hackmaster/Robovoice
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

v0.8: TCP transport via gatuna tunnel, session IDs, pre-synth playback, server Rust rewrite

Browse Source
This commit is contained in:
Mute
2026-08-13 09:58:33 +00:00
parent ecf00c1bc4
commit facbfe6a5c
16 changed files with 955 additions and 638 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Robovoice.App/AppConfig.cs
+1 -1
View File
@@ -12,7 +12,7 @@ public sealed class AppConfig
public int LengthScale { get; set; } = 100;
public int NoiseWScale { get; set; } = 800;
public bool MinimizeToTray { get; set; } = true;
public string InterfaceIp { get; set; } = string.Empty;
public string SttEndpoint { get; set; } = "127.0.0.1:6996";
public static string AppDataDir => Path.Combine(
Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData),
Robovoice.App/MainForm.Designer.cs
Generated
+8 -9
View File
@@ -17,7 +17,7 @@ partial class MainForm
private Button btnBrowseFile = null!;
private Label lblFileName = null!;
private Label lblServer = null!;
private ComboBox cmbInterface = null!;
private TextBox txtSttEndpoint = null!;
private Label lblLineStatus = null!;
private RichTextBox txtLog = null!;
private CheckBox chkMinimizeToTray = null!;
@@ -57,7 +57,7 @@ partial class MainForm
btnBrowseFile = new Button();
lblFileName = new Label();
lblServer = new Label();
cmbInterface = new ComboBox();
txtSttEndpoint = new TextBox();
lblLineStatus = new Label();
txtLog = new RichTextBox();
chkMinimizeToTray = new CheckBox();
@@ -143,15 +143,14 @@ partial class MainForm
lblFileName.ForeColor = Color.Gray;
// lblServer
lblServer.Text = "Interface:";
lblServer.Text = "STT:";
lblServer.Location = new Point(440, 48);
lblServer.Size = new Size(55, 23);
lblServer.Size = new Size(35, 23);
lblServer.TextAlign = ContentAlignment.MiddleLeft;
// cmbInterface
cmbInterface.Location = new Point(498, 45);
cmbInterface.Size = new Size(165, 23);
cmbInterface.DropDownStyle = ComboBoxStyle.DropDownList;
// txtSttEndpoint
txtSttEndpoint.Location = new Point(478, 45);
txtSttEndpoint.Size = new Size(185, 23);
// lblLineStatus
lblLineStatus.Text = "";
@@ -270,7 +269,7 @@ partial class MainForm
Controls.Add(btnBrowseFile);
Controls.Add(lblFileName);
Controls.Add(lblServer);
Controls.Add(cmbInterface);
Controls.Add(txtSttEndpoint);
Controls.Add(lblLineStatus);
Controls.Add(lblNoise);
Controls.Add(trkNoise);
Robovoice.App/MainForm.cs
+69 -101
View File
@@ -2,7 +2,7 @@ using NAudio.Wave;
using Robovoice.App;
using Robovoice.Core;
using Robovoice.Core.Voices;
using Robovoice.Stt.Dhcp;
using Robovoice.Stt.Tcp;
using Robovoice.Tts.LibPiper;
using System.Diagnostics;
@@ -16,7 +16,7 @@ internal sealed partial class MainForm : Form
private LibPiperTtsEngine? _tts;
private AudioOutput? _audioOutput;
private DhcpSttSource? _sttSource;
private TcpSttSource? _sttSource;
private Orchestrator? _orchestrator;
private PttHotkey? _pttHotkey;
private NotifyIcon? _trayIcon;
@@ -36,57 +36,64 @@ internal sealed partial class MainForm : Form
private async void OnLoad(object? sender, EventArgs e)
