diff --git a/CHANGELOG.md b/CHANGELOG.md index 8d07020..c1c51d7 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,15 @@ # Changelog +## 1.2.0 — v1.2.0 · The microphone stays where you put it + +- Add **Level** to Settings → Recording. Switched on, it starts at the level the microphone has now; the slider sets it from 0 to 100 %. While Utterform runs it looks every two seconds and sets the level back when Teams, Zoom, Chrome or anything else has changed it. Switched off, Utterform never touches the level. +- Windows: the microphone chosen for recordings, else the default one — the same slider as in Sound settings. Linux: the default input of PipeWire (`wpctl`) or PulseAudio (`pactl`). macOS: the default input. +- Choosing another microphone in the tray moves the held level to it. +- Where the level cannot be read, the switch stays off and says why. +- App version 1.2.0; the installers install v1.2.0 by default. + +See [1.2.0 release notes](docs/releases/v1.2.0.md). + ## 1.1.0 — v1.1.0 · The right microphone, one click away - Add **Microphone** to the tray menu: **System default (name)** and every microphone, the one in use checked. A click makes it the microphone for the next recordings — the laptop's own while the headset for calls stays plugged in. It is the same setting as Settings → Recording → Microphone, saved at once, and the window follows. diff --git a/README.md b/README.md index e38d17c..068677f 100644 --- a/README.md +++ b/README.md @@ -28,7 +28,7 @@ Utterform records a short voice clip, transcribes it with **OpenAI GPT Transcrib the cursor. The window can stay out of the way while a global shortcut controls the whole flow. -> **1.1.0 — The right microphone, one click away** switches the microphone from the tray menu and brings everything since 0.8.0: a first-run assistant, Settings with search, ChatGPT text actions, Context, an action per app and audio files. The Apple-Silicon build is Developer ID signed and notarized; Windows remains unsigned. See the [release notes](docs/releases/v1.1.0.md) and [downloads](https://github.com/jli-software/utterform/releases/tag/v1.1.0). +> **1.2.0 — The microphone stays where you put it** holds the microphone level you choose, even when Teams or another app turns it up or down. 1.1.0 brought the microphone in the tray menu and everything since 0.8.0: a first-run assistant, Settings with search, ChatGPT text actions, Context, an action per app and audio files. The Apple-Silicon build is Developer ID signed and notarized; Windows remains unsigned. See the [release notes](docs/releases/v1.2.0.md) and [downloads](https://github.com/jli-software/utterform/releases/tag/v1.2.0). ## How it works @@ -333,6 +333,7 @@ Push/PR CI runs Linux validation without release compilation. For a test binary - A vocabulary of your own terms, and a reasoning effort for the text step - Clipboard, TXT, Markdown, or combined output - Selectable microphone with a system-default fallback, in Settings and in the tray menu +- Hold the microphone level, so other apps cannot change how loud you record - Global dictation with a reserved key combination on Windows, macOS and X11, or a compositor binding on Wayland: either way without raising the window - Optional typing of the finished text into the focused window, next to clipboard and file, as one paste or as keystrokes (Linux and Windows) - A dictation key recorded by pressing it, read by physical key so the layout cannot change it diff --git a/docs/ARCHITECTURE.md b/docs/ARCHITECTURE.md index c2d2e00..a30e9b9 100644 --- a/docs/ARCHITECTURE.md +++ b/docs/ARCHITECTURE.md @@ -294,6 +294,33 @@ Linux has no such hover event; there the list follows the window. **Diagnostics.** `tray.microphone_chosen` (`default` or `device`) and `tray.microphones_failed` (step). Device names are not logged. +## Microphone level (1.2.0) + +**What it holds.** `settings.microphone_level`: a percentage, or `None` to leave the +level to the operating system (the default, and what an older settings file reads as). +`mic_level.rs` runs one thread, started in `setup`, that receives the target over a +channel — from `init` (the settings file), `save_settings`, `complete_onboarding` and the +tray's microphone choice — and, while a level is set, reads it every two seconds and +writes it back when it is more than one point away (rounding in the system moves it by +one). Nothing is read or written while no level is set. + +**Per system.