Using REW to measure DAC/ADC distortion?

[txbdan]

Hi all,

I'm planning to measure the distortion of my JBL subwoofer's passthrough circuit using my MOTU M4. I'm currently testing by looping my M4's output to its input and using REW to generate a 1khz tone and measure the FFT. It seems to be working well, but I was wondering what settings I should use to optimize this?

Same rate?
Generator tone dithering?
FFT window, length, averages, etc?

It seems you can just click the "Show distortion" button to get some nice specs.

Here's where I'm at now. One thing is that the FFT jumps around a bit so its a bit random as to when I click stop to collect specs. Should I increase averaging until its stable?

 

[Lambda]

Choose The Right FFT Window Function When Evaluating Precision ADCs

When evaluating the dynamic performance of precision ADCs using FFT analysis, coherent sampling provides the best results. In cases where coherent sampling cannot be achieved,...

www.electronicdesign.com

https://download.ni.com/evaluation/pxi/Understanding%20FFTs%20and%20Windowing.pdf

I would say use 1M FFT lengt with > 4 averages at 192khz

Generator tone dithering?

Yes and https://www.roomeqwizard.com/help/help_en-GB/html/siggen.html#locktofft

From there uses the same setting and you can make good relative comparisons.

 

[mdsimon2]

You want the sample rate to match whatever you have in Audio MIDI Setup for your capture / playback device to avoid Core audio resampling.

If you are running a single instance of REW you will want the input and output sample rates to match. If you have different capture and playback devices you can use different sample rates for playback and capture but it requires you to run two instances of REW.

If you are running at 24+ bit I do not find dither useful although it might be helpful it you were making digital in / out measurements. In general I use 128K FFT, Blackman-Harris 7, 8 averages and 20-22K. I also like use the Calibrate Level functionality by measuring output voltage with an accurate DMM and using that value for calibration, this is useful for noise measurements although the M4 has rather high noise itself so will be of limited use in an absolute sense.

As suggested earlier most important bit is to use the same settings so your results are equivalent.

Michael

 

[LTig]

Hi all,

I'm planning to measure the distortion of my JBL subwoofer's passthrough circuit using my MOTU M4. I'm currently testing by looping my M4's output to its input and using REW to generate a 1khz tone and measure the FFT. It seems to be working well, but I was wondering what settings I should use to optimize this?

Same rate?
Generator tone dithering?
FFT window, length, averages, etc?

It seems you can just click the "Show distortion" button to get some nice specs.

I think if you use 32k FFT with 4 averages at 44/48 kHz 24 bit with dither the resulting spectrum can be compared with @amirm's dashboard. For the multitone you need 192 kHz.

 

[acoustic1]

I have been using the same process to try and measure the performance of a cheap streamer (Tibo Bond Mini) with REW. The issue is, how to get a clean 1kHz tone into the streamer. I usually listen with Spotify so I've played test tones from their library. Most tones seem to be messy and distorted and its impossible to tell what is the streamer and what is the source signal. Is there a better way?

 

[Toltek]

I have been using the same process to try and measure the performance of a cheap streamer (Tibo Bond Mini) with REW. The issue is, how to get a clean 1kHz tone into the streamer. I usually listen with Spotify so I've played test tones from their library. Most tones seem to be messy and distorted and its impossible to tell what is the streamer and what is the source signal. Is there a better way?

If your streamer can play local files, just generate .WAV from REW

 

[acoustic1]

Ah yes! Thanks. I didn't think the Tibo could play local files, but it can (very awkwardly).

In case anyone is interested, here are some basic measurements of the Tibo Bond Mini (which has analogue out only), made using an old M-Audio USB Transit and REW:

M-Audio Transit loopback with 1kHz tone made in Audacity:

Tibo Bond Mini playing the same Audacity tone (.WAV) from the Tibo app:

Tibo Bond Mini playing a 1kHz tone by Audiolab on Spotify (ugh! what's going on there?):

Tibo Bond Mini streaming the cleanest 1kHz tone I can find on Spotify from an album called "Irritating Noises Vol.2" :

So in conclusion... I'm going to get a Wiim Mini!