{
ActiveControl = txtLog;
PopulateVoices();
PopulateOutputDevices();
try
{
ActiveControl = txtLog;
PopulateVoices();
PopulateOutputDevices();
_pttKey = (Keys)_config.PttKey;
if (_pttKey == Keys.None)
_pttKey = Keys.F8;
txtPttKey.Text = KeyToDisplayString(_pttKey);
_pttKey = (Keys)_config.PttKey;
if (_pttKey == Keys.None)
_pttKey = Keys.F8;
txtPttKey.Text = KeyToDisplayString(_pttKey);
if (!string.IsNullOrEmpty(_config.Voice) && cmbVoice.Items.Contains(_config.Voice))
cmbVoice.SelectedItem = _config.Voice;
else if (cmbVoice.Items.Count > 0)
cmbVoice.SelectedIndex = 0;
if (!string.IsNullOrEmpty(_config.Voice) && cmbVoice.Items.Contains(_config.Voice))
cmbVoice.SelectedItem = _config.Voice;
else if (cmbVoice.Items.Count > 0)
cmbVoice.SelectedIndex = 0;
if (!string.IsNullOrEmpty(_config.OutputDevice) && cmbOutput.Items.Contains(_config.OutputDevice))
cmbOutput.SelectedItem = _config.OutputDevice;
else
AutoSelectCableOutput();
if (!string.IsNullOrEmpty(_config.OutputDevice) && cmbOutput.Items.Contains(_config.OutputDevice))
cmbOutput.SelectedItem = _config.OutputDevice;
else
AutoSelectCableOutput();
trkNoise.Value = _config.NoiseScale;
trkSpeed.Value = _config.LengthScale;
trkNoiseW.Value = _config.NoiseWScale;
OnSliderScroll(null, EventArgs.Empty);
trkNoise.Value = _config.NoiseScale;
trkSpeed.Value = _config.LengthScale;
trkNoiseW.Value = _config.NoiseWScale;
OnSliderScroll(null, EventArgs.Empty);
chkMinimizeToTray.Checked = _config.MinimizeToTray;
PopulateInterfaces();
chkMinimizeToTray.Checked = _config.MinimizeToTray;
txtSttEndpoint.Text = _config.SttEndpoint;
btnBrowseFile.Click += OnBrowseFile;
btnTestVoice.Click += OnTestVoice;
btnManageVoices.Click += OnManageVoices;
btnClearLog.Click += (_, _) => txtLog.Clear();
txtPttKey.Enter += OnPttKeyFocus;
txtPttKey.KeyDown += OnPttKeyDown;
cmbOutput.SelectedIndexChanged += OnOutputChanged;
cmbVoice.SelectedIndexChanged += OnVoiceChanged;
cmbInterface.SelectedIndexChanged += OnInterfaceChanged;
btnBrowseFile.Click += OnBrowseFile;
btnTestVoice.Click += OnTestVoice;
btnManageVoices.Click += OnManageVoices;
btnClearLog.Click += (_, _) => txtLog.Clear();
txtPttKey.Enter += OnPttKeyFocus;
txtPttKey.KeyDown += OnPttKeyDown;
cmbOutput.SelectedIndexChanged += OnOutputChanged;
cmbVoice.SelectedIndexChanged += OnVoiceChanged;
txtSttEndpoint.Leave += OnSttEndpointChanged;
trkNoise.Scroll += OnSliderScroll;
trkSpeed.Scroll += OnSliderScroll;
trkNoiseW.Scroll += OnSliderScroll;
trkNoise.MouseUp += OnSliderReleased;
trkSpeed.MouseUp += OnSliderReleased;
trkNoiseW.MouseUp += OnSliderReleased;
trkNoise.Scroll += OnSliderScroll;
trkSpeed.Scroll += OnSliderScroll;
trkNoiseW.Scroll += OnSliderScroll;
trkNoise.MouseUp += OnSliderReleased;
trkSpeed.MouseUp += OnSliderReleased;
trkNoiseW.MouseUp += OnSliderReleased;
chkMinimizeToTray.CheckedChanged += (_, _) => SaveConfig();
chkMinimizeToTray.CheckedChanged += (_, _) => SaveConfig();
Resize += OnResize;
Resize += OnResize;
SetupTray();
SetupTray();
}
catch (Exception ex)
{