** Windows: `IMMDeviceEnumerator::GetDevice` with the endpoint id from the +cpal id `wasapi:`, else the default capture endpoint, then +`IAudioEndpointVolume::{Get,Set}MasterVolumeLevelScalar` — the Sound settings slider; +COM is initialised once on the thread. Linux: `wpctl get-volume/set-volume +@DEFAULT_AUDIO_SOURCE@`, else `pactl get-source-volume/set-source-volume +@DEFAULT_SOURCE@`; the ALSA ids Utterform records from do not map to PipeWire nodes, so it +is always the default source. macOS: `osascript` `input volume of (get volume settings)` / +`set volume input volume`, the default input. + +**Settings.** Recording → Level: the switch reads the current level +(`read_microphone_level`) and starts there; if that fails the switch stays off and the +reason is shown. The slider saves after 300 ms. + +**Diagnostics.** `mic_level.target` (level or `off`), `mic_level.restored` (from, to; at +most once a minute), `mic_level.failed` once until it works again, +`mic_level.read_failed`. No device names. + ## First-run assistant (0.14.0) **Who sees it.** `settings.onboarding_completed`. `AppSettings::default()` — what `load` diff --git a/docs/DEVELOPMENT.md b/docs/DEVELOPMENT.md index fdfec1b..718ab79 100644 --- a/docs/DEVELOPMENT.md +++ b/docs/DEVELOPMENT.md @@ -19,7 +19,19 @@ proposals are in [plans/v1.0-roadmap.md](plans/v1.0-roadmap.md). The plan of 202 (reliability, assistant, finish) is superseded; 0.7.13 shipped, 0.7.14 (`feat/simple-settings`) was rejected and will not be merged. -## Current handoff — 1.1.0 Microphone in the tray +## Current handoff — 1.2.0 Hold the microphone level + +Branch `work/utterform-1.2.0-hold-mic-level`. Jonas asked on 2026-11-13 for the +microphone to stay set (a RØDE NT-USB on Windows; its knobs only change the headphone +mix) without a separate tool, on Windows and Linux alike. Settings → Recording → Level +holds a level while Utterform runs. Design in +[ARCHITECTURE.md → Microphone level](ARCHITECTURE.md#microphone-level-120). Released +directly as 1.2.0 like 1.1.0. CBZIT007 has no Visual C++ Build Tools: the Windows code was +compiled by `Desktop builds`, the frontend and its Playwright tests ran locally (Edge via +`PLAYWRIGHT_CHROMIUM_EXECUTABLE`). Open: a real check with the NT-USB and Teams running, +and Linux on a PipeWire desktop. + +## Previous handoff — 1.1.0 Microphone in the tray Branch `work/utterform-1.1.0-tray-microphone`. Jonas asked on 2026-11-13 for a quick microphone switch in the tray (headset plugged in for calls, but dictating into the diff --git a/docs/plans/v1.0-roadmap.md b/docs/plans/v1.0-roadmap.md index 4ac30e6..95a3fe1 100644 --- a/docs/plans/v1.0-roadmap.md +++ b/docs/plans/v1.0-roadmap.md @@ -79,6 +79,7 @@ Proposals not yet confirmed by Jonas are marked **Proposal**. | 0.14.0 | First-run assistant | merged, ships with 1.0 | | 1.0.0 | Acceptance on Linux, Windows, macOS; release; WinGet | rc.1 published; not released on its own — ships as 1.1.0 | | 1.1.0 | Microphone in the tray; first stable release since 0.8.0 | released (2026-11-13) | +| 1.2.0 | Hold the microphone level | released (2026-11-13) | **Proposal (order):** the Omarchy backend first because it is small and the CLI already exists; Settings next so every later feature gets a fixed place; Codex before context diff --git a/docs/releases/v1.2.0.md b/docs/releases/v1.2.0.md new file mode 100644 index 0000000..aab99c9 --- /dev/null +++ b/docs/releases/v1.2.0.md @@ -0,0 +1,32 @@ +# Utterform 1.2.0 — The microphone stays where you put it + +**Hold the microphone level.** How loud a microphone records is set by the operating +system, and other programs change it without asking: Teams, Zoom and Chrome turn it up +and down on their own. Open Settings → Recording and switch on **Level**. It starts at the +level your microphone has now; move the slider to where it sounds right. From then on +Utterform keeps it there while it runs and sets it back within two seconds when something +else changed it — for dictation, for a call, for a video you record. + +Switch it off and Utterform leaves the level alone again. + +- **Windows:** the microphone chosen for recordings, otherwise the default one. It is the + same level as in Sound settings → Input. +- **Linux:** the default input of PipeWire or PulseAudio. +- **macOS:** the default input. + +A tip for USB microphones with knobs, such as the RØDE NT-USB: the knobs set what you +hear in the headphones. What is