Client Feedback on Audio Mix Passes: Getting Notes a Mixer Can Act On

"The music feels a bit loud, can the voice come more in front, and the ending needs more energy." That is the message, almost word for word, that lands in my agency's WhatsApp under the second mix pass of nearly every branded video we deliver. I'm pretty sure you've received the exact same sentence if you mix for clients, right, because it's basically the default shape of client feedback on audio mix passes when nobody has given the client a better way to say what they hear.

Now unpack it with me for a second. I don't know which part of the music they mean, the whole bed or just the swell around the product reveal, and I don't know if loud means loud against the voice or loud because they played it on a laptop at full volume. And "more energy" at the ending could mean push the music up, cut to a different cue, add a hit on the logo, or something in the edit that has nothing to do with the mix.

So the mixer guesses, and when the guess misses you're into round three on a job quoted for two. To be very honest, that loop comes from the format the notes arrive in, which we control as the people running the job, right, so this post is the system I use in my own agency to fix it.

Why client feedback on audio mix passes is so often vague

Clients hear a mix as one single thing, right, the way you hear a song in a café, so when something bothers them they describe the feeling rather than the layer causing it. That's completely normal, and it's exactly why sound design has a whole vocabulary that most marketing managers never needed to learn.

How they listen makes it worse. Most clients play the mix once on laptop speakers or a phone and type a summary from memory later, so the exact moment that bugged them melts into "the music feels loud" instead of "the music swells over the pricing line at 0:42".

And then mix notes and edit notes arrive tangled together, which is honestly the one that costs me the most rounds. "More energy" at the ending might really mean a different music cue or a faster cut into the logo, and those decisions belong earlier in the post-production pipeline. The catch here is that a mixer solving an edit problem with faders will push levels around for two rounds without making the client happy, because the problem was never a level.

Music videos are worse, right, with the artist, the manager and sometimes the label all listening on different setups, which is why I set those jobs up with PlayPause for music video production, and narrative films get their own sound mix approval workflow. Think of this post as the everyday version for branded, corporate and music video work.

“

A vague mix note is usually a very precise feeling that arrived without a timestamp attached to it.

A simple vocabulary for mix feedback notes

Before the first mix pass goes out, I send the client a really really short message on how to talk about sound, because anything longer gets skipped by a busy brand manager. Basically I ask them to name the layer they're reacting to, using one of these five words:

• Voice: the person on camera or the voiceover, including breaths, clicks and clarity.
• Music: the track underneath, its level, where it swells, cuts and ends.
• Effects: whooshes, hits, risers and any sound synced to something on screen.
• Room: the background ambience, crowd noise, street noise and so on.
• Overall: the loudness of the whole video against the videos it will sit next to.

Then I ask for a direction, up or down, earlier or later, cleaner or fuller, and a moment, which on a review link is just the frame the comment lands on. I also ask what they listened on, because notes from phone speakers and notes from studio headphones are very very different, and knowing that saves a lot of arguing about low end the client literally cannot hear.

A worked example from one vague message

Let me run the WhatsApp message from the top through that vocabulary, using a made-up 90-second product video so the timings are only illustrative, right. "The music feels a bit loud" usually means one stretch, so it becomes a range comment from 0:18 to 0:55 saying "Music down a little under the interview answer", and the intro keeps its level.

"Can the voice come more in front" becomes a frame comment at 0:42 saying "Voice, the word pricing gets swallowed", which is one word to fix rather than a whole voiceover to push up. And "the ending needs more energy" splits in two. One is an effects note on the logo frame at 1:24 saying "Hit on the logo feels small and late", and the other is an edit note asking to cut to the logo faster, which goes to the editor and never lands on the mixer's list.

So one vague sentence becomes four notes, three for the mixer and one for the editor, and the next pass answers all of them instead of guessing at one blended feeling.

On PlayPause I keep this vocabulary in a brand playbook for each client, a per-client document of creative style and direction you can share as a no-login page, so nobody digs through old emails to remember what "fuller" meant last time. Playbooks are on the Agency plan and up, and if you mix your own edits, my post on audio mixing basics for video editors covers the levels side before the client hears anything.

• Name the layer you are reacting to
• Say which direction you want it to move
• Leave the note on the exact moment
• Mention what you listened on
• Keep edit requests in a separate comment

Range comments for notes that span a whole section

Not every mix note is about one frame, right. Plenty cover a whole stretch, for instance "the music is fighting the interview for the entire testimonial section", and if the client drops that as one timestamp at the start, the mixer has to guess where the problem ends.

A range comment fixes that, because the client marks an in point and an out point, so the mixer sees the problem runs from 0:18 to 0:55 and dips the music across exactly those 37 seconds, instead of pulling the whole track down and wondering why the intro feels empty. Range comments are on every PlayPause plan, and you can see how they behave for sound on our audio review and time-coded annotation page.

The catch here is that some clients love ranges a bit too much, and you'll get one across a whole three-minute video saying "audio feels off", which is basically the WhatsApp message again with extra clicks. So I ask for one section per range, a verse, an interview answer or a product demo, and if the feeling really covers everything, it's an overall loudness note.

Review_Cut_v4.mp4In Review

212160p · ProRes

00:34 / 02:18

SR

Sarah 0:34

Frame-accurate note, everyone sees the exact same thing.

