Record/Playback Meter Reference Level?

What is the 0 dB reference level in the Playback and Record meters? Is is +4 dBu?

On a related matter, when measuring RMS of a portion of a recorded signal (Analyze => RMS), it displays “dB” and not in dBu or dBV. It seems the Audacity RMS measurement seems to be again a relative measurement (hence “dB”). Relative to what?

Sorry for such a simple questions (I am a newbie), but I was not able to locate an answer elsewhere

Audacity v2.4.1.



“None of the above.” :wink: It’s a digital level (just a number or count) and it’s correlated with a signal/voltage level but different ADCs (analog-to-digital converters) and DACs (digital-to-analog converters) are calibrated differently. (In a pro studio the digital levels might be calibrated with a dBu level and maybe with an SPL level.)

The digital reference is 0dBFS (zero decibels full scale) and it’s defined by the maximum count with a given number of bits. For example, with 16-bits you can “count” to −32,768 or +32,767.* You can’t count any higher without more bits. If your digital audio data hits those peaks you’re are at 0dB. (Everything is automatically scaled so a 0dB 24-bit file is no louder than a 0dB 8-bit file.)

0dB is the “digital maximum”. ADCs & DACs [u]clip[/u] (distort) if you try to go over 0dB. Regular (integer) WAV files and audio CDs are also hard-limited to 0dB.

Audacity uses floating-point** internally so it’s impossible to clip Audacity itself. For example, you can boost the bass and push the peaks over 0dB and as long as you adjust the overall volume down before exporting, your file won’t be clipped. (But Audacity will show potential clipping if you go over 0dB.)

[u]RMS[/u] is a kind of average on the same dBFS scale. The regular mathematical average is (about) zero since the signal is positive half the time and negative half the time. RMS corresponds to “loudness” better than the peak levels. (And there are other algorithms for measuring loudness.)


  • One bit is used for the +/- sign and there is no negative zero so you get one extra count on the negative side.

** With floating-point representation 0dBFS is defined as 1.0 and again everything is scaled when the file is converted to integer.

Thanks Doug.