Why Audacity 4.0 Broke My Android Sound-Asset Pipeline (And What I Switched To)

I update a handful of notification and UI sounds for an Android app every couple of releases — think button taps, a success chime, a couple of voice-over snippets for onboarding. Nothing fancy, but there are usually 30-40 raw WAV files that need trimming, normalizing, and converting to Ogg Vorbis before they land in res/raw. I've done this with Audacity for years, and when I saw the 4.0 release notes on GitHub I upgraded the same afternoon, mostly excited about the Qt rewrite and the new clip-editing model.

The interface really is nicer. But about twenty minutes into my usual workflow I went looking for the Macro Manager to batch-export the whole folder, and it just wasn't there. Neither was the scripting pipe I'd been driving from a small Python script for the last two years. That's the part of this release nobody puts in the headline screenshot, and it's the part that actually changed how I work.

This post walks through what Audacity 4.0 actually changes according to the release notes, what genuinely got better for hands-on clip editing, and where I had to build a workaround because the automation layer I depended on for Android asset prep simply isn't in this release yet.

What Audacity 4.0 Actually Rebuilds Under the Hood

The headline change, straight from the release notes, is that Audacity 4 rebuilds the entire application interface on Qt, with native high-DPI rendering. That is not a coat of paint — toolbars and panels can now be moved, docked, floated, shown or hidden, and the old fixed-mode toolbox is gone in favor of a layout you actually control.

Alongside the UI rewrite, Audacity switched project storage to a new .aup4 format. The old Select, Envelope, Draw and Multi-tool modes have been removed too, replaced by context-sensitive behavior — volume envelopes now live in a Clip Gain mode, sample drawing kicks in automatically when you zoom to individual samples, and splitting is triggered by holding S instead of switching tools. Sync-Lock is also gone, replaced by explicit gap-preserving and timing-preserving variants of delete, cut and paste.

None of this is subtle. If you have muscle memory built up over a decade of Audacity 2.x and 3.x, expect to relearn a chunk of it, even though the underlying editing concepts — tracks, clips, waveforms — are still recognizably the same program.

The New Clip-Editing Model Actually Helps When You're Cleaning Up a Folder of App Sounds

This is the part of 4.0 I have no complaints about. Every Android sound-asset pass involves the same tedious task: open a batch of short clips, trim silence off the front, normalize levels so the success chime isn't twice as loud as the tap sound, and nudge a few boundaries so nothing clips mid-word in the voice-over lines. The old Audacity made you do this one selection at a time, and it was easy to accidentally drag a boundary on the wrong track.

4.0's clip model treats each clip as a first-class object you click on directly, and a few specific changes save real time once you're working across a dozen files in one project:

• Multi-clip selection and editing: Shift-click several clip headers and moving, trimming, or time-stretching applies to all of them at once, instead of repeating the same drag four times.
• Grouping: clips can be grouped so a sound effect and its reverb tail stay locked together through moves, copies and duplicates.
• Freer placement: a clip can move between mono and stereo tracks, and dropping it onto another clip replaces the overlapped section instead of just blocking the move.
• A dedicated split tool: hold S and click the waveform or clip header, with split-cut, split-delete, split-at-silence and split-to-new-track available as separate commands.

Split-at-silence in particular replaced a manual step in my process — I used to eyeball silence gaps in a long voice-over take before slicing it into individual onboarding lines. Now that's one command instead of five careful selections.

Where It Broke My Workflow: Macro Manager and the Scripting Pipe Are Gone

Here's where the upgrade actually cost me time. My old pipeline drove Audacity from a Python script over the mod-script-pipe, importing each raw WAV, normalizing it, and exporting it as Ogg Vorbis into the app's res/raw folder in one unattended run. It looked roughly like this:

import pipeclient

client = pipeclient.PipeClient()

def export_clip(src_path, dest_path):
    client.write(f"Import2: Filename={src_path}")
    client.write("SelectAll:")
    client.write("Normalize: ApplyGain=1 PeakLevel=-1")
    client.write(f"Export2: Filename={dest_path} NumChannels=1")
    print(client.read())

for wav_file in raw_sound_assets:
    export_clip(wav_file, wav_file.replace(".wav", ".ogg"))

That script had been running unmodified for two Audacity major versions, which is exactly why I didn't think to check whether it still applied before upgrading. Audacity 4.0's compatibility notes are explicit that the Macro Manager and the scripting pipe are not available in this release, and there is no built-in replacement shipped alongside it — the notes just say they're working on adding it back later.

