Why your mix sounds crowded: what to turn down first

· Mixing demonstration with Songbird

Vocals can sound clear on their own and still get lost under the rest of the arrangement. Before raising their level, check the guitars, piano and backing vocals. Turning down a competing track may make the lyrics easier to hear.

In a sample of RateMySong (RMS) reviews from artist and producer accounts, 36% included advice to adjust the level of an existing vocal or instrument. Most also recommended other work on the song or mix.

This guide explains how to choose what to turn down, compare the result and recognize when a mix needs more than a level change. You'll hear two mix comparisons made with Songbird.

How we analyzed the reviews

Why a mix can sound crowded

A mix needs a clear focus. During a verse, that might mean keeping the lead vocals clear while the backing instruments provide rhythm and harmony. During a solo, the lead instrument takes over. Giving every track equal prominence can make it hard to hear what should stand out.

Start with the balance between the parts. The master volume changes the level of the whole recording; an individual track's fader changes its place within the mix. Turning the whole song down leaves those relationships intact.

EQ adjusts the balance of frequencies within a sound. It can help parts fit together, but a guitar that is simply too prominent may need a level change first. Engineer Chris Wainwright describes this approach in iZotope's guide to common EQ mistakes: establish levels and pan positions before processing, then judge changes with the other parts playing.

What to turn down first

Save your current mix, then loop 10–20 seconds of a verse or chorus where the vocals are hard to hear. You'll need separate vocal and instrument tracks to adjust their levels independently.

  1. Decide what should lead. If the vocals carry the verse, use them as your reference. Identify the backing parts you want to preserve, such as the guitar rhythm, bass line or chorus harmonies.
  2. Check the backing tracks. Listen for a track that makes the lead vocals harder to hear. Mute it briefly, then bring it back. Check both vocal clarity and what that track adds to the arrangement.
  3. Try a small reduction. Start by lowering that track by 1 dB while the whole mix plays. Listen to the same vocal line again. If the lyrics are easier to hear but the rhythm loses energy, adjust the level before changing another track.
  4. Compare both versions at a similar volume. Listen to the same section in each version. Check whether the vocals are clearer and whether the backing instruments still have enough presence. Keep the original if you prefer it.

Sometimes a backing track only needs to come down during a vocal line. Use volume automation to lower it for that line and bring it back between vocal phrases. Automation lets you change a track's level at specific points in the song. If the problem remains even at a lower level, investigate the sound or arrangement. Two busy parts may need different notes, timing or instruments to fit together.

Example: lowering the backing instruments and vocals

We used Roswell Pro Audio's recording of “The Weight,” written by Robbie Robertson. It has separate vocal and instrument tracks, so we could change the balance while keeping every performance and the full song intact. Songbird handled the mixing and scoring. The musicians' performances were already recorded.

For Mix A, we deliberately brought the acoustic guitar, electric guitar, piano and two backing-vocal tracks forward. For Mix B, we returned those five tracks to lower levels. These are our demonstration mixes, not Roswell's released mix.

Songbird's first listening assessment still described clear lead vocals in Mix A, with a prominent acoustic guitar. It suggested a smaller, optional guitar reduction. We had planned the broader level comparison ourselves; it does not demonstrate RMS diagnosing a crowded mix.

Here is the complete change from A to B. The lead vocals, drums, bass and other tracks stayed at the same settings.

Part Level change from A to B
Acoustic guitar Down 4 dB
Electric guitar Down 6 dB
Piano Down 6 dB
Two backing-vocal tracks Each down 4 dB
Lead vocals Unchanged

Those amounts describe this recording; they are not recommended settings for another session. We changed several supporting parts together, so this comparison cannot tell us which individual move mattered most. No EQ or other processing was added.

Listen to the same 25 seconds from each mix. Focus on the lead vocals, then the acoustic guitar and backing vocals. Check whether the lyrics are easier to hear in B and whether the backing tracks have become too quiet. Both versions are matched for playback loudness.

Same section in both mixes0:40–1:05 · 25 seconds

Compare the backing track levels

Mix A · Backing tracks raised

Matched playback loudness. Switching keeps your place.

Recording ©2020 Roswell Pro Audio · Demonstration mix made with Songbird

Hear the complete A and B mixes · 5:39 each
Mix A · Backing tracks raised

Recording ©2020 Roswell Pro Audio · Demonstration mix made with Songbird

Mix B · Backing tracks lowered

Recording ©2020 Roswell Pro Audio · Demonstration mix made with Songbird

What changed in the scores

We scored the complete recordings with Songbird's scoring tool, using the same Rock comparison group and scoring settings throughout.

Version Production Quality Overall
A · Backing tracks raised 54 40
B · Backing tracks lowered 58 40

Production Quality rose by four points. The Overall score stayed at 40. That is a narrower result than saying the song became better: Production Quality is a model's assessment, not a direct measure of how clearly you can hear the words.