recorded is set here. + +## Install + +Quit Utterform first. + +- **Linux (Omarchy / Arch):** + `curl -fsSL https://github.com/jli-software/utterform/releases/download/v1.2.0/install-linux.sh | sh` +- **Windows (PowerShell):** + `irm https://github.com/jli-software/utterform/releases/download/v1.2.0/install-windows.ps1 | iex` +- **macOS:** the [DMG](https://github.com/jli-software/utterform/releases/download/v1.2.0/utterform-macos-aarch64.dmg), + Developer ID signed and notarized. + +Windows builds remain unsigned and may trigger SmartScreen. Verify downloads against +`SHA256SUMS.txt`. diff --git a/package-lock.json b/package-lock.json index b7b282e..1aa1e5f 100644 --- a/package-lock.json +++ b/package-lock.json @@ -1,12 +1,12 @@ { "name": "utterform", - "version": "1.1.0", + "version": "1.2.0", "lockfileVersion": 3, "requires": true, "packages": { "": { "name": "utterform", - "version": "1.1.0", + "version": "1.2.0", "dependencies": { "@tauri-apps/api": "^2.11.1", "@tauri-apps/plugin-dialog": "^2.4.0", diff --git a/package.json b/package.json index 89af544..87806f4 100644 --- a/package.json +++ b/package.json @@ -1,7 +1,7 @@ { "name": "utterform", "private": true, - "version": "1.1.0", + "version": "1.2.0", "type": "module", "scripts": { "dev": "vite", diff --git a/scripts/install-linux.sh b/scripts/install-linux.sh index 6891cf6..259e1c1 100755 --- a/scripts/install-linux.sh +++ b/scripts/install-linux.sh @@ -2,7 +2,7 @@ set -eu -version="${UTTERFORM_VERSION:-v1.1.0}" +version="${UTTERFORM_VERSION:-v1.2.0}" release_base="${UTTERFORM_RELEASE_BASE_URL:-https://github.com/jli-software/utterform/releases/download/${version}}" asset="utterform-linux-x86_64-system.tar.gz" prefix="${UTTERFORM_PREFIX:-${HOME}/.local}" diff --git a/scripts/install-windows.ps1 b/scripts/install-windows.ps1 index 3ccabcd..26e66bd 100644 --- a/scripts/install-windows.ps1 +++ b/scripts/install-windows.ps1 @@ -1,6 +1,6 @@ # Verify the NSIS setup before opening it; no elevation or forced app shutdown. param( - [string]$Version = 'v1.1.0', + [string]$Version = 'v1.2.0', [switch]$VerifyOnly ) diff --git a/src-tauri/Cargo.lock b/src-tauri/Cargo.lock index e8270af..1d2afbd 100644 --- a/src-tauri/Cargo.lock +++ b/src-tauri/Cargo.lock @@ -5904,7 +5904,7 @@ checksum = "b6c140620e7ffbb22c2dee59cafe6084a59b5ffc27a8859a5f0d494b5d52b6be" [[package]] name = "utterform" -version = "1.1.0" +version = "1.2.0" dependencies = [ "anyhow", "base64 0.22.1", diff --git a/src-tauri/Cargo.toml b/src-tauri/Cargo.toml index 69c56e0..f17a50e 100644 --- a/src-tauri/Cargo.toml +++ b/src-tauri/Cargo.toml @@ -1,6 +1,6 @@ [package] name = "utterform" -version = "1.1.0" +version = "1.2.0" description = "Fast, lightweight voice-to-text transformation for the desktop" authors = ["Jonas"] edition = "2024" @@ -66,6 +66,8 @@ zbus = "5.9" [target.'cfg(target_os = "windows")'.dependencies] windows-sys = { version = "0.61", features = ["Win32_UI_Input_KeyboardAndMouse", "Win32_UI_WindowsAndMessaging", "Win32_UI_Accessibility", "Win32_System_Threading", "Win32_Security", "Win32_Foundation"] } winreg = "0.55" +# The microphone level (`mic_level`). Already in the build through cpal. +windows = { version = "0.62", features = ["Win32_Media_Audio", "Win32_Media_Audio_Endpoints", "Win32_System_Com", "Win32_System_Com_StructuredStorage"] } [target.'cfg(target_os = "macos")'.dependencies] # Typing at the cursor posts keyboard events through CoreGraphics; the frontmost diff --git a/src-tauri/src/commands.rs b/src-tauri/src/commands.rs index a445553..83041bb 100644 --- a/src-tauri/src/commands.rs +++ b/src-tauri/src/commands.rs @@ -114,6 +114,7 @@ pub fn save_settings(app: AppHandle, value: AppSettings) -> Result<(), String> { app.state::() .set(!value.onboarding_completed); crate::tray_microphones::set_selected(&app, value.input_device.clone()); + crate::mic_level::set(&app, value.microphone_level, value.input_device.clone()); diagnostics::info!( "settings.saved", engine = value.engine.as_str(), @@ -152,6 +153,7 @@ pub fn complete_onboarding( } app.state::().set(false); crate::tray_microphones::set_selected(&app, value.input_device.clone()); + crate::mic_level::set(&app, value.microphone_level, value.input_device.clone()); if skipped { diagnostics::info!