Timing reference has gone wonky

[Chuck Gerlach]

I started to do some measuring and when looking at the delay information available in the sweep data, I noticed that all of the numbers are now negative. To further test this, I used the same speaker (Right) as the timing reference (Right) , and instead of the delay being close to zero, it shows the following:

Delay -5.7560 ms (-1.974 m, -(6 ft 5.7 in))
using estimated IR delay relative to Acoustic reference played from Altitude R with no timing offset
Clock adjustment: -10.4 ppm

If I measure the right surround and use the right channel as the reference, it shows a smaller delay, but still a negative number. The mic is physically closer to the right surround - and the data looks like the following.

Delay -1.1044 ms (-379 mm, -(1 ft 2.91 in))
using IR start time relative to Acoustic reference played from Altitude SR with no timing offset
Clock adjustment: -10.7 ppm.

This just started and I am completely lost. The measurement file is attached. I have tried this on two different Macs and have restarted the Trinnov processor - all with the same result. I have also replaced the HDMI cable with the same results. Also tried it with room correction (Optimizer) off and though the delays times were slightly different, they were both wrong and both negative.

Any ideas?

 

[JStewart]

Probably not your answer, but you didn’t mention that all upmixers were off.

 

[Chuck Gerlach]

Initially they were not, but I did disable them and nothing changed.

 

[jtalden]

Do you have" loopback delay reference is IR peak" selected in REW preferences/analysis?
There is some sound preceding the main peak that may be setting that delay estimate.
- Just a guess.

 

[sm52]

Chuck Gerlach​

1. Do you need a 12kHz sample rate for some purpose or did it happen by accident?
2. Try installing a new version of REW. https://www.avnirvana.com/threads/v5-20-14-early-access-build.11154/

 

[John Mulcahy]

As mentioned above, in the Analysis preferences, Impulse response calculation select "Loopback delay reference is IR peak".

 

[Chuck Gerlach]

As mentioned above, in the Analysis preferences, Impulse response calculation select "Loopback delay reference is IR peak".

That is how it has been set for as long as I have been using REW. I am not using a USB Soundcard for these specific measurements.

 

[Chuck Gerlach]

Chuck Gerlach​

1. Do you need a 12kHz sample rate for some purpose or did it happen by accident?
2. Try installing a new version of REW. https://www.avnirvana.com/threads/v5-20-14-early-access-build.11154/

Curious? Where do you see that I am using a 12KHz sampling rate? In preferences, I have 48KHz set as the Sample Rate.

EDIT: I do see on the "Info" page that is shows a sampling rate of 12KHz??

 

[sm52]

Here.

 

[Chuck Gerlach]

I just tried the new software and am still getting negative delays. The first image below is the left speaker used as both the timing reference and the speaker to measure. In theory, the delay should be really close to zero, and as you can see, it is not.

The 2nd image is using the Left Surround as the timing reference and the left speaker is being measured. The LS is about 5 feet closer than the left speaker. So while the distance is reasonably close, the difference is negative.

I may be missing something here, but I am pretty confused at the moment.

 

[Chuck Gerlach]

Here.

Thanks. I just saw that prior to your post. Why it doesn't match the other setting shown is beyond me.

 

[John Mulcahy]

Sorry, didn't notice you were using the UMIK. The delay (and the reduced sample rate) are due to the very limited measurement range. Measuring over a limited frequency range affects the impulse response, creating artefacts that can impact delay estimates if the channel being measured is not itself limited in range (such as a subwoofer). Measure over the full frequency span for accurate delay estimates.

When the measurement end frequency is far below half the sample rate REW can lower the sample rate to reduce the amount of redundant data in the measurement. That is controlled by the "Decimate IR" Analysis preference.

 

[Chuck Gerlach]

Sorry, didn't notice you were using the UMIK. The delay (and the reduced sample rate) are due to the very limited measurement range. Measuring over a limited frequency range affects the impulse response, creating artefacts that can impact delay estimates if the channel being measured is not itself limited in range (such as a subwoofer). Measure over the full frequency span for accurate delay estimates.

When the measurement end frequency is far below half the sample rate REW can lower the sample rate to reduce the amount of redundant data in the measurement. That is controlled by the "Decimate IR" Analysis preference.

Some progress. Thanks. I am measuring to 20K now. Now when I measure using the same speaker for measurement and for the timing reference, I get a delay fairly close to zero (.14 inches to be exact). However, when I use, for example, the side surround as the timing reference and measure the front left speaker, which is about 5 feet further away than the left surround speaker, I still get a wonky delay (.06").

I'm still at a loss.

Thanks again for your assistance

 

[John Mulcahy]

If your processor has had its delays/distances correctly adjusted the delay should be close to zero for all channels.

 

[sam_adams]

Some progress. Thanks. I am measuring to 20K now. Now when I measure using the same speaker for measurement and for the timing reference, I get a delay fairly close to zero (.14 inches to be exact). However, when I use, for example, the side surround as the timing reference and measure the front left speaker, which is about 5 feet further away than the left surround speaker, I still get a wonky delay (.06").

I'm still at a loss.

Your measurement level seems a bit low:

Try raising it +6 to +10 dB and increasing the sweep length to 512K—also start at 20 Hz—as the noise level in the room seems rather high:

 

[JStewart]

@John Mulcahy , if bass management is active can that affect the delay calculation? I obviously don't know if this is the case with @Chuck Gerlach 's measurements, but I've always checked delays with speakers set to "Large" in the processor thinking that it might.

Thank you!

Edit: In the case of multiple subs phase aligned for best SPL response?

 

[Chuck Gerlach]

If your processor has had its delays/distances correctly adjusted the delay should be close to zero for all channels.

Wow! Color me "moronic". I had previously turned room correction off but it was on again. I apologize for not doing a better job of some "pre-analysis" prior to posting this.

Thank you so very much. All is well and working as it was and should.


And on a completely different subject, is there any near (mid; long) term plans to implement the "noise" option (vs sweep) when making a measurement?

 

[John Mulcahy]

if bass management is active can that affect the delay calculation?

It shouldn't, no.

 

[John Mulcahy]

is there any near (mid; long) term plans to implement the "noise" option (vs sweep) when making a measurement?

Not near term, but it will come eventually.