Strange RTA Measurement

[Ben99]

Hello,

Just started tuning my first car audio and this is probably a very noob question. Does my measurement via RTA look weird? It's almost completely flat from 20Hz onwards.

Setup:
ECM8000 RTA and BOMGE BMG22 for audio interface.
RTA is connected to Input #1
Output #2 loops back to Input #2
Output #1 is connected to my car DSP through RCA "in" ports. DSP setup is default meaning no EQ and no other filters applied.

Sound card was calibrated (see screenshot). Pink Periodic noise with full range setup was selected (see screenshot).

I have also tested Sine tones at 1,000 Hz and the RTA picks up that signal just okay.

Any help is appreciated.

Full range pink periodic noise RTA measurement

Sine tones at 1,000 Hz

Sound card calibration

 

[John Mulcahy]

Does my measurement via RTA look weird?

Yes. Looks like a loopback response with a mic cal file added, which makes sense since you say you have a loopback from output 2 to input 2 and you are measuring the R input, which is input 2 (L = 1, R = 2). To measure the mic input you need to select L.

A couple of observations:

• You are running quite an old version.
• You should use a window other than rectangular for sine tones which are not locked to the FFT length.

 

[Ben99]

Yes. Looks like a loopback response with a mic cal file added, which makes sense since you say you have a loopback from output 2 to input 2 and you are measuring the R input, which is input 2 (L = 1, R = 2). To measure the mic input you need to select L.

A couple of observations:

• You are running quite an old version.
• You should use a window other than rectangular for sine tones which are not locked to the FFT length.

Thank you, John. The measurement looks much better after adjusting the mic input. A couple of further questions:
- Do you think the setup from RTA window looks okay for measuring pink periodic noise?
- I have overlays of comparison between (green) line with mic calibration and (blue) line with no calibration. I hear mixed comments about using calibration file for ECM8000. what are your thoughts?
- I measure each speaker separately for now just to sort out the issues I had earlier. Going forward, should I be measuring left then right speakers separately or all speakers at once is okay?

 

[John Mulcahy]

Do you think the setup from RTA window looks okay for measuring pink periodic noise?

Yes, though you don't need averaging unless you are moving the mic around to make a spatial average.

I hear mixed comments about using calibration file for ECM8000. what are your thoughts?

I wouldn't use the generic ECM8000 cal file.

should I be measuring left then right speakers separately or all speakers at once is okay?

Measure speakers separately.

 

[Ben99]

Yes, though you don't need averaging unless you are moving the mic around to make a spatial average.


I wouldn't use the generic ECM8000 cal file.


Measure speakers separately.

Appreciate your tips, John.

I also stumble on the time alignment topic.

The measurement I took with my car rear right speaker indicates negative delay distance. My guess is the setting is incorrect, but I can't seem to figure out why. At the moment output 2 loops back to Input 2, RTA mic = input 1, and output 1 connects to DSP.

 

[John Mulcahy]

The measurement I took with my car rear right speaker indicates negative delay distance.

A negative delay just means the sound from that speaker appears to arrive sooner than the sound from the reference speaker. The two measurements whose images you posted are using different methods to estimate timing, the FL is using estimated IR delay (the usual method for a full range measurement) but RR is using the IR start time (unusual, typically only applied to subwoofer measurements). Reflections may be confusing things, but it isn't possible to say much based on screenshots. You should attach an mdat file if you have a query about some measurements.

 

[Ben99]

A negative delay just means the sound from that speaker appears to arrive sooner than the sound from the reference speaker. The two measurements whose images you posted are using different methods to estimate timing, the FL is using estimated IR delay (the usual method for a full range measurement) but RR is using the IR start time (unusual, typically only applied to subwoofer measurements). Reflections may be confusing things, but it isn't possible to say much based on screenshots. You should attach an mdat file if you have a query about some measurements.

Hi John,
I have attached the mdat time alignment file using the acoustic timing reference option. I am using loopback of output 2 to input 2 as reference. What I don't understand is why the result shows as negative delay? I would imagine the reference signal should arrive before the actual sound coming from my left speaker.

Another question on EQ tuning

The DSP i'm using is very limited in its tuning capability. It doesn't allow me to tune each channel separately, but only combined front channels then rear channels separately. It also has fixed EQ frequency adjustment.
Other than that, the speakers are stock from factory (Front = tweeter and 6.5" mid bass for each left and right channel. Rear = only mid bass for each left and right channel)

Below is the measurement snapshot from both front left + right speakers. I am lost on the target settings and basically not sure where to get started.

 

[John Mulcahy]

The measurement was made with the timing mode set to acoustic reference.

Change the timing mode on the Measure dialog to use the loopback reference.

 

[Ben99]

The measurement was made with the timing mode set to acoustic reference.

View attachment 63109

Change the timing mode on the Measure dialog to use the loopback reference.

Ah, I got it now. Measurement Looks much better. However, the measurement from RR(rear right) speaker seems to be quite off as signal took the longest out of all other signals to reach the microphone. (car is right hand drive so the microphone is at driver's position). I tried those time delay numbers with DSP and it didn't sound right. I understand that reflection might play a role here, but is there any technique to help find the right distance?

Car is a regular saloon. All speakers are stock from factory, so each mid-bass on each door and tweeters on each A pillar.

 

[John Mulcahy]

In a car the largest peak can be from later reflections rather than the direct sound, so the delay estimate gets confused. You can use the IR stat time figures, if you open the Info panel (Big "i" above the list of measurements) the IR start time is shown just over half way down. select each measurement in turn to see its data.

 

[sm52]

Seems like the polarity is reversed in RR.

 

[Ben99]

In a car the largest peak can be from later reflections rather than the direct sound, so the delay estimate gets confused. You can use the IR stat time figures, if you open the Info panel (Big "i" above the list of measurements) the IR start time is shown just over half way down. select each measurement in turn to see its data.

Thanks for the tips, John.
Based on that, the rear right speaker took the longest to reach where the RTA mic is positioned (front right seat):
Speaker | start time(ms)
FR | 5.89
FL | 6.12
RL | 6.58
RR | 7.48

This is interesting because the measurement I took with plain measuring tape shows the opposite:
Speaker | length(cm)
RL 126
FL 130
FR 90
RR 86

I'm starting to think if there is a problem with the RTA microphone. As a backup, can you recommend an inexpensive calibrated microphone? The Dayton Audio EMM6 microphone from Cross Spectrum is sold out.

 

[Ben99]

Seems like the polarity is reversed in RR.

I just noticed that too. Checked with polarity app and it says phase is positive. I tried reversing the phase through DSP and it sounded out of phase with that..

 

[sm52]

Checked with polarity app and it says phase is positive.

Other speakers have positive polarity. All 4 measurements were taken with one microphone. So RR actually has negative polarity. Reverse the polarity at the RR terminals and measure the 4 speakers again with your microphone.

 

[John Mulcahy]

It can be pretty difficult to determine polarity from the impulse response in a car with so many reflections. You can use the signal generator to make a polarity test signal and watch it in the scope with the mic close to the driver, see https://www.roomeqwizard.com/help/help_en-GB/html/siggen.html#sinewave

 

[John Mulcahy]

Are you sure your left and right connections from REW haven't got swapped?

 

[Ben99]

Are you sure your left and right connections from REW haven't got swapped?

I don't quite understand it. Can you please explain more?

 

[John Mulcahy]

When you play something on the left channel in REW does it come out of the left speaker?

 

[Ben99]

When you play something on the left channel in REW does it come out of the left speaker?

Yes and I do control channels via a DSP

 

[John Mulcahy]

Perhaps the DSP is applying some delays.