Simple exercise for assessing the noise levels in your system


A few weeks ago I wanted to know what is the noise level of my system, so I created a couple of WAV files containing white noise at different volumes.

Idea is simple. By playing a bunch of WAV files and adjusting the volume knob in the headphone amp until the noise can be heard. Then move to next quieter file – see F.A.Q. down below for detailed instructions.

Here’s the link for the files – :warning: use at your own risk

Just out of curiosity, how noisy is my system? Well below my own “if-I-don’t-hear-and-it-does-not-bother-me–I-simply-don’t-care” threshold. And this happens with the -96dBFS file (and amp cranked all the way up). So that is how far my objectivism goes with the gear I have. :smiley:

Anyway, hope this is helpful for the folks out there.

Cheers. :beers:


How to use those files?

Start with the loudest noise file, i.e., the one with the smallest negative dB number (e.g.: -12dbFS).

  1. Before play, turn the amp volume all the way down;
  2. Start playing the file;
  3. Slowly start cranking up the volume;
  4. If white noise is heard (as similar fashion than the louder files), turn the volume down;
  5. Move to the next quieter file;
  6. Repeat until ambient noise is higher than the ones presented by the headphones/IEMs, or an anomaly can be heard.

Why the files are 1 minute long?

Just enough time to crank the volume up and try to hear something weird, and then turn volume back down again. When done, move to the next file.

Are there any compression? Is this a lossless file?

AFAIK, this is a pure WAV file and therefore it is lossless (no compression).

How the files were created? What those numbers mean?

Files were created using a DAW (Digital Audio Workstation) with a noise generator plugin. The number in the file name means the peak (max) noise level. It is not RMS.

What kind of issues should I expect?

The simplest one is EMI/RFI. I ended up replacing a set of RCA cables because of this. Much quieter now. :smiley:

Another very common anomaly is ground loop.

I don’t want to do this.

So don’t do it then.

OK, I did it and I’m not happy with the result.

Don’t blame me. That’s the price of curiosity. :wink:

I’m sure there’s plenty of knowledge in the community to address eventual issues on a case by case basis.


Admittedly it took me a minute or 2 to figure out what the point of this would be, but actually this is quite useful, simple yet effective. Thanks for sharing!

With my Bifrost 2 (PC - USB as input) > Euforia > Empyrean I’m noise-free until -108 dB, at around 4 o’clock on the volume wheel. :smiley: Yet with my HD 660S I start getting some noise at the -96 dB file, at around 12 o’clock. Both well beyond the levels I typically listen to, so I guess my system passed the test. :slight_smile:


You got the drill. There’s actually a few other exercises that can come out from this. I’ll explain another one. I have setup my Modi/Magni 3+ combo with a cheap’o RCA cable connecting the two in a temporarily fashion as I move things around. Signal is coming via TOSLINK from a FiiO K3.

If I pick my most efficient headphone, the SRH1540 – I’m chickening out doing this exercise with IEMs for obvious reasons :hear_no_evil: – I can hear RFI at the same SPL as the recorded white noise with the -96 dBFS file. It’s quiet but it can be heard. Magni is at maximum volume and on HIGH gain.

So another exercise is simply to find the maximum volume knob which this RFI can no longer be heard – assuming I won’t replace that RCA cable nor there is something else happening. Well, the volume knob position – specifically for this setup – is at 2 o’ clock in the Magni 3+ (HIGH gain).

If I start browsing for the gear specs, this is what I have after filtering out the important bits:

Magni 3+

Maximum Power, 50 ohms: 1.6W RMS per channel


Impedance (@ 1 kHz): 46 Ohms
Maximum Input Power 1000 mW

Conclusion: So at 2 o’ clock on HIGH gain I don’t know what would damage first, my hearing or the headphone. Not that this matters under that extreme condition anyway. :smile:

Question 1: should I worry about noise on this setup? I wouldn’t.

Question 2: can you deduce SINAD from exercise above? I can try. Since I perceived both RFI and white noises at the same volume with the -96 dBFS file, a very wild guess would be a SINAD in the 80 dB ballpark. Still a ton of headroom nonetheless. See question 1 above.

Just out of curiosity, Magni 3+ with the SRH1540 is never past 10 o’clock on LOW gain.

Hope this helps.


What makes this useful imo is that it’s a way to sorta quantify audible noise levels more precisely than just by volume knob position. Most modern music plays at around -20 dBFS give or take, so with noise matching up at -96 dBFS you’d effectively be able to play music at 76 dB SPL with zero noise.

In my case with the setup I already described, my observations are as follows;

  • With my Empyrean, noise becomes audible at 4 o’clock / around -108 dBFS, when listening to music I rarely go past 12 o’clock.
  • With my HD660S, noise becomes audible at 12 o’clock / around -96 dBFS, when listening to music I rarely go past 11 o’clock.
  • With my FiiO FH7, noise becomes audible at 9 o’clock / around -84 dBFS, when listening to music I rarely go past 9 o’clock.
  • The ambient noise in my room at night when not typing/moving around etc. is around 25 dB SPL on average.

Since my maximum listening volume is right about 85 dB SPL, assuming my music most commonly plays at around -20 dBFS, it seems that in my case noise becomes audible from roughly around 20 dB SPL and up. This seems to correlate quite consistently with the masking effect that would occur from ambient noise in my room.

So what’s interesting is that even though this isn’t exact science by any means, the numbers seem to correlate with the expectations and observations to a reasonably accurate degree. Based on these numbers it would seem that in my situation a noise floor at -85 dBFS would be inaudible for me for most of my listening, and noise levels at -100 dBFS would cover even the most dynamic recordings with an inaudible noise floor at my highest listening volume level.

Again this all isn’t exact science, and it’s not like this makes me feel any differently about how my gear performs, but imo it’s kinda fun and interesting to make a bit more sense of all this.