("onboarding.skipped"); } else { @@ -178,6 +180,16 @@ pub fn list_input_devices(app: AppHandle) -> Result, String Ok(devices) } +/// The microphone's level now, in percent, for Settings → Recording → Level +/// to start from. An error says why this system cannot hold the level. +#[tauri::command] +pub fn read_microphone_level(device: Option) -> Result { + crate::mic_level::read(device.as_deref()).map_err(|error| { + diagnostics::warning!("mic_level.read_failed", error = error); + error + }) +} + /// Open the microphone and play the start cue. /// /// Asynchronous so the work leaves the main thread: a synchronous command runs diff --git a/src-tauri/src/domain.rs b/src-tauri/src/domain.rs index caf2993..6e923cb 100644 --- a/src-tauri/src/domain.rs +++ b/src-tauri/src/domain.rs @@ -143,6 +143,9 @@ pub struct AppSettings { pub engine: TranscriptionEngine, pub cloud_model: CloudModel, pub input_device: Option, + /// The microphone level to hold, in percent; `None` leaves it to the + /// operating system. See `mic_level`. + pub microphone_level: Option, pub output_directory: Option, pub copy_to_clipboard: bool, pub save_to_file: bool, @@ -227,6 +230,7 @@ impl Default for AppSettings { engine: TranscriptionEngine::OpenAi, cloud_model: CloudModel::GptTranscribe, input_device: None, + microphone_level: None, output_directory: None, copy_to_clipboard: true, save_to_file: false, diff --git a/src-tauri/src/lib.rs b/src-tauri/src/lib.rs index 5c34184..d929961 100644 --- a/src-tauri/src/lib.rs +++ b/src-tauri/src/lib.rs @@ -15,6 +15,7 @@ mod hotkey; mod live; #[cfg(target_os = "macos")] mod macos; +mod mic_level; mod models; mod output; mod platform; @@ -271,6 +272,7 @@ fn setup(app: &mut tauri::App, startup_intent: Intent) -> Result<(), Box, + pub device: Option, +} + +/// The way to the watcher thread; `None` until [`init`] started it. +#[derive(Default)] +pub struct MicLevel(Mutex>>); + +/// Start watching with the level in the settings file. +pub fn init(app: &AppHandle) { + let (sender, receiver) = mpsc::channel(); + let started = thread::Builder::new() + .name("mic-level".into()) + .spawn(move || watch(receiver)); + if let Err(error) = started { + diagnostics::warning!("mic_level.thread_failed", error = error); + return; + } + if let Ok(mut slot) = app.state::().0.lock() { + *slot = Some(sender); + } + if let Ok(value) = settings::load(app) { + set(app, value.microphone_level, value.input_device); + } +} + +/// Hold `level` (or stop holding with `None`) on `device`. Applied at once. +pub fn set(app: &AppHandle, level: Option, device: Option) { + let target = Target { + level: level.map(|level| level.min(100)), + device, + }; + if let Ok(slot) = app.state::().0.lock() + && let Some(sender) = slot.as_ref() + { + let _ = sender.send(target); + } +} + +fn watch(receiver: mpsc::Receiver) { + platform::init_thread(); + let mut target = Target::default(); + let mut last_log: Option = None; + let mut failing = false; + loop { + let wait = if target.level.is_some() { + INTERVAL + } else { + Duration::from_secs(3600) + }; + match receiver.recv_timeout(wait) { + Ok(next) => { + if next.level != target.level { + diagnostics::info!( + "mic_level.target", + level = next + .level + .map_or_else(|| "off".to_string(), |level| level.to_string()), + ); + } + target = next; + failing = false; + } + Err(mpsc::RecvTimeoutError::Timeout) => {} + Err(mpsc::RecvTimeoutError::Disconnected) => return, + } + let Some(level) = target.level else { continue }; + match hold(level, target.device.as_deref()) { + Ok(Some(previous)) => { + failing = false; + if last_log.is_none_or(|at| at.elapsed() >= LOG_INTERVAL) { + last_log = Some(Instant::now()); + diagnostics::info!("mic_level.restored", from = previous, to = level); + } + } + Ok(None) => failing = false, + Err(error) => { + if !failing { + failing = true; + diagnostics::warning!("mic_level.failed", error = error); + } + } + } + } +} + +/// Set the level back if it moved; `Some(previous)` when it had to. +fn hold(level: u8, device: Option<&str>) -> Result, String> { + let current = platform::read(device)?; + if !needs_correction(current, level) { + return Ok(None); + } + platform::write(device, level)?; + Ok(Some(current)) +} + +/// Rounding in the operating system moves a level by a point; that is no change. +fn needs_correction(current: u8, target: u8) -> bool { + current.abs_diff(target) > 1 +} + +/// The level now, for Settings to start the slider at. +pub fn read(device: Option<&str>) -> Result { + platform::read(device) +} + +fn percent(fraction: f32) -> u8 { + (fraction.clamp(0.0, 1.0) * 100.0).round() as u8 +} + +/// `wpctl get-volume`: "Volume: 0.40" or "Volume: 0.40 [MUTED]". +#[cfg_attr(not(target_os = "linux"), allow(dead_code))] +fn parse_wpctl(output: &str) -> Option { + let value = output + .trim() + .strip_prefix("Volume:")? + .split_whitespace() + .next()?; + value.parse::().ok().map(percent) +} + +/// `pactl get-source-volume`: "Volume: front-left: 26214 / 40% / -23.88 dB, …". +#[cfg_attr(not(target_os = "linux"), allow(dead_code))] +fn parse_pactl(output: &str) -> Option { + output + .split_whitespace() + .find_map(|word| word.strip_suffix('%')?.parse::().ok()) + .map(|value| value.min(100) as u8) +} + +/// `osascript -e "input volume of (get volume settings)"`: "75", or +/// "missing value" for a microphone without a level. +#[cfg_attr(not(target_os = "macos"), allow(dead_code))] +fn parse_osascript(output: &str) -> Option { + output + .trim() + .parse::() + .ok() + .map(|value| value.min(100) as u8) +} + +#[cfg(target_os = "windows")] +mod platform { + use windows::{ + Win32::{ + Media::Audio::{ + Endpoints::IAudioEndpointVolume, IMMDevice, IMMDeviceEnumerator, + MMDeviceEnumerator, eCapture, eConsole, + }, + System::Com::{CLSCTX_ALL, COINIT_MULTITHREADED, CoCreateInstance, CoInitializeEx}, + }, + core::HSTRING, + }; + + pub fn init_thread() { + // The watcher thread is Utterform's own; nothing else initialised COM on it. + let _ = unsafe { CoInitializeEx(None, COINIT_MULTITHREADED) }; + } + + /// The chosen microphone (cpal names it `wasapi:`), else + /// the default input — the same fallback a recording takes. + fn endpoint(device: Option<&str>) -> Result { + unsafe { + let enumerator: IMMDeviceEnumerator = + CoCreateInstance(&MMDeviceEnumerator, None, CLSCTX_ALL) + .map_err(|error| format!("enumerator: {error}"))?; + let chosen: Option = device + .and_then(|id| id.strip_prefix("wasapi:")) + .and_then(|id| enumerator.GetDevice(&HSTRING::from(id)).ok()); + let device = match chosen { + Some(device) => device, + None => enumerator + .GetDefaultAudioEndpoint(eCapture, eConsole) + .map_err(|error| format!("no microphone: {error}"))?, + }; + device + .Activate::(CLSCTX_ALL, None) + .map_err(|error| format!("volume: {error}")) + } + } + + pub fn read(device: Option<&str>) -> Result { + let volume = endpoint(device)?; + let level = unsafe { volume.GetMasterVolumeLevelScalar() } + .map_err(|error| format!("read: {error}"))?; + Ok(super::percent(level)) + } + + pub fn write(device: Option<&str>, level: u8) -> Result<(), String> { + let volume = endpoint(device)?; + unsafe { volume.SetMasterVolumeLevelScalar(f32::from(level) / 100.0, std::ptr::null()) } + .map_err(|error| format!("write: {error}")) + } +} + +#[cfg(target_os = "linux")] +mod platform { + use std::process::Command; + + pub fn init_thread() {} + + fn run(program: &str, arguments: &[&str]) -> Option { + let output = Command::new(program).args(arguments).output().ok()?; + output + .status + .success() + .then(|| String::from_utf8_lossy(&output.stdout).into_owned()) + } + + pub fn read(_device: Option<&str>) -> Result { + if let Some(output) = run("wpctl", &["get-volume", "@DEFAULT_AUDIO_SOURCE@"]) { + return super::parse_wpctl(&output).ok_or_else(|| "wpctl: unreadable".into()); + } + if let Some(output) = run("pactl", &["get-source-volume", "@DEFAULT_SOURCE@"]) { + return super::parse_pactl(&output).ok_or_else(|| "pactl: unreadable".into()); + } + Err("neither wpctl (PipeWire) nor pactl (PulseAudio) answered".into()) + } + + pub fn write(_device: Option<&str>, level: u8) -> Result<(), String> { + let fraction = format!("{:.2}", f32::from(level) / 100.0); + if run( + "wpctl", + &["set-volume", "@DEFAULT_AUDIO_SOURCE@", &fraction], + ) + .is_some() + { + return Ok(()); + } + let percent = format!("{level}%"); + if run( + "pactl", + &["set-source-volume", "@DEFAULT_SOURCE@", &percent], + ) + .is_some() + { + return Ok(()); + } + Err("neither wpctl (PipeWire) nor pactl (PulseAudio) set the level".into()) + } +} + +#[cfg(target_os = "macos")] +mod platform { + use std::process::Command; + + pub fn init_thread() {} + + fn osascript(script: &str) -> Result { + let output = Command::new("osascript") + .args(["-e", script]) + .output() + .map_err(|error| format!("osascript: {error}"))?; + if !output.status.success() { + return Err("osascript failed".into()); + } + Ok(String::from_utf8_lossy(&output.stdout).into_owned()) + } + + pub fn read(_device: Option<&str>) -> Result { + let output = osascript("input volume of (get volume settings)")?; + super::parse_osascript(&output).ok_or_else(|| "this microphone has no level".into()) + } + + pub fn write(_device: Option<&str>, level: u8) -> Result<(), String> { + osascript(&format!