Null Test Your Master Against a Reference: What the Result Tells You
Published · by Loopin
A null test shows you exactly what a master changed: line up two files, flip polarity on one, match the level and listen to what is left. Silence means identical; whatever you hear is the difference.
Know what a null test can and cannot do
It compares the same music. Master against the unmastered mix tells you what the processing added; master A against master B from two services tells you how they differ. A master against a different released song (a true reference) will not null, because the notes differ and you hear only mush.
So for a different-song reference, use loudness and spectrum comparison instead. Use nulling when you want to hear the change itself, which is the point when you check an automatic master before sending it to a client.
Line the files up to the sample
Drop both files on separate tracks starting at 0:00. Zoom to the waveform and compare the first transient; if one file is offset, nudge until the first kick sits at exactly the same sample. Any offset, even one millisecond, gives you a comb-filtered mess instead of a null.
Use the same sample rate on both. If the rates differ, convert one first, otherwise the test measures the conversion, not the master.
Flip polarity and match the level
Invert polarity on one track (a utility or gain plugin has a phase-invert button) and sum both to the master. Now match the levels: turn the louder master down until its loudness equals the quieter one. If you skip this, the leftover is dominated by the gain difference.
A good starting point is loudness within 0.1 dB, read on the same meter. Loopin’s Original-vs-Mastered switch is loudness-matched for listening, but a null test needs the two exported files matched by hand. Use free online mastering, download the master and test it against your mix.
Read the residue: four results
Silence: the files are the same, or the processing changed only gain. A faint, tonal residue that follows the music: tone changes, such as a bass or treble shift. A rhythmic, pumping residue that sits on the kicks and snares: heavy limiting or compression; compare it to your mix to see where it clamped.
A constant hiss at very low level: format differences, such as 16-bit versus 24-bit, which is normal and sits around −90 dBFS. Residue at normal volume, loud and wide: the master changed the stereo image. Listen with headphones and note which element you hear.
Use the result to make a decision
If the residue is mostly low-end and the client hates a muddy master, go back to the mix and fix the kick and bass balance. If it is pumping, try a less dense feel or less loudness rather than shipping it. If you want a target to hold it against, see what a reference track is.
Log the result in the session notes with the date and the settings. Next month, you can compare the same sound without guessing what you did.
Do not null across formats and call it proof
MP3 against WAV never nulls to silence, because the lossy file differs. Compare the exact files you deliver. A null test is a diagnostic for engineers; it does not tell you which master sounds better. Your ears do, after level matching.
Automatic mastering can be checked this way but cannot be directed; if the null reveals something you dislike, take it to an engineer. Free downloads are capped at 1 per rolling 24 hours (2 with a free account), so run the null on the download you intend to deliver, not on a previous file. Master the track free first.
Frequently asked questions
What is a null test in mastering?
You invert polarity on one of two aligned files and sum them. What remains is the difference between them, so you hear exactly what changed.
Can I null a master against a commercial reference?
No, not meaningfully. A different song will not cancel. Use loudness and spectrum comparison for that, and keep nulling for the same music in two versions.
Why must I match levels first?
A gain difference of even 1 dB leaves a residue that drowns out the real differences. Match within 0.1 dB on one meter before you listen.
What does a pumping residue mean?
It follows kicks and snares and points to dense limiting or compression. Compare it with your mix to see how hard the master pushed.
Does 16-bit vs 24-bit show up in a null?
Yes, as a very low level hiss around -90 dBFS. That is normal and not a flaw in the master.
Why does my null leave a loud comb-filter sound?
The files are not sample-aligned or have different sample rates. Align the first transient exactly and convert to one rate.