System.Diagnostics.Debug.WriteLine($"OnLoad failed: {ex}");
}
await InitializeEngineAsync();
BeginInvoke(async () => await InitializeEngineAsync());
}
private Keys _pttKey = Keys.F8;
@@ -232,7 +239,8 @@ internal sealed partial class MainForm : Form
Log($"Loading voice: {voiceName}...");
_audioOutput?.Dispose();
_tts?.DisposeAsync().AsTask().Wait();
if (_tts is not null)
await _tts.DisposeAsync();
_tts = new LibPiperTtsEngine(
modelPath,
@@ -242,21 +250,21 @@ internal sealed partial class MainForm : Form
noiseWScale: trkNoiseW.Value / 1000.0f);
_audioOutput = new AudioOutput();
if (_sttSource is null)
{
string ifaceIp = cmbInterface.SelectedItem as string ?? "";
_sttSource = new DhcpSttSource { InterfaceIp = ifaceIp, Log = Log };
await _sttSource.StartAsync();
}
_orchestrator?.DisposeAsync().AsTask().Wait();
_orchestrator = new Orchestrator(_tts, _audioOutput, _sttSource, Log)
{
OutputDeviceName = cmbOutput.SelectedItem as string ?? string.Empty,
};
try
{
if (_sttSource is null)
{
_sttSource = new TcpSttSource { Endpoint = txtSttEndpoint.Text, Log = Log };
await _sttSource.StartAsync();
}
if (_orchestrator is not null)
await _orchestrator.DisposeAsync();
_orchestrator = new Orchestrator(_tts, _audioOutput, _sttSource, Log)
{
OutputDeviceName = cmbOutput.SelectedItem as string ?? string.Empty,
};
await _orchestrator.InitializeTtsAsync();
Log("Engine ready. Press PTT to send to STT server.");
SetupHotkey();
@@ -426,59 +434,19 @@ internal sealed partial class MainForm : Form
SaveConfig();
}
private void PopulateInterfaces()
private async void OnSttEndpointChanged(object? sender, EventArgs e)
{
cmbInterface.Items.Clear();
foreach (var (ip, name) in DhcpSttSource.GetAvailableInterfaces())
{
cmbInterface.Items.Add(name);
cmbInterface.Items[^1] = name;
cmbInterface.Items[cmbInterface.Items.Count - 1] = name;
}
if (!string.IsNullOrEmpty(_config.InterfaceIp))
{
for (int i = 0; i < cmbInterface.Items.Count; i++)
{
if (cmbInterface.Items[i] is string s && s.Contains(_config.InterfaceIp))
{
cmbInterface.SelectedIndex = i;
return;
}
}
}
if (cmbInterface.Items.Count > 0)
cmbInterface.SelectedIndex = 0;
}
private void OnInterfaceChanged(object? sender, EventArgs e)
{
if (cmbInterface.SelectedItem is not string selected)
return;
string? ip = ExtractIpFromDisplay(selected);
if (ip is null) return;
_config.InterfaceIp = ip;
_config.SttEndpoint = txtSttEndpoint.Text;
SaveConfig();
if (_sttSource is not null)
{
_sttSource.DisposeAsync().AsTask().Wait(2000);
await _sttSource.DisposeAsync();
_sttSource = null;
_ = InitializeEngineAsync();
await InitializeEngineAsync();
}
}
private static string? ExtractIpFromDisplay(string display)
{
int start = display.IndexOf('(');
int end = display.IndexOf(')');
if (start < 0 || end <= start) return null;
return display[(start + 1)..end];
}
private void SetupTray()
{
if (_trayInit) return;
@@ -544,7 +512,7 @@ internal sealed partial class MainForm : Form
_config.LengthScale = trkSpeed.Value;
_config.NoiseWScale = trkNoiseW.Value;
_config.MinimizeToTray = chkMinimizeToTray.Checked;