("set volume input volume {level}")).map(|_| ()) + } +} + +#[cfg(not(any(target_os = "windows", target_os = "linux", target_os = "macos")))] +mod platform { + pub fn init_thread() {} + + pub fn read(_device: Option<&str>) -> Result { + Err("not available on this system".into()) + } + + pub fn write(_device: Option<&str>, _level: u8) -> Result<(), String> { + Err("not available on this system".into()) + } +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn a_point_of_rounding_is_no_change() { + assert!(!needs_correction(70, 70)); + assert!(!needs_correction(69, 70)); + assert!(!needs_correction(71, 70)); + assert!(needs_correction(68, 70)); + assert!(needs_correction(100, 70)); + assert!(needs_correction(0, 70)); + } + + #[test] + fn wpctl_volume_is_read_muted_or_not() { + assert_eq!(parse_wpctl("Volume: 0.40\n"), Some(40)); + assert_eq!(parse_wpctl("Volume: 0.73 [MUTED]\n"), Some(73)); + assert_eq!(parse_wpctl("Volume: 1.50\n"), Some(100)); + assert_eq!(parse_wpctl("nothing\n"), None); + } + + #[test] + fn pactl_volume_takes_the_first_channel() { + let output = "Volume: front-left: 26214 / 40% / -23.88 dB, front-right: 26214 / 40% / -23.88 dB\n balance 0.00\n"; + assert_eq!(parse_pactl(output), Some(40)); + assert_eq!( + parse_pactl("Volume: mono: 98304 / 150% / 10.57 dB"), + Some(100) + ); + assert_eq!(parse_pactl("No such entity"), None); + } + + #[test] + fn osascript_level_or_none() { + assert_eq!(parse_osascript("75\n"), Some(75)); + assert_eq!(parse_osascript("missing value\n"), None); + } + + #[test] + fn fractions_round_to_whole_percent() { + assert_eq!(percent(0.0), 0); + assert_eq!(percent(0.704), 70); + assert_eq!(percent(1.2), 100); + assert_eq!(percent(-0.1), 0); + } +} diff --git a/src-tauri/src/tray_microphones.rs b/src-tauri/src/tray_microphones.rs index 05f09a6..9a27852 100644 --- a/src-tauri/src/tray_microphones.rs +++ b/src-tauri/src/tray_microphones.rs @@ -182,14 +182,19 @@ pub fn refresh(app: &AppHandle, throttled: bool) { pub fn choose(app: &AppHandle, selected: Option) { let saved = settings::load(app).and_then(|mut value| { value.input_device = selected.clone(); - settings::save(app, &value) + settings::save(app, &value).map(|()| value.microphone_level) }); - if let Err(failure) = saved { - diagnostics::warning!("tray.microphone_choice_failed", class = failure.class()); - // Show the choice that is still in force. - update(app, |_| {}); - return; - } + let level = match saved { + Ok(level) => level, + Err(failure) => { + diagnostics::warning!("tray.microphone_choice_failed", class = failure.class()); + // Show the choice that is still in force. + update(app, |_| {}); + return; + } + }; + // A held level follows the microphone to the one now chosen. + crate::mic_level::set(app, level, selected.clone()); diagnostics::info!( "tray.microphone_chosen", microphone = if selected.is_some() { diff --git a/src-tauri/tauri.conf.json b/src-tauri/tauri.conf.json index d323c24..a045f41 100644 --- a/src-tauri/tauri.conf.json +++ b/src-tauri/tauri.conf.json @@ -1,7 +1,7 @@ { "$schema": "https://schema.tauri.app/config/2", "productName": "Utterform", - "version": "1.1.0", + "version": "1.2.0", "identifier": "software.jli.utterform", "build": { "beforeDevCommand": "npm run dev", diff --git a/src/App.svelte b/src/App.svelte index 2e2cae5..8323b84 100644 --- a/src/App.svelte +++ b/src/App.svelte @@ -64,6 +64,7 @@ /// How long a text field waits after the last keystroke before it is written. const TYPING_SAVE_DELAY_MS = 500; + const LEVEL_SAVE_DELAY_MS = 300; const ENGINES: Array<{ mode: TranscriptionMode; label: string; title: string }> = [ { mode: "gpt_transcribe", label: "Cloud", title: "GPT Transcribe: finished text, audio is sent to OpenAI" }, { mode: "gpt_live_transcribe", label: "Live", title: "GPT Live Transcribe: text appears at the cursor while you speak" }, @@ -119,6 +120,7 @@ // appears and then turns out to be dead is worse than one that arrives late. let hotkeySupport: HotkeySupport = { supported: false, default: "Ctrl+Alt+D", explanation: "", failure: null }; let hotkeyError = ""; + let levelMessage = ""; /// While the recorder is listening, the key it is listening for must not be /// held by the operating system on Utterform's behalf, or pressing it would /// start a recording instead of being read. Released for that moment only, @@ -1007,6 +1009,23 @@ /// Put a change into effect and write it. `delay` is for text fields: the /// write waits for a pause in typing, and the next change restarts the wait. + /// Holding starts at the level the microphone has now, so switching it on + /// changes nothing you hear until the slider moves. + async function toggleMicrophoneLevel(toggle: HTMLInputElement) { + levelMessage = ""; + if (!toggle.checked) { + change({ microphone_level: null }); + return; + } + try { + const level = await api.readMicrophoneLevel(settings.input_device); + change({ microphone_level: Math.max(0, Math.min(100, Math.round(level))) }); + } catch (error) { + levelMessage = `Utterform cannot set the microphone level here: ${String(error)}`; + toggle.checked = false; + } + } + function change(patch: Partial, delay = 0) { settings = { ...settings, ...patch }; scheduleSave(delay); @@ -1658,6 +1677,20 @@
{@render rowHead("microphone", "Microphone")}
change({ input_device: value || null })} />
{/if} + {#if shownRows.has("mic-level")} +
{@render rowHead("mic-level", "Level", settings.microphone_level !== null ? "mic-level" : null)} +
+ {#if settings.microphone_level !== null} + change({ microphone_level: Number(event.currentTarget.value) }, LEVEL_SAVE_DELAY_MS)} /> + {settings.microphone_level}% + {/if} + toggleMicrophoneLevel(event.currentTarget)} /> +
+ {#if levelMessage}{/if} +
+ {/if} {#if shownRows.has("shortcut")}
{@render rowHead("shortcut", "Shortcut")}
diff --git a/src/app.css b/src/app.css index 2bcf5c7..35973ef 100644 --- a/src/app.css +++ b/src/app.css @@ -118,6 +118,10 @@ main:not(.has-result) footer { margin-top:20px; } .row-control.wide { flex:1; max-width:400px; } .row-message { flex-basis:100%; margin:0; text-align:right; } .row-note { color:var(--muted); font-size:11px; line-height:1.5; } +.level-slider { width:140px; height:20px; min-height:0; padding:0; border:none; background:transparent; accent-color:var(--accent); cursor:pointer; } +.level-slider:focus { border:none; outline:none; } +.level-slider:focus-visible { outline:2px solid var(--accent); outline-offset:2px; } +.level-value { min-width:32px; text-align:right; font-variant-numeric:tabular-nums; } .row-input { width:min(300px,100%); min-height:34px; height:34px; font-size:12px; } .row-input.narrow { width:76px; } .row-button { min-height:30px; padding:0 10px; border:1px solid var(--line); border-radius:5px; background:transparent; color:var(--text); font-size:12px; white-space:nowrap; } diff --git a/src/lib/api.ts b/src/lib/api.ts index 1df6462..7a21e25 100644 --- a/src/lib/api.ts +++ b/src/lib/api.ts @@ -61,6 +61,8 @@ export const api = { copyText: (text: string) => invoke("copy_text", { text }), saveSettings: (value: AppSettings) => invoke("save_settings", { value }), listInputDevices: () => invoke("list_input_devices"), + // The microphone's level now, in percent; rejects where it cannot be held. + readMicrophoneLevel: (device: string | null) => invoke("read_microphone_level", { device }), startRecording: ( inputDevice: string | null, engine: Engine, diff --git a/src/lib/settings-catalog.test.ts b/src/lib/settings-catalog.test.ts index 605e03f..da642bd 100644 --- a/src/lib/settings-catalog.test.ts +++ b/src/lib/settings-catalog.test.ts @@ -9,7 +9,7 @@ describe("the settings catalog", () => { }); it("shows a category's own rows when nothing is searched", () => { - expect([...visibleRows("recording", " ")]).toEqual(["microphone", "shortcut", "sounds"]); + expect([...visibleRows("recording", " ")]).toEqual(["microphone", "mic-level", "shortcut", "sounds"]); }); it("finds rows across categories by label, keyword or category, ignoring case", () => { diff --git a/src/lib/settings-catalog.ts b/src/lib/settings-catalog.ts index 86dccc0..c09931f 100644 --- a/src/lib/settings-catalog.ts +++ b/src/lib/settings-catalog.ts @@ -14,7 +14,7 @@ export const SETTINGS_CATEGORIES: Array<{ id: SettingsCategory; label: string }> ]; export type SettingsRowId = - | "microphone" | "shortcut" | "sounds" + | "microphone" | "mic-level" | "shortcut" | "sounds" | "engine" | "api-key" | "language" | "vocabulary" | "context" | "local-model" | "actions" | "text-provider" | "text-model" | "context-selection" | "context-clipboard" | "app-actions" @@ -31,6 +31,7 @@ export interface SettingsRow { export const SETTINGS_ROWS: SettingsRow[] = [ { id: "microphone", category: "recording", label: "Microphone", keywords: ["mic", "input", "device", "audio"] }, + { id: "mic-level", category: "recording", label: "Level", keywords: ["volume", "gain", "loud", "quiet", "input level", "hold", "lock", "teams", "pegel"] }, { id: "shortcut", category: "recording", label: "Shortcut", keywords: ["hotkey", "key", "keyboard", "dictation", "global"] }, { id: "sounds", category: "recording", label: "Sounds", keywords: ["sound", "cue", "beep", "chime", "click", "test"] }, { id: "engine", category: "transcription", label: "Engine", keywords: ["cloud", "live", "local", "whisper", "gpt", "transcribe", "mode", "offline"] }, diff --git a/src/lib/types.ts b/src/lib/types.ts index 3e96ed3..0b387ff 100644 --- a/src/lib/types.ts +++ b/src/lib/types.ts @@ -33,6 +33,8 @@ export interface AppSettings { engine: Engine; cloud_model: CloudModel; input_device: string | null; + /// The microphone level Utterform holds, in percent; `null` leaves it alone. + microphone_level: number | null; output_directory: string | null; copy_to_clipboard: boolean; save_to_file: boolean; @@ -181,6 +183,7 @@ export const DEFAULT_SETTINGS: AppSettings = { engine: "open_ai", cloud_model: "gpt_transcribe", input_device: null, + microphone_level: null, output_directory: null, copy_to_clipboard: true, save_to_file: false, diff --git a/tests/ui.e2e.ts b/tests/ui.e2e.ts index bf3d9ff..c24702b 100644 --- a/tests/ui.e2e.ts +++ b/tests/ui.e2e.ts @@ -81,6 +81,7 @@ test.beforeEach(async ({ page }) => { case "codex_status": return { signedIn: !Reflect.get(window, "__codexSignedOut"), models: [{ id: "gpt-6.1-sol", name: "GPT-6.1-Sol" }, { id: "gpt-6-luna", name: "GPT-6-Luna" }], defaultModel: "gpt-6-luna" }; case "open_codex_setup": (Reflect.get(window, "__supportCalls") as string[]).push("codex-setup"); return; case "list_input_devices": return [{ id: "test-mic", name: "Studio microphone", isDefault: true }]; + case "read_microphone_level": if (Reflect.get(window, "__levelUnavailable")) throw "neither wpctl (PipeWire) nor pactl (PulseAudio) answered"; return 63; case "list_local_models": return models; case "download_local_model": case "delete_local_model": { const model = models.find((model) => model.id === args.modelId); @@ -927,6 +928,39 @@ test("text actions can run on the ChatGPT subscription, marked experimental", as await page.screenshot({ path: testInfo.outputPath("ai-chatgpt-model.png") }); }); +test("the microphone level is held from where it is, and moved with the slider", async ({ page }, testInfo) => { + await page.goto("/"); + await page.getByRole("button", { name: "Open settings" }).click(); + const hold = page.getByRole("switch", { name: "Hold the microphone level" }); + await expect(hold).not.toBeChecked(); + await expect(page.getByRole("slider")).toHaveCount(0); + await hold.click(); + const level = () => page.evaluate(() => (Reflect.get(window, "__savedSettings") as { microphone_level?: number | null } | null)?.microphone_level); + await expect.poll(level).toBe(63); + const slider = page.getByRole("slider", { name: "Level" }); + await expect(slider).toHaveValue("63"); + await slider.fill("80"); + await expect.poll(level).toBe(80); + await expect(page.getByText("80%")).toBeVisible(); + await page.mouse.move(0, 0); + await page.screenshot({ path: testInfo.outputPath("microphone-level.png") }); + await hold.click(); + await expect.poll(level).toBeNull(); + await expect(page.getByRole("slider")).toHaveCount(0); + expect(await page.evaluate(() => document.documentElement.scrollWidth <= window.innerWidth)).toBe(true); +}); + +test("where the level cannot be set, the switch stays off and says why", async ({ page }) => { + await page.addInitScript(() => Object.assign(window, { __levelUnavailable: true })); + await page.goto("/"); + await page.getByRole("button", { name: "Open settings" }).click(); + const hold = page.getByRole("switch", { name: "Hold the microphone level" }); + await hold.click(); + await expect(page.getByRole("alert")).toContainText("cannot set the microphone level"); + await expect(hold).not.toBeChecked(); + expect(await page.evaluate(() => (Reflect.get(window, "__savedSettings") as { microphone_level?: number | null } | null)?.microphone_level ?? null)).toBeNull(); +}); + test("the Context action's sources are switched in their own category, at once", async ({ page }, testInfo) => { await page.goto("/"); await page.getByRole("button", { name: "Open settings" }).click();