Timestamped audio feedback on words, hits and transitions

The other half of mix notes are tiny and precise, right, like a harsh "s" on one word, a breath before a key line, a whoosh landing two frames after the cut, or a music edit that jumps at a transition. A hit that lands a couple of frames late tends to feel wrong even when nobody can say why, and a note saying "around the one minute mark" sends the mixer scrubbing around for ages.

With time-coded annotations for video review, the client pauses where they hear it and types, and the comment sticks to that exact frame, so "the hit on the logo is late" arrives pinned to the logo and the mixer nudges it by eye against the picture. I went through why that beats typing timecodes by hand in my explainer on frame-accurate commenting in video review. For speech notes, AI transcription on the Agency plan gives the client a clickable transcript beside the player, so they find the exact line where the voice dips without scrubbing at all.

Where the notes go next depends on where the mix happens. Our Premiere Pro panel brings the comments into Premiere and jumps the playhead to the frame when you click one, and I covered the timecode side in syncing timestamped comments to Premiere. If your mixer works in Resolve or in a DAW, the flow is export, upload, review, and then they work through the notes by timestamp in their own session, which is why keeping your Resolve timelines and client versions organised matters so much.

The review file needs decent audio too, so I export a proper stereo track instead of a tiny proxy, as I explain in my notes on the best codec for client review.

Name the layer, the direction and the moment

A note like music down a little under the voice at forty two seconds gets fixed in one pass, while the music feels loud usually costs a whole extra round.

Routing mix notes to the mixer without email chains

In my agency the editor and the mixer are often two different people, and sometimes the mixer is a freelancer who has never spoken to the client. Rewriting the client's notes into an email for them is the worst thing I can do there, because every rewrite loses a bit of what the client meant and adds a bit of what I think they meant.

So the notes go to the mixer directly, right. The client comments in the browser with no account and no install, I @mention the mixer wherever they're needed, and when the mixer isn't sure whether "cleaner" means less reverb or less music, they reply in the thread on that comment, so the answer lives on the frame instead of in a WhatsApp side chat, you see what I mean here.

Slack posts each comment to a channel on every plan, and that's how my mixers actually see notes, since a freelancer lives in Slack and not on a review page. Guest version upload, on Agency and up, lets the mixer put the next pass onto the card without an account, with their name and email collected. And the who-watched analytics, which every plan has, show who opened the link, when and from which city, which is really useful when a client replies "sounds good" and nobody actually played the mix.

I also use custom statuses like "Mix notes open", "With mixer" and "Mix approved", so everyone sees where the pass stands without asking. If a note turns into a request for a new cue or a recut ending, I flag it in the thread as a change of scope, right, because that's exactly how scope creep sneaks into mix rounds.

1Client comments on the link with layer, direction and moment

2Editor tags the mixer on the notes that need them

3Mixer replies in the thread when a note is unclear

4Slack posts every new comment to the mix channel

5Mixer uploads the next pass onto the same card

Comparing passes in a sound mix review before sign off

Every mix pass goes onto the same card as a new version, MV1, MV2, MV3 and so on, with the old notes staying attached to the old version, so when a client says "it was better before" I open MV1 and listen instead of hunting for a file called final_final_v2. To me that's the simplest form of version control a client understands without a lesson.

Before sign off I check the mix revision notes on the previous version one by one against the new pass, because a mixer fixing nine of ten notes and missing the tenth is how you end up in round four. On the Agency plan, side-by-side version compare puts the temp mix next to the final, or MV2 next to MV3, so I can confirm the dip under the interview actually happened and that fixing the voice didn't thin the music out somewhere else, right.

One honest point on listening, the browser plays a streaming copy while your original stays untouched, so I treat the review link as the place for balance, timing and clarity notes, and at the end of the day the final loudness check happens on proper monitors in the mixer's room.

Old way

notes forwarded by email, passes saved as loose files and nobody sure which mix the client heard

With PlayPause

every pass stacked on one card, notes pinned to frames, and temp and final played side by side

Frequently asked questions

How do I get a client to give specific notes on an audio mix?

Give them a tiny vocabulary before the first pass goes out, right, the layer, the direction and the moment, and ask them to leave each note on the exact frame or as a range across one section. In my agency that one short message does more than any amount of chasing afterwards, because the client learns what a useful note looks like before writing a vague one.

Should clients review a mix on headphones or speakers?

Ideally both, but mostly on whatever their audience will use, which for branded and social work is usually a phone or a laptop. I just ask them to tell me what they listened on, so I can weigh a bass note from phone speakers differently from one given on good headphones, and the mixer's own monitors stay the final technical reference.

Can a freelance mixer upload a new mix pass without an account?

On the Agency plan and up, guest version upload lets someone add the next version to the card without an account, and PlayPause collects their name and email so you know who uploaded it. The new pass stacks on top of the previous one, and the old notes stay on the old version where they belong.

Which PlayPause plan do I need for mix review?

Range comments, frame-accurate comments, threaded replies, @mentions, Slack and version stacks are on every plan, starting with Creator at $9 a month. For real mix rounds I'd pick Agency at $19 a month, because side-by-side compare, AI transcription, guest version upload and Playbooks start there, and its share links last 90 days instead of 30. Pricing is flat per workspace, so adding your mixer as a member doesn't change the bill.

If your mix rounds keep stretching past two because the notes arrive as one vague paragraph, set up the vocabulary and send the next pass as a link, and trust me on any level, your mixer will feel the difference in the first round. You can see every plan on our PlayPause pricing page, and each one starts with a 7-day free trial.

So yeah. That's my way of saying it.