My fix was to stop asking Audacity to do the batch conversion at all. I still open individual clips in 4.0 for the manual trimming and split-at-silence work, but the bulk normalize-and-export step moved to a plain shell script using sox for loudness normalization and ffmpeg for the Ogg Vorbis conversion, which is what most of these command-line audio tools are for anyway:

#!/usr/bin/env bash
# Batch normalize + convert app sounds for res/raw, no Audacity automation involved
for wav in raw_sound_assets/*.wav; do
  name=$(basename "$wav" .wav)
  sox "$wav" "/tmp/${name}_norm.wav" norm -1
  ffmpeg -y -i "/tmp/${name}_norm.wav" -c:a libvorbis -q:a 4 \
    "app/src/main/res/raw/${name}.ogg"
done

It's a longer chain of tools than before, but it's also not tied to whatever Audacity's automation story looks like in a given release. If you're prepping Android sound resources at any real volume, I'd honestly recommend keeping that part of the pipeline outside Audacity regardless of version — it just happened that 4.0 forced the decision on me earlier than planned.

.aup3, .aup4, and What Happens When You Open an Old Project

Audacity 4 introduces a new .aup4 project format, and the migration path only goes one direction. Opening an existing .aup3 project converts it to .aup4 without touching the original file on disk, which is a sensible default — but the converted copy cannot be saved back to .aup3, so there's no round-tripping once you've opened it in 4.0. Older .aup projects (the pre-3.0 format) can still be imported, just not opened directly.

That one-way door matters if you work with anyone still on 3.x, or if you keep long-lived project files around for revisiting old app releases. I keep archived Audacity projects per app version for exactly that reason, and I now have to decide, per project, whether I want it staying editable in 3.x or moving forward into 4.0 for good.

Practically, this means I duplicate the project folder before opening anything I might need to hand back to a 3.x install, since the conversion doesn't leave a way back short of keeping a separate copy myself.

Workspaces, Themes, and Actually Usable High-DPI Rendering

The Qt rewrite shows up most obviously in how the interface handles a high-resolution display. I do most of this editing on a laptop with a 4K panel hooked up to an external monitor, and previous Audacity versions had the usual mixed-DPI awkwardness — blurry icons on one screen, oddly-scaled toolbars on the other. That's gone in 4.0.

Audacity 4 also adds UI layouts saved as Workspaces, with Modern, Classic and Music presets built in, plus light, dark and high-contrast themes with configurable accent colors, track colors and clip styles. I switched to the Classic workspace out of habit for the first week and then just gave up and used Modern once I stopped hunting for buttons in their old 3.x positions.

The new Home screen showing recent projects with preview thumbnails is a small thing, but genuinely useful once you have a dozen archived app-sound projects sitting in the same folder and can't remember which one has the notification chime you're looking for.

Recording and Playback Changes That Matter During a Voice-Over Pass

A chunk of my sound-asset work is re-recording a single onboarding line after a script tweak, which means punching in on top of an existing track without redoing the whole take. Audacity 4.0 rebuilt Punch and Roll, lead-in recording, latency compensation, software playthrough and per-track input monitoring for the new interface, and recording can now start anywhere on the timeline and create a clip right at that position instead of forcing you into a fixed insertion point.

The playhead also stays visible during navigation now and can be dragged directly to a new position, and playback can seek without stopping first. Those sound like small conveniences until you're doing take after take of a two-second voice line and every extra click adds up over an afternoon.

On the plugin side, supported formats are VST3, Nyquist, LV2 on Linux and Audio Units on macOS, and Audacity can generate a fallback control panel when a plugin doesn't provide its own interface. I don't lean heavily on third-party plugins for short app sounds, but if your pipeline depends on a specific VST or AU compressor for consistent loudness across assets, it's worth confirming it still loads before you commit a whole project to 4.0.

What's Still Missing, and Whether I'd Upgrade Your Build Pipeline Today

Beyond the Macro Manager and scripting pipe, the compatibility notes list several other Audacity 3 features not yet available in 4.0: Time Tracks, Note/MIDI tracks, the Mixer, and VAMP and LADSPA plugin hosting are all missing, along with Play-at-speed. The notes frame these as things being worked on for future releases, not permanent removals, but there's no date attached to any of it.

If any of those are load-bearing for how you work — MIDI tracks for scoring, LADSPA plugins you've relied on for years, Play-at-speed for proofing dialogue at a faster tempo — that's worth checking against your own project list before you touch the upgrade button on a machine you actually depend on.

For my own setup, I've landed on running 4.0 for the actual clip editing — trimming, splitting, multi-clip moves — because that part of the experience is a clear improvement, and keeping the batch normalize-and-export step in shell scripts outside Audacity entirely, which turned out to be more resilient to this kind of release churn anyway. I'm not reinstalling 3.x, but I'm also not deleting it yet, and I'd suggest the same dual-track approach to anyone whose Android asset pipeline currently assumes Audacity can drive itself end to end.

Sources: Hacker News — Audacity 4.0