Different REW Measurements

[dark25.mc]

Hello, I’m Marco and I take my REW measurements with: my MacBook, Umik-1 microphone, my Focusrite Scarlett 6i6 second gen. connected with the coaxial cable to my Onkyo A9010 amplifier and my Wharfedale Mach 9 loudspeakers.

Recently taking new measurements, I noticed on REW All SPL & impedance graphic a strange behavior in lowest frequencies department.

From 60 Hz down there is a decay of frequency curve compared to my old measurements than a month ago.

I changed nothing in my listening room: such as loudspeakers or furnitures, ecc…

I examined carefully REW to know each single function well and to be sure that I have not mistaken some setting. So I decided to take new measurements with only the single channels, L and R and the frequency curve response of the single channels from 60 Hz down is higher than exactly should be, identical to my old measurements, taken only on both channels, L+R.

If I take new measurements as usual with both channels L+R, the frequency curve is again below 60 Hz.

In addition, according to my recent measurements with just the single channels L and R and with my friend suggestion I checked the connection phase of my loudspeakers and everything is ok.

I decided to make an attempt, a further test, reversing just one loudspeaker phase and listening some music I noticed that from the other loudspeaker, the in phase one, the lowest frequencies were suddenly more presents as never before, it seemed like a bass booster and woofer went down very much.

Maybe there is a sound waves phase shift due to wrong loudspeakers position: a delay or an advance between a loudspeaker and the other one can provoke waves bouncing, they come late and the frequencies themselves emitted from the loudspeaker are reduced.

Finally, I have some questions about how REW works.

How do I take the measurements? With both channels or the single channels separately?

While I was taking the measurements, I understood what reference output do: the sweep starts from the reference loudspeaker and it ends to the channel test.

I would like to understand what is the output reference useful for? Is it necessary for the measurements or can I avoid it?

Among the options there’s also the possibility to take measurements without reference output, disabling the timing reference the sweep is just emitted from the indicated channel.

I attach REW measurements files: an old measurement L+R only and recents ones, L, R and L+R too.

 

[John Mulcahy]

In the new measurements the right speaker has inverted polarity, the impulse response is inverted (use the percent axis on the Impulse graph to see that).

There is something odd with the old measurement, it has a very large clock adjustment value suggesting a problem with the timing reference detection.

Measure speakers individually. The reference speaker has the timing reference signal sent to it, it does not receive the measurement sweep unless it is also the speaker you are measuring. Using a timing reference allows meaningful comparison of the relative timing of different measurements, provided the mic was in the same position for each of them.

 

[dark25.mc]

I tried to take new measurements but always with the same results!

 

[skid00]

Swap the cable leads on *one* of the speakers. Put the positive cable lead into the negative speaker port. That will fix your phase difference.

 

[dark25.mc]

Ok, I tried your solution, from the loudspeaker in phase the lowest frequencies were more pronounced but the stereo image was strange, out of phase in the middle of the listening point.