound, right?

Yes. Both REW and the Excel were set to 300 m/s. I thought that would convert distances correctly?

These are the cross corr align results performed by REW with SOS set as 300 m/s. I am quite confident of their accuracy however, I aligned them all to center channel and if it's having problems with the very HF, this could have caused issues. The highest peaks you see in the IR window belong to the highest frequency measured i.e. 24000Hz. Volume levels are results of REW "Align SPL" for 2.5 octaves each way around 100kHz which gave the least variation between your 1/1 smoothed speaker responses.


You certainly can but this will decimate the IR and cause its sampling rate to drop (unless you have unticked "decimate IR" in REW default settings). IR peak shape will look more handsome but will be less accurate.

The distance settings look identical to mine, except for SLA and SBL.

Ok, I'll look into what might cause the difference in SPL alignment.

Decimate IR is unticked. Learnt that the hard way, as REW wouldn't do cross-corr align if measurements had different sampling rates.

These are the correct settings derived from original Audyssey measurements in your ady file:

View attachment 68075

Many of them are identical to REW measurements but there are some differences. Is your "adjust clock with acoustic reference" option ticked in REW? That's a must with USB mics otherwise they are prone to clock deviations from measurmeent to measurement. I'd trust more on Audyssey mic's distance settings.

Audyssey has inverted both your subs and has decided to apply a 0.78m relative delay to the second sub (2.6ms in 300m/s sos terms) and moved the combined response to 9.12m which is not applicable. The correct setting is 12.41m which is even more inapplicable. Your extreme subwoofer delay remains a mystery unless it's caused by the center speaker (which also strangely seems to need around -6-7dB volume alignment). I'd try using another speaker as acoustic timing reference and remeasure the sub to opt out some possible causes. Also, if there's a phase dial on the subs, move it to 0 degrees.

I sure hope so, but I will double check the adjust clock setting. The distances in the .ady file are off partly because I moved the MLP back a bit once I measuree with REW. The laser measured distance from the center speaker to the Audyssey mic tip in that .ady file is 2,277 meters. Sorry for misleading you.

I use external amplifiers for my center, and my FL and FR. I suppose that's the cause of the SPL difference?

 

[serko70]

I thought that would convert distances correctly?

REW does distance calculations accordingly when sos is set to 300m/s but you still have to enter the acoustic reference channel distance as it should be in the AVR to the XL and that requires manual 343/300 adjustment on your part. There's no easy way for the XL file to adjust for that but I noted it for the next Audyssey video (if ever!).

 

[Spennis]

REW does distance calculations accordingly when sos is set to 300m/s but you still have to enter the acoustic reference channel distance as it should be in the AVR to the XL and that requires manual 343/300 adjustment on your part. There's no easy way for the XL file to adjust for that but I noted it for the next Audyssey video (if ever!).

Ok. So if I understand you correctly, are you saying that I should manually convert the actual distance (2,32 meters) to what the AVR thinks the distance is? 2,32 m / 343 m/s = 6,76 ms, so that equals 2,029 m (6,76 ms / 300 m/s)?

Is that the only manual SOS adjustment I need to do?