Problems with Rev V5.19 Beta 7 on my Mac

[mike w]

Hi John: I noticed that when I measure using "Both" speakers the results is a much higher dB response than measuring either Right or Left speaker by itself for higher frequencies only. This just started happening when I downloaded the latest version (V5.19 Beta 7). In previous versions, there was no difference. I can post examples if you like.

regards,

mike

 

[John Mulcahy]

Worth posting some mdat files showing the differences. Would expect a 'both' measurement to be higher by around 3 dB on average, but with lots of comb filtering at high frequencies due to the two sources - if the distances to each speaker are significantly different that comb filtering would extend lower in frequency. No reason there should be different behaviour between REW versions, but results will be very sensitive to exact microphone position when there are two sources.

 

[mike w]

Ok, I just ran three scans, L speaker, R speaker and Both speakers. Please look. I also included a scan done a couple months ago using both speakers.

Another difference in the new SW is that when I initially go to measurement mode, the output drop down menu is not highlighted (right, left, both). I have to run the first scan in whatever mode its in then it becomes highlighted in subsequent measurements to allow for change.

BTW, this is a fun tool. However, I've not been able to make any headway against the room modes. Floyd Toole's recommendations are a little expensive for me right now (multiple subwoofers). I'm reminded of that saying that ignorance is bliss. ;-)

 

[John Mulcahy]

The previous example you attached was for left only (you can see the output selection in the measurement info). The output device and output have not been selected for any of the measurements, which may explain the output selector behaviour - best selecting them in the Soundcard preferences.

The measurements look as expected. The Both measurement has two peaks corresponding to the signals from each speaker, offset by about 113 us.

If I delay the right channel measurement by 113 us and add it to the left, I get a trace almost identical to your Both measurement (it is the purple trace below, overlaying the dark blue trace of the Both measurement).

Where the signal from both channels is exactly in phase and has much the same level, the sum will be about 6dB higher than the individuals. Where they are out of phase there will be sharp dips, such as at 53.6 Hz (with no smoothing the level at 53.6 Hz in the Both measurement is about 20 dB below either of the individual measurements).

 

[mike w]

Thanks John. I'll check the settings on the sound card.