v1.1.2 - FFmpeg download persistence fix (issue #12)

SpeechCraft Studio v1.1.2

Fixes FFmpeg download going to a temp folder that was wiped on every launch.

What was wrong

When SpeechCraft runs as a packaged EXE, __file__ resolves to PyInstaller's MEIPASS — a per-launch extraction folder that Windows wipes on exit. The FFmpeg download copied ffmpeg.exe there, the current session "worked", but the file disappeared on the next launch. The success message was technically correct (the copy succeeded) but the file vanished immediately. Reported in issue #12 by PaulDrake777.

What's fixed

• New download destination: %LOCALAPPDATA%\SpeechCraft\ffmpeg.exe in the installed EXE (or next to audio_editor.py in dev runs). Survives across launches.
• setup_ffmpeg() now looks at the persistent location first, then the legacy path, then PATH.
• on_ffmpeg_download_complete re-verifies the file exists before showing the success message, and tells you the path + manual fallback URL on failure.
• Six regression tests in tests/test_ffmpeg_path.py pin the path resolution in place.

Manual workaround (if you already have the broken build)

1. Download ffmpeg.exe from https://www.gyan.dev/ffmpeg/builds/ (essentials build is enough).
2. Create the folder C:\Users\<your-user>\AppData\Local\SpeechCraft\ if it doesn't exist.
3. Drop ffmpeg.exe into that folder.
4. Restart SpeechCraft — it'll detect it and skip the download dialog.

Bundled assets

• SpeechCraft_Studio_Setup.exe (606 MB) — Windows installer with edition-choice page
• SpeechCraft_Studio_Core.exe (172 MB) — portable, Core only (no Whisper/pedalboard)
• SpeechCraft_Studio.exe (436 MB) — portable, Full edition with local AI

v1.1.1 - NSIS installer + edition-switch menu

SpeechCraft Studio v1.1.1

Adds a proper Windows installer plus the in-app Edition Switcher.

What's new

NSIS installer (SpeechCraft_Studio_Setup.exe, 606 MB)

A single-file Windows setup wizard that walks you through:

• Welcome page
• MIT license
• Edition choice (Core / Full radio buttons, Core pre-selected)
• Install progress bar
• Finish page with optional "Launch SpeechCraft Studio"

The installer places SpeechCraft Studio in C:\Program Files\SpeechCraft Studio\ with:

• Start Menu shortcuts
• A proper uninstaller (Control Panel → Programs → Uninstall)
• Registry entries for Add/Remove Programs

Edition switcher (Help → Switch Edition...)

No need to re-run the installer if you change your mind. The Switch Edition menu item:

• Opens the same Core/Full choice dialog as first-run onboarding
• Saves your new preference
• Prompts you to restart for the change to take effect

Both EXEs (SpeechCraft_Studio_Core.exe and SpeechCraft_Studio_Full.exe) are bundled, so the menu can switch between them instantly.

Fixed

• Core startup crash: removed eager import of pedalboard in dialogs/effects_dialogs.py. The Core EXE no longer triggers an "unhandled exception: pedalboard" when launched from the Start Menu.
• NSIS installer hang: File directives moved into the actual Section (where they compile correctly) and removed the post-install MessageBox that blocked the finish page from advancing.

Pick the right asset for you

Asset	Size	Use when
SpeechCraft_Studio_Setup.exe	606 MB	First-time install on Windows — recommended
SpeechCraft_Studio_Core.exe	172 MB	Portable use, no install needed, no Whisper/pedalboard
SpeechCraft_Studio.exe	435 MB	Portable use, no install needed, all features

If you already have v1.1.0 installed, you can keep using the EXE files directly — v1.1.1 only adds installer + menu polish, no behavioral changes inside the app.

v1.1.0 - Core + Full + Onboarding

SpeechCraft Studio v1.1.0

Two variants for the first time — pick what fits your needs.

Core — 172 MB (recommended)

Small download, fast launch. Includes:

• Recording with real-time transcription
• TTS (Edge TTS + Piper)
• Effects: compressor, de-esser, EQ, noise gate, normaliser, room tone match
• Breath smoothing
• Auto-ducker
• In-app bug reporting

Full — 435 MB

Adds everything in Core, plus:

• Local AI transcription (Whisper)
• Advanced effects (pedalboard)

First-run onboarding

On first launch, a dialog asks you which bundle type you're running. The choice is saved to %APPDATA%/SpeechCraft/setup.json. Next launches skip the dialog.

Default is Core — smaller bundle wins if you're not sure.

Try Core first

The Core EXE launches in roughly a third of the time of the Full bundle because Windows Defender has less to scan. If you find you need local Whisper or pedalboard later, you can switch from the menu bar (future PR).

Splash + lazy imports

The launch splash now stays up while heavy modules load. Each chunk of work announces itself so you always know if the app is stuck or still loading.

Looking for an installer?

A one-click NSIS installer (SpeechCraft_Studio_Setup.exe, 606 MB, bundles both variants) ships in v1.1.1.

v1.0.0 - Public Beta

First public release of SpeechCraft Studio.

Features

• Audio editing with multi-track support
• Text-to-speech (Edge TTS, Piper TTS)
• Effects: Compressor, De-esser, EQ, Noise Gate, Normalise, Room Tone Match
• Breath smoothing and auto-ducking
• Studio recording with real-time transcription
• In-app bug reporting

Requirements

• Windows 10/11
• NVDA 2024.1+ (recommended for screen reader users)

Installation

Download the EXE and run directly. No installer required.

Piper TTS Setup

For offline TTS, download Piper from https://github.com/rhasspy/piper/releases
Extract piper.exe and DLLs to the same folder as the EXE.

Support

• Bug reports: https://github.com/trasles16-ux/speechcraft-audio/issues
• Email: tracy@tracysmith.co.za