Unexpected results when time aligning subwoofers

[CBM]

I have a Denon 3400H receiver connected to Revel M105 speakers with two SVS SB12-NSD subwoofers. I ran Audyssey to establish distances then turned it off because I didn't like the sound. I then manually aligned the subwoofers by changing the distance of one sub until I got the smoothest graph up to 100 Hz where the crossover is. Then I changed the distance of the subwoofers together to get the smoothest graph up to 200 Hz. The result sounds pretty good especially the base.

What surprised me is the effect of these adjustments on the full-range graph.

The orange line is the original setting and the green is after the distance adjustment.

My question is: what caused the deviation on the full-range graph when I only adjusted the distances of the subs. The graphs are 1/24 smoothing and the mic did not move.

 

[RayDunzl]

My question is: what caused the deviation on the full-range graph when I only adjusted the distances of the subs.

Probably:

Relative phase of the two wavefronts at the measuring position.

 

[CBM]

Probably:

Relative phase of the two wavefronts at the measuring position.

Interesting. Not much changed when I delayed one sub relative to the other, but it did when I delayed both subs relative to the mains.

 

[Jdunk54nl]

Do you adjust delay by distance on the denons? If so, when you increase distance on the sub, it actually delays the other speakers respectively (farthest speaker is at 0 ms of delay). So if this is the case, it could be that the other speakers were delayed differently from the beginning (they usually are some step like 0.2ms).

It also could be that using a single spot and doing a single sweep will usually have variations at the upper frequencies.

It also throws me off with your graphs linear spacing instead of log spacing on the x-axis....

 

[DonH56]

If you delay the subs with respect to the mains you are adjusting the phase at the crossover point to the main speakers. It sounds like (sorry) you had to tweak distance to get the subs and mains in phase at the MLP. Audyssey, and some other room correction systems, seem to have trouble with multiple subs, or sometimes even one. That is why so many measure after running room correction to verify (and/or fix) the bass response.

 

[CBM]

Do you adjust delay by distance on the denons? If so, when you increase distance on the sub, it actually delays the other speakers respectively (farthest speaker is at 0 ms of delay). So if this is the case, it could be that the other speakers were delayed differently from the beginning (they usually are some step like 0.2ms).

It also could be that using a single spot and doing a single sweep will usually have variations at the upper frequencies.

It also throws me off with your graphs linear spacing instead of log spacing on the x-axis....

Oops. Sorry about the x-axis. These are log.

 

[CBM]

Do you adjust delay by distance on the denons? If so, when you increase distance on the sub, it actually delays the other speakers respectively (farthest speaker is at 0 ms of delay). So if this is the case, it could be that the other speakers were delayed differently from the beginning (they usually are some step like 0.2ms).

It also could be that using a single spot and doing a single sweep will usually have variations at the upper frequencies.

It also throws me off with your graphs linear spacing instead of log spacing on the x-axis....

The Denon uses distance. Sub 1 was originally 17 ft and I reduced it to 10 ft. Sub 2 was 13.9 ft and I reduced it to 2.2 ft. So I reduced the delay of the mains.

 

[Jdunk54nl]

The Denon uses distance. Sub 1 was originally 17 ft and I reduced it to 10 ft. Sub 2 was 13.9 ft and I reduced it to 2.2 ft. So I reduced the delay of the mains.

What are all the other distances before and after? This could easily be why some small changesin frequency response. Not to mention just a single sweep will have variations too.

 

[CBM]

What are all the other distances before and after? This could easily be why some small changesin frequency response. Not to mention just a single sweep will have variations too.

LF is 10.2 ft, RF is 9.9 ft before and after.

 

[DonH56]

Note that due to phase shift in the crossover and whatever filters are applied the "distance" setting for optimal response might not match the physical distances.

 

[Jdunk54nl]

Do you have a null at 95hz in your house? Can you post single speaker measurements? Left speaker by itself, right by itself, sub by itself. Then left + sub, right + Sub.

 

[CBM]

Do you have a null at 95hz in your house? Can you post single speaker measurements? Left speaker by itself, right by itself, sub by itself. Then left + sub, right + Sub.