_config.InterfaceIp = ExtractIpFromDisplay(cmbInterface.SelectedItem as string ?? "") ?? "";
_config.SttEndpoint = txtSttEndpoint.Text;
_config.Save();
}
Robovoice.App/Orchestrator.cs
+7 -33
View File
@@ -19,8 +19,6 @@ internal sealed class Orchestrator : IAsyncDisposable
private bool _playing;
private Task? _synthTask;
private CancellationTokenSource? _synthCts;
private System.Threading.Timer? _segmentTimer;
private static readonly TimeSpan SegmentTimeout = TimeSpan.FromMilliseconds(250);
public string OutputDeviceName { get; set; } = string.Empty;
@@ -51,7 +49,6 @@ internal sealed class Orchestrator : IAsyncDisposable
_log($"SEGMENT: \"{msg.Text}\" ({msg.Text.Length} chars)");
_pendingTexts.Enqueue(msg.Text);
EnsureSynthTask();
ResetSegmentTimer();
}
}
else
@@ -67,33 +64,11 @@ internal sealed class Orchestrator : IAsyncDisposable
_log("FINAL: (empty)");
}
CancelSegmentTimer();
TransitionToPlaying();
}
}
}
private void ResetSegmentTimer()
{
_segmentTimer?.Dispose();
_segmentTimer = new System.Threading.Timer(_ => OnSegmentTimeout(), null, SegmentTimeout, Timeout.InfiniteTimeSpan);
}
private void CancelSegmentTimer()
{
_segmentTimer?.Dispose();
_segmentTimer = null;
}
private void OnSegmentTimeout()
{
_log("STT: segment timeout, starting playback early");
lock (_stateLock)
{
TransitionToPlaying();
}
}
private void TransitionToPlaying()
{
if (_playing)
@@ -104,13 +79,11 @@ internal sealed class Orchestrator : IAsyncDisposable
if (_synthTask is null || _synthTask.IsCompleted)
{
_log("TTS: nothing to play");
CancelSegmentTimer();
}
return;
}
_playing = true;
CancelSegmentTimer();
int sampleRate = _bufferSampleRate;
var chunks = _audioBuffer.ToList();
@@ -189,7 +162,6 @@ internal sealed class Orchestrator : IAsyncDisposable
{
lock (_stateLock)
{
CancelSegmentTimer();
_synthCts?.Cancel();
_synthCts?.Dispose();
_synthTask = null;
@@ -209,7 +181,7 @@ internal sealed class Orchestrator : IAsyncDisposable
public async Task InitializeTtsAsync()
{
_log("Initializing TTS engine...");
await _tts.InitializeAsync();
await Task.Run(() => _tts.InitializeAsync());
_log($"TTS ready (sample rate: {_tts.SampleRate} Hz)");
}
@@ -269,15 +241,17 @@ internal sealed class Orchestrator : IAsyncDisposable
public async ValueTask DisposeAsync()
{
if (_disposed) return;
_currentCts?.Cancel();
_disposed = true;
try { _currentCts?.Cancel(); } catch { }
_currentCts?.Dispose();
_synthCts?.Cancel();
try { _synthCts?.Cancel(); } catch { }
_synthCts?.Dispose();
CancelSegmentTimer();
_sttSource.TranscriptReceived -= OnTranscript;
await _sttSource.DisposeAsync();
await _tts.DisposeAsync();
_audioOutput.Dispose();
_disposed = true;
}
}
Robovoice.App/Robovoice.App.csproj
+1 -1
View File
@@ -3,7 +3,7 @@
<ItemGroup>
<ProjectReference Include="..\Robovoice.Core\Robovoice.Core.csproj" />
<ProjectReference Include="..\Robovoice.Tts.LibPiper\Robovoice.Tts.LibPiper.csproj" />
<ProjectReference Include="..\Robovoice.Stt.Dhcp\Robovoice.Stt.Dhcp.csproj" />
<ProjectReference Include="..\Robovoice.Stt.Tcp\Robovoice.Stt.Tcp.csproj" />
</ItemGroup>
<ItemGroup>