Helping to solve Phase measurements being inconsistent with acoustic timing reference and Analog Loopback question

[rajopotamus]

Hello All-

Part 1:

I wanted to pick your brains to help resolve the issue im having regarding the phase plots. First I will give you a little bit of a background on my setup and goals. Im currently trying to time align, and eq my JBL 4722 LCR two way active speakers. My ultimate goal is to use REW to take accurate frequency and phase measurements and export them to FIR designer and create an FIR Filter to linearize the phase and eq my LCR speaker drivers (HF and LF units) as well as tiime align them.

Current Setup:

Microphone - Umik2
Computer - MacBook Pro running MacOS Sequoia
Avr - Monolith HTP-1
DSP - QSC Core510i
50 foot Optical USB-C cable
3 foot HDMI cable

Currently I have the microphone directly connected to the MacBook via a long (50 foot) optical usb-c and that is connected to my HTP-1 via HDMI which goes to the QSC Core DSP and out to my amps/speakers. The MacBook picks up the umik2 right away and everything seems fine with that. I output audio using my HTP-1 over HDMI from my MacBook/REW

I get consistent frequency measurements, but I do not get consistent phase measurements when taking measurements of the respective HF or LF units; even when using the acoustic time reference option. After I take the measurement using an acoustic reference (my side surround speaker - not the driver im measuring for the LCR), I estimate the delay and unwrap the phase. The general shape is pretty consistent but it moves up and down on the phase plot vertically and I can't seem to get the same repeatable results which makes me unsure of what actual phase plot is accurate.

I am not sure what is causing this but some past posts have stated there may be issues with the fact that the UMIK2 has its own clock, and the Output sound card has its own clock; which could lead to inconsistencies. I do have a long run (40 feet) as my equipment rack is immovable and is well outside my theater. I am currently connecting the UMIK 2 to my MacBook with an optical Usb - C cable since it's a long distance. The HDMI connection from my MacBook to the HTP-1 processor is only a couple feet. I am lost on what if anything I could do to get more consistent phase measurements. I would appreciate any help in troubleshooting this.

Part 2:
I read in prior posts that a USB interface and XLR mic might be a better idea for more absolute timing references using an analog loopback.

I ordered a Motu ultra lite M5 and an earthworks M23R analog XLR mic.

I understand that I could create an analog loopback by plugging my mic into the first XLR mic input of the motu, then using a TRS cable from output 1 into XLR mic input 2 to effectively create the loop back timing reference. This is where things start to get a little muddy for me.

In REW I would be selecting Motu M5 as the output correct? So how would I get the appropriate channels ( Left, right, center, side left, side right, etc..) as if I was playing sound through the HTP1 when it was selected as the output? Would I be bypassing my HTP-1 altogether and trying to connect directly to my QSC Core 510i inputs (balanced euroblock connectors?) I guess I am a little confused how to implement this solution in my setup. If anyone can shed some light on that for me that would be great! Thank you!

Sincerely,
Raj

 

[John Mulcahy]

After I take the measurement using an acoustic reference (my side surround speaker - not the driver im measuring for the LCR), I estimate the delay and unwrap the phase.

If you use Estimate IR delay on each measurement you are abandoning the timing reference and removing relative timing information. Measure your tweeter using the acoustic ref, then use Estimate IR delay and choose the option that updates the timing offset. Don't use Estimate IR delay again unless you change the position of the mic relative to the reference speaker (better if you don't). Unwrapping phase for room measurements will typically produce a plot that has a very large span as phase can undergo complete rotations at deep notches. Unwrapping also happens relative to the cursor position, so make sure it is at the same frequency whenever you use it. Not sure why you need to unwrap through.

As long as you have the Analysis preference option "Adjust clock with acoustic ref" selected REW will take core of any clock rate differences.

 

[rajopotamus]

Thanks for getting back to me so quickly!

I was only unwrapping so I could see it relatively. I didn’t realize it was dependent on cursor position.

 

[rajopotamus]

edit

 

[rajopotamus]

If you use Estimate IR delay on each measurement you are abandoning the timing reference and removing relative timing information. Measure your tweeter using the acoustic ref, then use Estimate IR delay and choose the option that updates the timing offset. Don't use Estimate IR delay again unless you change the position of the mic relative to the reference speaker (better if you don't). Unwrapping phase for room measurements will typically produce a plot that has a very large span as phase can undergo complete rotations at deep notches. Unwrapping also happens relative to the cursor position, so make sure it is at the same frequency whenever you use it. Not sure why you need to unwrap through.

As long as you have the Analysis preference option "Adjust clock with acoustic ref" selected REW will take core of any clock rate differences.

Thank you so much for clearing this up for me!

One more questions on how I should measure these drivers.

I am m measuring each driver in the speaker independently; in my case the HF and LF drivers. If I wanted to take a measurements at the MLP of all my LF units in the L, C,and R. Will each driver need its own acoustic timing reference since the speakers are in different locations relative to the mic? Addiotjnally; if I take measurements of the LF and HF drivers of the same speaker, with the same mic location; does each driver needs its own acoustic timing reference?

 

[John Mulcahy]

No, they can and should all use the same reference since the mic should be staying in the same place.