I think so. It shows up in all my measurements. There's also one about 77 Hz. These are what I was trying to reduce with two subs. I won't be able to take the measurements you asked for until this weekend. What are you thinking?

 

[Jdunk54nl]

It is interesting to me that you had to set your subs to a shorter distance. I think this is going to cause you phase issues ultimately. In all of the different setups I have time aligned (using SMAART or open sound meter to see phase in real time), I have always had to add more delay to the main speakers (or set the subs to a farther distance).

 

[CBM]

It is interesting to me that you had to set your subs to a shorter distance. I think this is going to cause you phase issues ultimately. In all of the different setups I have time aligned (using SMAART or open sound meter to see phase in real time), I have always had to add more delay to the main speakers (or set the subs to a farther distance).

When I increased the distance the dips got worse. Maybe I didn't go far enough. I'll try again this weekend. Thanks.

 

[Jdunk54nl]

When I increased the distance the dips got worse. Maybe I didn't go far enough. I'll try again this weekend. Thanks.

It is possible that you have it right, I just haven't experienced that in my limited number of cases. I obviously haven't done every scenario so it may be that your situation, you do need to have subs at some delay and no delay on your mains.

 

[ernestcarl]

Perhaps you can check your step response in REW to see where the sub’s peak is in relation to the mains.

 

[CBM]

Perhaps you can check your step response in REW to see where the sub’s peak is in relation to the mains.

That sounds like a good plan. I did try the time align tool and it wanted me to move the subs 20 ft which is more than the Denon will allow.

 

[CBM]

It is possible that you have it right, I just haven't experienced that in my limited number of cases. I obviously haven't done every scenario so it may be that your situation, you do need to have subs at some delay and no delay on your mains.

When you've tried this procedure in the past, was your full-range FR affected like mine? Do you think it's caused by phase or harmonics or something else?

 

[DonH56]

If you get the subs and mains somewhat out of phase it will accentuate peaks and dips where they interact (across the frequency band -- a crossover is not a brick wall that falls to zero on either side of the crossover). If the subs and mains are 180 degrees out of phase, you'll get a null, and similarly a peak when they are in phase. Since you can only double the signal, more or less, peaks tend to be 6 dB or less, but nulls can be very deep (20 dB or more) and thus hard to correct (virtually impossible with just EQ).

As you adjust delay you should see the curves flattening where the delay is optimal, and increasingly "wiggly" as you move away from optimal phase (in either direction). You need the subs and mains in phase at the main listening position (MLP) with the added constraint that the subs must be adjusted to reduce the impact of room modes if possible. That usually means placing the subs appropriately to counter the nulls (see e.g. Todd Welti's papers) and adjusting their phase (or delay) relative to the main speakers to dial everything in.

It can be tedious but is usually rewarding.

FWIWFM - Don

 

[CBM]

If you get the subs and mains somewhat out of phase it will accentuate peaks and dips where they interact (across the frequency band -- a crossover is not a brick wall that falls to zero on either side of the crossover). If the subs and mains are 180 degrees out of phase, you'll get a null, and similarly a peak when they are in phase. Since you can only double the signal, more or less, peaks tend to be 6 dB or less, but nulls can be very deep (20 dB or more) and thus hard to correct (virtually impossible with just EQ).

As you adjust delay you should see the curves flattening where the delay is optimal, and increasingly "wiggly" as you move away from optimal phase (in either direction). You need the subs and mains in phase at the main listening position (MLP) with the added constraint that the subs must be adjusted to reduce the impact of room modes if possible. That usually means placing the subs appropriately to counter the nulls (see e.g. Todd Welti's papers) and adjusting their phase (or delay) relative to the main speakers to dial everything in.

It can be tedious but is usually rewarding.

FWIWFM - Don

That makes sense. What's the best method to determine if the subs and mains are out of phase? I was only looking at below 200 hz and I improved it somewhat. Then I zoomed out to see the full-range and was surprised that adjusting the subs distance made a difference in the higher frequency response.