To compare two audio files, play them in sync from the same start point, match their integrated loudness, and switch between them so you hear the same bar of music from each. Listen to short sections, switch often, and write down what changes. If you need to know whether a difference is audible, such as WAV against MP3, hide which file is which and repeat the choice several times. Opening two files one after the other in a media player gives you two different moments of the song at two different levels, which tells you very little.
What people compare
The method is the same for every pair. What changes is the question you ask.
| Pair | Question |
|---|---|
| Two bounces of a mix | Did the vocal 1 dB up, the new reverb or the other bus compressor work? |
| WAV against MP3 or AAC | Can I hear what the lossy encoder removed? |
| Two versions of a song | Which arrangement, take or edit is better? |
| A mix against a master | What did mastering change, apart from loudness? |
| Your mix against a commercial track | How does my balance compare with a release I trust? |
The last two have their own guides: how to compare masters and how to use reference tracks.
Line up the files
Switching only works when both files start at the same point. Export every bounce from the same start position in the DAW, with the same length. If one file has half a second of extra silence at the start, every switch jumps half a second in the song, and you compare different notes.
Keep the sample rate and format the same for bounces you compare, unless the format is the thing you are testing. Name the files by what changed, for example "song vocal +1.wav" and "song vocal +0.wav", so the notes you take later make sense.
Match the loudness first
The louder file usually sounds better. Ears are less sensitive to lows and highs at low levels (ISO 226, the Fletcher-Munson curves), so the louder version sounds fuller and brighter. A difference of about 1 dB is enough to bias the choice. A new bus compressor that adds 1.5 dB of level will win against the old one even if it sounds worse at the same level.
Match the files by integrated loudness in LUFS, not by peak level. Peaks say how high the loudest sample reaches. LUFS measures the loudness of the whole file, K-weighted to approximate how ears hear. Measure each file, then turn the louder one down by the difference. A file at −12 LUFS against one at −14 LUFS needs 2 dB of gain reduction. Turn the louder file down rather than the quieter one up, so nothing clips. The level matching guide covers this in more detail.
Switch at the same moment
Small differences in sound are hard to remember. Switching instantly, at the same point in the song, lets you compare the same note from both files while the first one is still fresh. Stopping file A, opening file B and pressing play from the start puts 30 seconds of music and a gap between the two things you want to compare.
- Pick a section that shows the difference: the chorus for a vocal change, a sparse part for reverb, a kick and bass passage for low end.
- Loop that section, 4 to 8 bars long.
- Switch every bar or two. Listen to one element at a time: first the vocal, then the low end, then the cymbals.
- Note which file you prefer for each element, and why.
ABSeq does these steps on iPhone and iPad. It plays up to four files in sync and looping, the A/B key switches to the next track at the same moment in the song, and Auto Gain measures each file's integrated loudness and turns every track down to the quietest one. It opens WAV, AIFF, AIFC, CAF, FLAC, MP3, M4A and AAC from the Files app. The ABSeq walkthrough shows each step.
WAV vs MP3 and AAC
A WAV file holds the audio as it was bounced. MP3 and AAC are lossy formats. The encoder removes content it predicts you won't hear, so the file is smaller. Whether you can hear the loss depends on the music, the encoder settings, your playback system and your ears. Only a test answers it for you.
Watch the peaks when you encode. True peak is the highest level of the waveform between samples. A master at 0 dBFS sample peak can have inter-sample peaks above 0 dBTP, which can clip when encoded to AAC, MP3 or Ogg. Spotify recommends masters peak no higher than −1 dBTP, and −2 dBTP for masters louder than −14 LUFS. If an encoded file sounds harsh or crackly on loud hits and the WAV doesn't, check the WAV's true peak first. The true peak guide explains why.
Blind tests show whether you can hear it
When you know which file is the WAV, you tend to hear it as better. A blind test removes that. Someone else, or software, plays the two files in random order without saying which is which, and you pick one each round.
There are two kinds of test. An ABX test asks you to identify which file X is. A preference test asks which of two versions you prefer. In both, chance matters. In 16 ABX trials, 12 or more correct has under a 5% chance by guessing (p ≈ 0.038). In 10 trials, 9 or more correct has about a 1% chance (p ≈ 0.011). The same reasoning works for preference. If you pick the same file 9 times in 10, chance is an unlikely explanation. If your picks split 5 and 5, you can't hear a difference you care about, and either file will do.
ABSeq's Blind Test is a preference test. It plays the same short passage from tracks A and B each round, shuffled and labelled Version A and Version B, with Auto Gain on by default. You set 1 to 50 rounds and the seconds per passage, and the file names appear only in the results. The blind A/B test guide covers how many rounds to run and how to read the score.
Questions
How do I compare two audio files?
Export both from the same start point, match their integrated loudness in LUFS, then play them in sync and switch at the same moment in the song. Loop a short section, switch every bar or two, and note which file you prefer for each element.
How do I compare the audio quality of two files?
Match their loudness first, because the louder file usually sounds better. Then switch between them on the same section, and run a blind test so you don't know which is which. If you pick the same file 9 times in 10, the difference is probably real.
Can you hear the difference between WAV and MP3?
It depends on the music, the encoder settings, your playback system and your ears. Test it blind at matched loudness. In 16 ABX trials, 12 or more correct answers has under a 5% chance of coming from guessing.
Why does one file sound better just because it is louder?
Ears are less sensitive to lows and highs at low levels, so a louder version sounds fuller and brighter. A difference of about 1 dB can bias a comparison. Match integrated LUFS before you judge.
How do I compare two audio tracks of different lengths?
Trim or re-export them so they start at the same point in the song, then loop the section you want to judge. If the start points differ, switching jumps to a different moment and you compare different notes.