How to give feedback to an audio engineer (that actually results in what you want)
23/05/2026
The revision process is where a lot of projects either go smoothly or start to go sideways. Not because of a lack of skill on either side, but because the communication between a musician and an audio engineer often defaults to a language that does not translate into specific, actionable changes.
“Make it sound better” is the most common example. It is honest — the person giving the feedback genuinely wants it to sound better — but it tells the engineer nothing about what “better” means to them. What follows is usually a guess on the engineer’s part, another version that may or may not be closer to what the musician wanted, and a revision round that could have been avoided.
This is not a criticism of musicians who give vague feedback. It is a skill that no one explicitly teaches. Here is a practical guide to doing it well.
The two kinds of feedback
Feedback on an audio edit or mix generally falls into two categories: descriptive and directive.
Descriptive feedback describes what you are hearing — what the current version sounds like to you. “The guitars feel too dominant in the chorus” or “the vocal sounds far away” or “the whole thing feels a bit muddy in the low end” are all descriptive.
Directive feedback tells the engineer specifically what to change. “Turn the guitars down in the chorus” or “bring the vocal forward” or “reduce the low end around the 200–400Hz range” are directive.
Both are useful. Descriptive feedback is useful because it tells the engineer what you are hearing, which may or may not match what they were hearing. Directive feedback is useful when you know specifically what you want. Descriptive feedback is usually better when you are not sure exactly what the fix is — that is the engineer’s job to work out. Directive feedback is better when you have a clear preference.
What is neither useful nor directive is a global instruction: “it needs more energy” or “it should hit harder.” These say nothing about what specifically needs to change.
The reference track
The most efficient tool for communicating a desired sound is a reference track — a song that represents something close to what you are aiming for. It does not need to be the same genre or the same instrumentation. It just needs to represent some quality that you want your track to have.
“Something like the vocal clarity in this song” or “the low end weight of this mix” or “the overall brightness of this production” — these are specific, communicable targets. An engineer can listen to a reference and understand what quality you are pointing at, even if the music itself is very different.
If you do not have a specific reference track, describing the characteristic you want — bright and airy, warm and dense, punchy and controlled — is still more useful than a general “it should sound better.”
Describing what you hear vs describing what you want
One of the more common feedback problems is a confusion between describing the symptom and describing the desired outcome.
“The snare sounds weak” is describing what you hear. “Make the snare more punchy” is describing what you want. Both are useful, and giving both is even better: “the snare sounds weak, I’d like it to feel more punchy and present.”
Some examples of effective feedback combinations:
“The vocals are getting lost behind the guitars in the chorus — I’d like them to sit more in front.”
“The whole mix sounds a bit dark and dull — I want it to feel brighter and more open, something like the top end on [reference track].”
“The bass feels inconsistent — some notes jump out while others disappear. I’d like it to feel more even.”
Each of these describes what the person is hearing and what they want to hear instead. That is enough information for an engineer to work with.
What to avoid
A few patterns that slow revision rounds down and usually lead to less satisfying results:
Vague global instructions. “It needs more life” or “it should hit harder” or “it’s missing something” — these do not point to anything specific and require the engineer to guess what you mean.
Changing what you want between rounds. If the first version was reviewed and feedback was given, then the revision is reviewed and the feedback contradicts what was asked for in the first round, the process starts to spiral. Before giving feedback, listen carefully and give considered feedback rather than immediate reactions.
Too many things at once. A list of fifteen changes is harder to work through cleanly than five specific, prioritised notes. If there are many things you want to address, start with the most important ones.
Describing a feeling without any specifics. “It doesn’t feel right” or “I’m not sure but something is off” — these are honest reactions, but they do not help. If you can identify that something feels wrong but cannot pinpoint what, listening again on a different system or coming back to it after a break often helps clarify.
A simple structure for revision notes
If you are not sure how to structure your feedback, this format works well:
- What you are hearing (the symptom or characteristic)
- What you would like instead (the desired result)
- Where in the track this applies (throughout, just in the chorus, the first verse only)
For example: “The kick drum sounds a bit thin (1). I’d like it to have more weight and impact (2). Particularly in the verses where it feels like it disappears behind the bass (3).”
Three pieces of information. Specific, actionable, and clear about where the issue is. That is the kind of feedback that leads to a good result in fewer rounds.
A note on trust
The revision process works best when there is a degree of trust between the musician and the engineer. You bring the knowledge of the music, the intent, and the feeling you are trying to achieve. The engineer brings the technical tools and the trained ear to get there. The feedback process is the bridge between those two things.
Precise feedback is not about micromanaging. It is about communicating clearly so that the result reflects what you actually wanted — rather than what the engineer guessed you might have wanted. That benefits everyone.