Transcribers: control your playback app's volume against everything else on Mac

When you transcribe on a Mac, the audio you're typing needs to sit at one steady, audible level while music, Slack pings, and system alerts stay well below it. macOS won't do that on its own — it routes every app through a single master volume, so turning down a ping turns down the voice too. The fix is a per-app mixer: pin your playback app (Express Scribe, oTranscribe, VLC, a browser) to a fixed level, cap everything else, and nudge the source up or down from the keyboard without breaking your typing rhythm.

Transcription is the rare job where audio balance directly costs you time. Miss a mumbled word because a notification chimed over it and you're scrubbing back five seconds to catch it again. Set the source too hot and every consonant clips. And unlike most listening, your hands are pinned to the keyboard — often a foot pedal is driving play, pause, and rewind — so anything that makes you reach for the mouse to adjust a slider is a small tax you pay hundreds of times a shift.

Why the Mac fights you here

Most transcription playback tools have their own volume control. Express Scribe has a level slider and variable-speed playback with a constant-pitch mode; oTranscribe, VLC, and browser-based players all carry a volume slider too. So setting the source level is usually solved inside the app.

The trouble is everything around it. Your reference music, a second video you're checking against, Slack and Mail alerts, calendar chimes — they all share the one master output with your playback app. macOS gives you a single system volume and no per-app mixer, so there's no native way to say "keep the dictation at a comfortable 70% and hold notifications at 15%." Drop the master to quiet a ping and the voice you're transcribing drops with it. Push it back up and the next alert startles you. This is the same wall anyone hits who wants independent control of one app's volume on a Mac.

The free ways, and where they stop

You can get partway with no extra software. Each of these helps, and for the occasional file they may be all you need — but each also runs out.

• Set the source in the playback app. Fix the dictation to a level you can type to and leave it. This handles the source, not the interruptions around it.
• Turn on Do Not Disturb / a Focus mode. This silences most notification banners and their sounds, which helps a lot. But it's all-or-nothing per category, it doesn't touch background music or a second media app's level, and some alert sounds still slip through.
• Quit the noisy apps. Effective and free, but you lose your reference music and you're relaunching things every time you switch tasks.

None of these give you what the work actually wants: the source app steady and audible, the distractions present but quiet, and a way to shift the balance in a keystroke without stopping the pedal-and-type loop.

The real fix: a per-app volume mixer

A per-app mixer sits at the Mac's audio layer and gives every running app its own independent level in one place — the feature macOS never shipped. For a transcriber that means: playback app at your fixed working level, background music down low, notification sounds capped so a ping is a soft tap instead of a stab over a key phrase. Set it once and it holds across the whole file.

Several tools do this, and they're honestly close on the core ability. Where they differ is how you reach the control — which matters more than usual when your hands never leave the keyboard.

Background Music is genuinely free and open source, and it will hold your playback app and your music at separate levels from its menu-bar list — a solid choice if budget is the whole decision. SoundSource, from Rogue Amoeba, is the deep paid option at $49 one-time: per-app volume plus EQ and routing, controlled from a menu-bar panel, and worth it if you also want to shape the sound. Both are good tools. The question for a transcriber isn't whether they can set a level — it's how many actions it takes to change one mid-file.

Why the keyboard matters for this job specifically

A menu-bar mixer means: let go of the keyboard, go to the menu bar, click the icon, find the app's row, drag its slider, dismiss the panel, get your hands back. That's fine once. Across a shift of "that speaker got quiet, nudge it up — now the music's too loud, drop it — now a ping" it's constant small interruptions to a flow that lives or dies on rhythm.

Fader's premise is that the fastest controller you already own is the keyboard under your hands. You press a toggle hotkey and the keyboard becomes the mixer — one letter column is your playback app, the next is your music, and so on — so you ride levels with keystrokes and never leave the document. Set a per-app mute or a cap on notifications once and it stays. For work where your hands are already on the keys and a foot pedal owns transport, that's the difference between a mixer you fight and one you forget is even there.

A setup that works for a transcription shift

1. Set the dictation level inside your playback app once, to a volume you can type to comfortably for an hour.
2. In your per-app mixer, give the playback app its own channel and leave it at 100% — it's already set internally, so this becomes your stable reference.
3. Put background music and any second media app on their own channels, low. You want them present, not competing.
4. Cap notification and system sounds hard — or bind a mute key — so a Slack ping can't land over a key phrase. If your alerts are the main offender, see how to quiet Slack notifications without muting your music.
5. Keep the pedal for transport, and use the keyboard to nudge the source up or down when a speaker drops their voice — without stopping to mouse over to a slider.

Bottom line

The source app usually sets its own level; the real problem is holding it steady against music, alerts, and a second media app that all share one master output — and macOS gives you no way to split them. A per-app mixer fixes that. Background Music does it for free if you're patient with a menu, and SoundSource does it with more depth for $49. But for transcription, where your hands stay on the keyboard and your feet stay on the pedal, the thing that pays off every minute is reaching the control without stopping — which is exactly what Fader is built to do.