How to - Distortion Measurements with REW

You may want to do a calibration curve to normalise the response of the soundcard, as it may roll-off the amplitude quite significantly below 10Hz. On my EMU0404, I also get a lower inherent roll-off corner frequency with 44kHz sample rate compared to 192kHz sample rate (due to the AK5385 HPF configuration). That could allow easier assessment of harmonic levels if the fundamental is down towards 2Hz.
 
  • Like
Reactions: sajadho1988
My 2 cents: The Solo is fine for single channel measurements like amplifiers. However, for measuring loudspeakers you'd probably want the 2i2 so you could take advantage of dual channel measurements. The Solo limits you to single cha
Can't imagine why you'd want to do 2-channel microphone measurements. The Solo is a 2-channel interface with the mic on Channel 1 and you can use Channel-2 (TS/TRS) for loopback.
 
Last edited:
Having taken measurements but never really setting my rig up properly or learning anything other taking distortion measurements I've decided to start from the beginning with probably really stupid obvious questions! My rig consists of at the moment a Motu M2 and Autoranger.

So first do I need to calibrate the sound card and if so how to do this? Hitting the calibrate soundcard button brings up this window
calibrate screen 1.png


So reading this correctly output 1 to input 1, I'm assuming that's what they call external loop back?

So next screen
calibrate screen 2.png


Am I using the generator to create a (for instance) a -10dbfs wave? If so the meters on the calibrate page look like this
calibrate screen 3.png


hitting the next button again I'm met with this change on the meters
calibrate screen 4.png


I take it the aim is to match the in signal to match the out? Do I need to take notice of the Ref in level?
calibrate screen 5.png


I'm guessing this is close enough? Pressing next makes REW make measurements not sure what I'm looking for / at here so don't know if this is correctly set up?
calibrate screen 6.png


calibrate screen 7.png


If this is all the correct way to do this do I select 'finish' and then save the cal file?
 
I will be following along since I am even more of a novice. Could you define what you want to measure and what your objectives are. What is “a Motu M2 and Autoranger”?

I think we actually need a separate thread for beginners. This thread is rather daunting for the uninitiated..
 
I've been reading through this thread and finding my way around the REW software and finally got round to hooking up my F4 to the Focusrite 2i2 and took some distortion measurements this evening. The image shows the F4 in parallel mono connection (both channels connected together in parallel) running with a bias giving ~320mV across the source resistors in both channels. Driving an 8 ohm load.

PXL_20241215_070751448.jpg


Nice clear harmonics off the 1kHz fundamental. By changing the bias I can get the distortion to reduce as the bias increases so all looking good and as expected 🙂 I'm a little concerned about the noise - looks like its power supply related with 60 and 120Hz and associated harmonics. There's also a load of HF hash, which seems to get better / worse depending on the routing of the twisted pair going from the RCA connectors to the amp PCBs.

F4 (B2+B1 parallel) 320mV bias 2164V in 2102mV at 8 ohm load - signal input pink twisted pair ...jpg

I'm using 400VA dual 18V torroid and the store's universal power supply. Is what I'm seeing here to be as expected in terms of noise or is the noise a bit high and is there something I can do to reduce this?
 
Perhaps worth appreciating that afaik the soundcard's output and input use a common ground, but you have connected each of those signal line grounds to different points in your test setup. That is likely one reason for the extraneous signals. You could modify the amps' wiring to suppress that noise loop, or remove one ground connection (such as the signal input gnd to the soundcard that appears to come from your load), or just not be concerned about those artefacts.
 
  • Like
Reactions: Mobes123
I will be following along since I am even more of a novice. Could you define what you want to measure and what your objectives are. What is “a Motu M2 and Autoranger”?

I think we actually need a separate thread for beginners. This thread is rather daunting for the uninitiated..
Hi François,

What is want to do is measure distortion of amps I've repaired over 1khz and then over a sweep of frequency and be confident the right is set up correctly.

The Motu M2 is an external audio interface like the focusrite 2i2 and the Autoranger is a Jan Didden design where it takes the output from an amp and automatically limits the voltage seen by the Audio interface so it doesn't damage it
 
  • Like
Reactions: Francois G
To get REW to show the input and outputs to be the same with a loopback test, I had to add a mic cal file.
Its in Preferences, Cal Files, Mic cal section and click on gear wheel.

I checked the actual levels with a true rms multimeter. Most are reasonably accurate upto 1khz but I recently bought a Unit-T UT61E+ which has a 40-10kHz bandwidth.
(It's this value i.e. the input that you enter into the box)
 
Hi

Sorry for the newbie’s question, I’ve never tried that type of measurement.

Would the Scarlett Solo 3rd generation be ok ? I have no clue if it’s supported by REW and includes any Java supported card…

Picture of what I intend to buy

IMG_4891.png
IMG_4890.png
IMG_4888.png
IMG_4889.png


Thanks
Eric
 
I use a Behringer UCA222 - you can't get much cheaper but it works well at 44 & 48Khz and as there are no controls, it is easy to use.

Importantly for me,. it has an spdif toslink port that is ideal for testing Dacs.

I made a -20/-26db attenuator for testng amps. (you;ll need one whichever audio interface you use)

I found the problem with those with input and output level controls is that you must also connect an RMS voltmeter otherwise adjusting them after doing the calibration in REW defeats the purpose of this. Far, far easier not to have controls!

If you want higher res sampling, an iRig HD is good for 96Khz and only has an input gain control but no spdif.
 
Last edited:
  • Like
Reactions: e_fortier