Why Songs on Your Album Sound Different, and the Fix

You play your record from start to finish and something is off: one song is louder, another is darker, a third feels like it came from a different band. It usually came from a different day. Here is why it happens, how to find out exactly what differs, what mastering can fix and what has to go back to the mix.

The short answer

Songs on an album sound different because they were mixed at different times, in different rooms or at different listening levels, so they differ in loudness and tone. Mastering fixes the differences that sit across the whole song: level, a gentle tonal match toward the album's average, and the gaps and fades between songs. It cannot fix a problem inside a mix, such as a buried vocal or a muddy low end.

You hear	Usual cause	Fix it in
One song is louder or quieter than the rest	A different level or amount of limiting when it was bounced	The master (album level)
One song is brighter or duller	A different room, monitors or listening level when it was mixed	The master, gently
One song has much more bass	A different room, or a mix never checked against the others	The master if it is small, the mix if it is big
The vocal is lost in one song	A balance decision in that mix	The mix
One song sounds muddy or boxy	A problem in that mix	The mix
Too much or too little silence between songs	Gaps and fades never set	The master

Why it happens

Mixed months apart. Taste, plugins and ears change between sessions. A song finished in spring and one finished in autumn rarely share a balance.
Different rooms or engineers. Every room and every pair of monitors hides some frequencies and exaggerates others, so a mix that sounds right in one place is brighter or bassier in another.
Different listening levels. At low volume the ear hears less bass and top, so a song mixed quietly tends to get more of both than one mixed loud.
Different references. If each mix was chased toward a different favorite song, the mixes end up in different places.
One song brighter or bassier by choice or by accident: a different microphone, a different sample pack, a different plugin chain.
Different loudness. One song bounced with a limiter working hard, another with none; one with peaks at 0 dBFS, another at −6.

How to tell what is different

Do these in order. Each one isolates one kind of difference:

Match the loudness first. Louder always sounds better for a few seconds, so two songs at different levels tell you about level, not about tone. Turn the louder one down until they feel equal, or use a loudness meter.
Flip between the end of one song and the start of the next, at one volume setting. The jump you feel is the jump a listener feels.
Compare the tone curves. A spectrum analyzer with the same settings on each song shows where one has more low end or more top than the others. Look at the average over the whole song, not at one moment.
Measure the integrated LUFS and the true peak of each final file. Songs several LU apart will not sit together, whatever their tone.
Listen to the record in order on headphones and on a phone speaker. The odd song out on a phone is the one to look at first.

What mastering can fix

Mastering works on the finished stereo file, so it can move the whole song, not the parts inside it. For an album that means three things:

Level: the songs follow one level plan. The most intense song sets the level and the quieter, ballad-like songs sit a little under it, as on commercial records. Not every song at the same LUFS.
Tone: the songs that stand out are nudged toward the average of the album. A small EQ move on the outliers, not a copy of one song's curve onto all the others.
Flow: the gaps and fades are set, so the silence between songs feels the same all the way through.

How to set each of these, step by step, is in the guide on how to master an album.

What needs a remix

A vocal that is buried, or too loud, over the music. That is a balance decision in the mix.
A muddy or boxy mix. Cutting at the master is a patch, not a cure.
Harsh S sounds. De-ess the vocal in the mix; taming them at the master also dulls everything else up there.
A song that needs a big EQ move, much more than a couple of dB, to sit with the rest. Go back to its mix and compare it with a song on the record that works.
Clipping, noise and bad takes.

Mixing vs mastering: which problems belong to the mix and which to the master

Fix it without making every song the same

The goal is a record that sounds like one band in one room, not songs cloned from one curve. Fix the outliers and leave the rest. A ballad should stay softer than the closer. An interlude should sound like an interlude. A record where every difference has been flattened is tiring to hear.

Or do it in your browser

OBRA's Album mode does this for 2 to 20 songs, in any order you drag them into. The most intense song carries the album level and the others sit a little below it, ballads lower, capped so nothing drops too far (up to 4.5 LU in the Natural master, 3.5 in OBRA, 2.5 in Loud). Only songs that stand out from the album's average are nudged in tone, half the way, by at most ±2 dB with at most two EQ bells; interludes and skits are left alone. Gaps default to 2 seconds, adjusted for tails, cut endings and segues, and you can change the gap and the fade of every song.

You hear the record in Natural, OBRA and Loud side by side. The download is a ZIP with every master, a CUE sheet and an album report with the LUFS and peak of each song. Listening is free. The album download is $9 / €9 / £9 once, with the downloads available for 30 days, or included with OBRA PRO. Nothing is uploaded: the engine runs in your browser.

Master your album →

If you only need to check a single song, master it free: two songs a day.

Master a single song free →

Keep reading

How to master an album or EP: order, level, tone, gaps, peaks and delivery, step by step
How loud should you master for Spotify?: a target by genre and how to hit it
LUFS for Spotify, Apple Music and YouTube: what LUFS and true peak mean

Frequently asked questions

Why do the songs on my album sound so different?

Usually because they were mixed at different times, in different rooms or at different listening levels. That leaves one song louder, brighter or bassier than the rest. It is common and fixable when the differences are in level and tone.

Can mastering make all the songs on my album sound the same?

It can bring them to one level plan and nudge the songs that stand out toward the album's average. It does not make them identical, and it should not: a ballad stays softer than a closer, and an interlude stays an interlude.

Should every song on an album have the same loudness?

Usually not. The most intense song sets the level and the quieter songs sit a little lower, as on commercial records. Forcing every song to the same LUFS makes the quiet ones feel pushed.

What cannot be fixed by mastering?

Problems inside the mix: a vocal that is buried, a muddy or boxy mix, harsh S sounds, clipping and noise. Those go back to the mix. A song that needs a big EQ move to fit is a mix problem too.

How do I compare songs fairly?

Match their loudness first, because the louder one always sounds better for a few seconds. Then flip between the end of one song and the start of the next, and compare their tone curves with a spectrum analyzer.

How much does it cost to fix an album with OBRA?

Listening is free. The album download is $9 / €9 / £9 once, with the downloads available for 30 days, or included with OBRA PRO. Single songs are free for two a day at /mastering.

Is my music uploaded?

No. The engine runs in your browser, so your songs stay on your device.

Why the songs on your album sound different, and how mastering fixes it