REW subwoofer delay measurements don't make sense

[paul8]

I followed to a degree this article to measure sub delay: https://www.minidsp.com/applications/acoustic-measurements/292-rew-timing-ht.
The problem is when sending LFE signal in measurement window with L front acting as acoustic reference the system delay value comes as negative for sub, something like -8ft. OK, I then increase distance to my sub in my Denon receiver, actually maxing it out - then measure again and all I get is same -8ft delay, sometimes -7ft, and next time -11ft, the results are not consistent. Crossover is set to 40 Hz, sub out is LFE only, I use a Macbook Pro for this exercise.
Can the "system delay" value be even relied on in REW? Or am I doing something wrong?

Thanks.

 

[John Mulcahy]

Probably best attaching some example measurements in an mdat file.

 

[paul8]

Problem solved, apparently default channel mapping on Mac is wrong, center and lfe need to be swapped in REW settings.

 

[John Mulcahy]

Your LFE measurements don't cover a sufficient frequency span, measure to at least 200 Hz, higher if the response has not dropped off by then.

 

[paul8]

One observation: after Audissey calibration applied REW shows LFE system delay as 4 ft. If I fix this by increasing sub distance, FR shows early drop off in bass freq, so looks like Audissey is right.

 

[paul8]

Actually, if I fix the system delay made by Audyssey and then boost sub level - bingo FR and sound is even better! REW rules.

 

[paul8]

So, I aligned my sub using the system delay value, however when I use alignment tool at most frequencies it shows at least 2 ms difference between my fronts and the sub.
What would be more accurate method for the sub alignment, the alignment tool or the system delay vs an acoustic reference?

 

[John Mulcahy]

Delay measurements are unreliable for low bandwidth devices like subwoofers, the alignment tool is best for that.

 

[paul8]

I'm trying to measure a new subwoofer, if I align the sub by using delay value against an acoustic reference sound wise it actually makes sense - I clearly see SPL increase, so looks like sub and speakers are aligned and SPL goes up. However looking at the REW's alignment tool at 60 Hz which is my crossover - it still shows that my sub is off by no less than 4 ft!
What gives?
Attaching my mdat.

 

[sm52]

Try 2 options. If you haven't changed anything since the last published measurement. Sub -4ms or -1.376m. Second. Sub -5ms or -1.548m.

 

[JStewart]

However looking at the REW's alignment tool at 60 Hz which is my crossover - it still shows that my sub is off by no less than 4 ft!
What gives?

Different latency on the two different electrical paths is a common reason. For example, the popular MiniDSP 2x4 HD adds 2-3ms iirc. Does your new sub have any DSP on board?

 

[paul8]

Different latency on the two different electrical paths is a common reason. For example, the popular MiniDSP 2x4 HD adds 2-3ms iirc. Does your new sub have any DSP on board?

Which different electrical paths? REW shows near 0 delay in its IR delay measurement, yet 4ft in its alignment tool using same IR data. My question is why the difference?
BTW, I flipped polarity on the sub to get more SPL at the crossover frequency, aligned again using the delay value, now the alignment tool shows -30 ft difference, which makes 0 sense to me, my AVR cannot even adjust to such distance.

 

[JStewart]

Which different electrical paths?

Between the sub and the mains which is how I understood the question and made specific mention of a product that is often used between an AVR and subwoofer.

REW shows near 0 delay in its IR delay measurement, yet 4ft in its alignment tool using same IR data. My question is why the difference?

John M did already say that delay measurements are unreliable for subwoofers and to use the alignment tool. When using the alignment tool with a subwoofer one would align by phase not IR (peak). Is this how you're doing it?

Apologies if I'm still not understanding. If that's the case perhaps a couple of screen shots will help clarify for myself or someone else.

An mdat file with sub only, left only, and right only and with the crossovers in place on the avr, may also be useful to anyone who may wish to help.

 

[sm52]

REW shows near 0 delay in its IR delay measurement

Because t=0 for a subwoofer impulse REW sets at the beginning of the impulse. And the subwoofer impulse is long, so its peak or another place close to the peak is far from the beginning of the impulse. Therefore, you must independently establish the position of the subwoofer in time.

 

[paul8]

Between the sub and the mains which is how I understood the question and made specific mention of a product that is often used between an AVR and subwoofer.



John M did already say that delay measurements are unreliable for subwoofers and to use the alignment tool. When using the alignment tool with a subwoofer one would align by phase not IR (peak). Is this how you're doing it?

Apologies if I'm still not understanding. If that's the case perhaps a couple of screen shots will help clarify for myself or someone else.

An mdat file with sub only, left only, and right only and with the crossovers in place on the avr, may also be useful to anyone who may wish to help.

I tried alignment tool with both phase and impulse, they show approximately same value, which is hard to make sense of. The mdat is attached to post #9

 

[JStewart]

I'm having trouble making sense of the measurements. Perhaps someone else can provide some definitive insight.

First measurements 1,3,4,5 show a reflection approx 5.5ms after the initial peak that is equal to the initial peak. Its the same delay for all the measurements. Was all upmixing off in the AVR when the measurements were done? That's all I can think to guess.

Looking at the frequency response I don't see a right speaker only measurement that doesn't include the sub. At least that's what it looks like. Right measurements 3,4,5 look like right with sub at different levels.

There's guidance for collecting measurements to align a sub to mains in this post if it helps.

Regarding multi-sub alignment, sub to main alignment and EQ set up

Hi, Just want to get a few things clear regarding multi-sub alignment, sub to main alignment and EQ set up. I been using REW a few time and know the software I believe quite good, but don’t always know how to set up measurements or how to do the analyses. Just did some playing around a few...

www.avnirvana.com

 

[sm52]

I'm having trouble making sense of the measurements.

I think the author of the thread is not using a filter for the right and left sides. Allows them to play whatever they know how.

 

[JStewart]

I think the author of the thread is not using a filter for the right and left sides. Allows them to play whatever they know how.

I think the same.
Might you have a possible cause for that 2nd impulse peak in the measurements?

 

[sm52]

Might you have a possible cause for that 2nd impulse peak in the measurements?

Up to 5.5 ms theim pulse starts normally. Then the second peak and the thickness of the noise increases. If it were a reflection, the thickness of the noise would not increase. Therefore, I assume the appearance of a second source producing another impulse.

 

[paul8]

I think the author of the thread is not using a filter for the right and left sides. Allows them to play whatever they know how.

Yes, for those measurement for L and R sub was enabled, but since timing signal is limited in frequency and not going through sub anyway what difference would it make to exclude the sub? Acoustic reference which is L speaker would still measure same right?

 

[sm52]

since timing signal is limited in frequency and not going through sub anyway what difference would it make to exclude the sub?

A whistle, with acoustic synchronization, is not a measuring signal; while it is sounding, REW remembers only the moment of arrival of this signal. The graphs we see are obtained using the second signal, which typically has a full bandwidth of 0 to 24 kHz.
To correctly match the sub and main speakers, you need to have normal measurements of each of them separately with the same acoustic reference. Such a reference could be L. Without a connected subwoofer.