The automated listening comparison did not establish a preferred mix. The score also assessed the original exports, whose overall loudness changed with the faders; we matched only the listening copies. We therefore cannot attribute the score increase solely to clearer separation. Your comparison of the recordings remains part of the decision.

Would lowering the lead vocal level help further?

We made one more version. After listening to B, Songbird tried lowering Bret Alexander's lead vocal by 1 dB to bring it a little further back in the mix. Every other track stayed at the same level.

This Mix C scored 55 for Production Quality and 39 Overall, down from B's 58 and 40. The additional adjustment did not improve either score, so we restored B as our working version. That result applies to this particular vocal change; it does not mean vocals should always stay louder.

You can hear the B/C comparison here, again at matched playback loudness. Listen to the lead vocals against the backing instruments and decide which balance you prefer.

Same section in both mixes0:40–1:05 · 25 seconds

Compare the lead vocal level

Mix B · Original lead vocal level

Matched playback loudness. Switching keeps your place.

Recording ©2020 Roswell Pro Audio · Demonstration mix made with Songbird

Hear the complete B and C mixes · 5:39 each
Mix B · Original lead vocal level

Recording ©2020 Roswell Pro Audio · Demonstration mix made with Songbird

Mix C · Bret's vocal lowered 1 dB

Recording ©2020 Roswell Pro Audio · Demonstration mix made with Songbird

Get feedback on your mix

If you are unsure which track to adjust, upload your current mix to RateMySong, choose Mixing with Deep and describe where the balance sounds wrong, including a timestamp. Explain how you want the vocals to sit in the mix and whether you can edit the separate tracks. For example:

In the chorus at 0:40, I want the lead vocals to stay clear without losing the guitar rhythm or backing harmonies. I have the separate vocal and instrument tracks. Is anything competing with the lead vocals, and what should I check first?

Use the feedback to choose a change you can hear and compare. The aim is a balance that serves the song: a clear focus, with enough of the other parts left to support it.

You'll need to sign in and have review credits. RMS shows the cost before you run the review. For more help describing your intention, see how to get useful feedback on your music.

Review my mix

Review sample, recording and comparison notes

Review sample. We randomly selected 200 recordings from 6,659 completed original submissions made August 23–September 21, 2026, using Pacific dates. The accounts identified as artists, singer-songwriters, rappers, bands or producers. Internal/test accounts, declared revisions and exact audio reuploads within the period were excluded. Sixty selected reports covered songwriting rather than mixing. Of the remaining 140 Full Review or Mixing reports, one disclosed AI-generated accompaniment and was excluded, leaving 139 reports from 131 accounts. Account profiles do not verify that a recording is non-AI.

What we counted. AI-assisted review of the saved feedback identified 50 of 139 reports (36%) recommending an explicit relative-level change: a fader adjustment, clip gain, volume automation or volume ducking. Fifteen (11%) included one in the first suggested fix; 35 included one later or in supplementary coaching. EQ or compression alone, instructions for newly added parts, section dropouts and changes to the whole recording's volume did not count.

Two ambiguous cases stayed in the denominator but were not counted as positives. A 95% sampling interval for the 36% estimate is approximately 28–44%; it does not account for errors in the feedback or our classification. We checked the coding against source passages and rechecked a random set of 20 reports using the same AI-assisted process. This measures written recommendations. We did not independently listen to these submissions or test whether level changes alone would resolve their problems.

Recording and credits. “The Weight,” written by Robbie Robertson. Recording ©2020 Roswell Pro Audio LLC. Vocals: Bret Alexander, Stacia Abernatha and Paul Smith. Guitar, mandolin and keys: Bret Alexander. Bass and acoustic guitar: Paul Smith. Drums: AJ Jump. Violin: Nyke Van Wyk. Recording engineering: Paul Smith / 8DAWS. Recording source and original mix. These demonstration mixes were made with Songbird from the supplied session; the original performances were not generated by Songbird.

How we made the comparison. All three mixes contain the same 16 source tracks and last 5:39. One duplicate bass file was excluded. Timing, pan positions and the master setting stayed fixed. The A/B changes were five track-level reductions; C changes only Bret's lead level from B. Sample checks verified each rendered recipe. No additional EQ, compression, limiting or automation was used. All complete versions were scored with the same model, Rock comparison group and excerpt positions. Playback copies were separately loudness matched using constant gain and encoded as MP3s; scored WAV files were left unchanged. The short comparisons cover 0:40–1:05.

Selection and limits. We retained this level comparison after two earlier, smaller guitar and vocal tests produced unchanged scores in a separate RMS web-review series. Those results are not comparable with the Songbird scores shown here. The revised balance was planned by us; the first listening assessment did not independently prescribe it. The automated A/B assessment found no preference, and no listener study has established one. These results do not estimate how much another song's score would change.