Another request for REW setup and use help

[Mikey757]

Hi,
I’m a retired software engineer, audio enthusiast, built many speakers over the years, a few amps, play bass guitar, etc.
Finally got around to getting actual test gear with REW. Great tool. Thanks!!
However, a little, um, confusing on setup.
I have a laptop running W11 (and a desktop Intel pc with ubuntu), a umik-1, several good amps to choose from.

So, first question: I did setup and soundcard calibration on the laptop, seemed to work ok (java devices; didn't want to get into a research project with ASIO).
Set up speaker/amp levels and did initial sweep. Levels ok, but get a warning that distortion was high, about 48%. ???
I could not hear any significant distortion so should I ignore or am I missing something? TIA
--
mike

 

[Keith_W]

1. Don't post .TXT files. Post MDAT's.
2. You may be measuring too loud if distortion is that high.

 

[sm52]

Levels ok, but get a warning that distortion was high, about 48%. ???

Something, some additional signal, is mixed into the test signal. Maybe from the OS mixer. Maybe you specified the wrong inputs and outputs on the REW sound card tab. If you choose to use Java drivers, select those that begin with Excl.

 

[Mikey757]

1. Don't post .TXT files. Post MDAT's.
2. You may be measuring too loud if distortion is that high.

Ok.
Tried lowering levels but level set shows too low. Even so, I still get high distortion. Same on an Intel laptop and Intel desktop running Ubuntu. Both using java devices.
Dumb question maybe, but should I disable all other audio sources on the pc?

 

[Mikey757]

Something, some additional signal, is mixed into the test signal. Maybe from the OS mixer. Maybe you specified the wrong inputs and outputs on the REW sound card tab. If you choose to use Java drivers, select those that begin with Excl.

Tried both Excl and non. Same results. Could be OS related. Sigh. Not looking for a research project, just want it to work. Should I get a different platform?
Thanks.

 

[Keith_W]

Ignore REW's warning and post your MDAT. We may be able to determine the cause of the distortion from examining your measurement.

In general, causes of high distortion in REW measurements may include:

- it's real. You have a really bad speaker.
- USB power/clock interference - see this thread and read John's comments. Try a different port, try a shorter cable, etc.
- High noise floor being interpreted as distortion (especially low freqs)
- OS issues, e.g. not using WASAPI Exclusive mode (as already alluded to)
- Microphone issues
- Unfortunate coincidences (e.g. if you measure a null at 100Hz, and there is a peak at 200Hz, it will appear that harmonic distortion is massive).

 

[sm52]

Should I get a different platform?

No. REW should work on Win11. Look in the sound properties for your sound card for all the tabs where there may be Windows sound 'improvements'. If they are enabled, disable them. And check 2 boxes for exclusive rights. I turn off all audio devices except those used for measurement when I take measurements.

 

[Mikey757]

Ignore REW's warning and post your MDAT. We may be able to determine the cause of the distortion from examining your measurement.

In general, causes of high distortion in REW measurements may include:

- it's real. You have a really bad speaker.
- USB power/clock interference - see this thread and read John's comments. Try a different port, try a shorter cable, etc.
- High noise floor being interpreted as distortion (especially low freqs)
- OS issues, e.g. not using WASAPI Exclusive mode (as already alluded to)
- Microphone issues
- Unfortunate coincidences (e.g. if you measure a null at 100Hz, and there is a peak at 200Hz, it will appear that harmonic distortion is massive).

Mdat is from 8" 2 way spkr measured 1 meter away, vertical. hth

 

[Keith_W]

I think your "distortion" is noise floor.

Let's look at this null at 66Hz.

It corresponds to this peak in the distortion graph.

Nulls bring the signal closer to the noise floor, as you can see here.

In fact, you can map every null in your SPL graph to every peak you see in the distortion graph.

Take away: ignore it. You have a tonne of comb filtering going on there. You need to take individual measurements of each to make sure that you haven't accidentally wired one of them with inverse polarity.

 

[sm52]

Mdat

The input device is listed as a computer microphone. But there must be Umik. The test signal level is only 20 dB above the noise level. Try raising the test signal level by 10-20 dB.

 

[Mikey757]

I think your "distortion" is noise floor.

View attachment 92571

Let's look at this null at 66Hz.

View attachment 92572

It corresponds to this peak in the distortion graph.

View attachment 92573

Nulls bring the signal closer to the noise floor, as you can see here.

In fact, you can map every null in your SPL graph to every peak you see in the distortion graph.

Take away: ignore it. You have a tonne of comb filtering going on there. You need to take individual measurements of each to make sure that you haven't accidentally wired one of them with inverse polarity.

Wow, interesting. All that info from the mdat file? Apparently I need to rtfm more. :-) Very helpful, thanks!
The crossover is at 3kHz and w and t are wired same polarity (according to the crossover connections). I had reversed and got a big dip at 3kHz. That's at 1 meter. Further out the "cone" narrows and response gets better. However, the mid-highs slant down and I would prefer straight or up. I could test each driver separately but it sounds more like room/enclosure/mic placement issue(s). Pointers on help files/docs to start with much appreciated. Thanks for your time, much obliged.