How to measure speaker distortion in REW

[Dj7675]

In the M106 review thread the topic of distortion came up. I have both the M106 and M16 speakers and would like to measure their distortion to see if, or when distrotion is an issue at various volume levels. Amir's distortion measurements are at 86 and 96 decibels at 1m. I would like to measure distortion at some higher levels as well as both full range and crossed over at 80/90/100hz to see what affect that has on distortion. I would like to see if distortion is an issue for both the M106 and M16 at higher volume levels if using with subs.
I have a laptop with REW and a calibrated UMIK-1. What type of measurement should I do in REW for distortion?

 

[Blumlein 88]

Do the regular sweep. One of the tabs in the results will be distortion. Click on that tab and it displays distortion levels. You can choose which harmonics and thd. Be very careful going above 96 db as that's potentially dangerous for the tweeters.

 

[Dj7675]

Do the regular sweep. One of the tabs in the results will be distortion. Click on that tab and it displays distortion levels. You can choose which harmonics and thd. Be very careful going above 96 db as that's potentially dangerous for the tweeters.

If I wanted to measure at a louder volume would it make sense to limit the frequencies during the sweep? Something like 40hz-10k? Would this make it safer on the speakers?

 

[QMuse]

In the M106 review thread the topic of distortion came up. I have both the M106 and M16 speakers and would like to measure their distortion to see if, or when distrotion is an issue at various volume levels. Amir's distortion measurements are at 86 and 96 decibels at 1m. I would like to measure distortion at some higher levels as well as both full range and crossed over at 80/90/100hz to see what affect that has on distortion. I would like to see if distortion is an issue for both the M106 and M16 at higher volume levels if using with subs.
I have a laptop with REW and a calibrated UMIK-1. What type of measurement should I do in REW for distortion?

As recommended and explained by REW author earlier on this forum stepped sine distortion measurement in RTA window will give you more accurate result than sweep.

For better accuracy you may also want to hold mic vertically (pointed to the ceiling). Be sure to apply 90deg mic calibration file in that case. Distance to the speaker should be 1m or less.

 

[Blumlein 88]

As recommended and explained by REW author earlier on this forum stepped sine distortion measurement in RTA window will give you more accurate result than sweep.

For better accuracy you may also want to hold mic vertically (pointed to the ceiling). Be sure to apply 90deg mic calibration file in that case. Distance to the speaker should be 1m or less.

Stepped gives more accurate results, but is more dangerous for the tweeters at high levels.

 

[Blumlein 88]

If I wanted to measure at a louder volume would it make sense to limit the frequencies during the sweep? Something like 40hz-10k? Would this make it safer on the speakers?

Yes, but tweeters usually work at 3khz or even lower. 40 hz will help a lot on the low end. You'll know you've gone too far just as soon as you have. Which is then to late.

You're doing something dangerous you'll have to decide the level of risk you're willing to take. Say I as someone who has blown both tweeters and woofers testing accidentally to destruction.

 

[Dj7675]

Yes, but tweeters usually work at 3khz or even lower. 40 hz will help a lot on the low end. You'll know you've gone too far just as soon as you have. Which is then to late.

You're doing something dangerous you'll have to decide the level of risk you're willing to take. Say I as someone who has blown both tweeters and woofers testing accidentally to destruction.

Thank you both for your comments. I will now be much more careful. It would be good to know, but not at the expense of the speakers!

 

[ernestcarl]

If you were using active monitors with protection limiters in place, you probably aren't likely to blow anything. Still, at very high SPLs, I'm not sure I'd try the stepped sine measurement suggested by QMuse even with active monitors... maybe at lower or medium SPLs, perhaps.

I've only ever done the usual quick (128k length) sine sweeps and this is what I got with the Sceptre S8 when pushed at it's limit:

(taken at 1 meter from HF horn opening)

*Forgot to mention -1.5dB HF shelving and -1.5dB "Acoustic Space" wide notch filter centered around 250Hz in the monitor's internal DSP were enabled. I forgot to reset it to completely factory default 'flat' -- well, it's not exactly flat as I was able to do a better job with my manual EQ to linearize the speaker's response -- before taking these measurements.

I was absolutely leery of doing this for a long time -- and wanted to only try this to see if I get better results if using a professional 4V RMS interface audio output instead of the 2V RMS I normally have connected -- answer is: Yes... and you also get significantly lower noise with balanced cables to boot. Still, I'm not completely convinced of that 116dB SPL max SPL Presonus is claiming for these monitors. Perhaps what it really means are transient peaks/bursts rather than anything sustained. But whatever... Distortion is still pretty high at such high SPLs, regardless of one's hunger for more amp power.

Doesn't make much sense to go higher.

 

[QMuse]

Stepped sine is fine with levels up to 96db/1m, but using it I wouldn't push the level higher than that.

 

[Trdat]

But how do we calibrate the level in the Generator? It mentions that "the rms output must be measured while the signal generator is playing the sine. A frequency near the mains is recommened for best results with a voltmeter."

So are we putting the two rods of the volt meter on the outputs of the amplifier to get a voltage reading then entering it?

And a frequency near the "mains" what that mean?