Add leading silence detection and removal logic

nvdaHelper/local/silenceDetect.h

@@ -1,3 +1,8 @@
+// A part of NonVisual Desktop Access (NVDA)
+// This file is covered by the GNU General Public License.
+// See the file COPYING for more details.
+// Copyright (C) 2025 NV Access Limited, gexgd0419
+
 #ifndef SILENCEDETECT_H
 #define SILENCEDETECT_H
 
@@ -173,42 +178,6 @@ size_t getLeadingSilenceSizeMono(
 	return size;
 }
 
-/**
- * Return the trailing silence wave data length, in bytes.
- * Assumes the wave data to be of one channel (mono).
- * Uses a `WaveFormat` type (`Fmt`) to determine the wave format.
- */
-template <class Fmt>
-size_t getTrailingSilenceSizeMono(
-	const unsigned char* waveData,
-	size_t size,
-	typename Fmt::SampleType threshold
-) {
-	using SampleType = Fmt::SampleType;
-	constexpr size_t bytesPerSample = Fmt::bytesPerSample;
-
-	if (size < bytesPerSample)
-		return 0;
-
-	constexpr SampleType zeroPoint = Fmt::zeroPoint();
-	const SampleType minValue = zeroPoint - threshold, maxValue = zeroPoint + threshold;
-
-	// Check each sample in reverse order
-	const unsigned char* p = waveData + (size - (size % bytesPerSample));
-	SampleType smp = SampleType();
-	do {
-		p -= bytesPerSample;
-		memcpy(&smp, p, bytesPerSample);
-		smp = Fmt::signExtend(smp);
-		// this sample is out of range, so the trailing silence starts at the next sample
-		if (smp < minValue || smp > maxValue)
-			return size - (p - waveData) - bytesPerSample;
-	} while (p > waveData);
-
-	// The whole data block is silence
-	return size;
-}
-
 /**
  * Invoke a functor with an argument of a WaveFormat type that corresponds to the specified WAVEFORMATEX.
  * Return false if the WAVEFORMATEX is unknown.
@@ -272,29 +241,6 @@ inline size_t getLeadingSilenceSize(
 	return len - len % wfx->nBlockAlign;  // round down to block (channel) boundaries
 }
 
-/**
- * Return the trailing silence wave data length, in bytes.
- * Uses a `WAVEFORMATEX` to determine the wave format.
- */
-inline size_t getTrailingSilenceSize(
-	const WAVEFORMATEX* wfx,
-	const unsigned char* waveData,
-	size_t size
-) {
-	size_t len;
-	if (!callByWaveFormat(wfx, [=, &len](auto fmtTag) {
-			using Fmt = decltype(fmtTag);
-			len = getTrailingSilenceSizeMono<Fmt>(
-				waveData, size, Fmt::defaultThreshold());
-		}))
-		return 0;
-
-	size_t align = wfx->nBlockAlign;
-	len += align - 1;
-	len -= len % align;  // round up to block (channel) boundaries
-	return len;
-}
-
 }  // namespace SilenceDetect
 
 #endif  // SILENCEDETECT_H

source/gui/settingsDialogs.py

@@ -3755,7 +3755,7 @@ def __init__(self, parent):
 		#  Advanced settings panel.
 		label = _("Trim leading silence in speech audio")
 		self.trimLeadingSilenceCheckBox = speechGroup.addItem(wx.CheckBox(speechBox, label=label))
-		self.bindHelpEvent("AdvancedSettingsTrimLeadingSilence", self.trimLeadingSilenceCheckBox)
+		self.bindHelpEvent("TrimLeadingSilence", self.trimLeadingSilenceCheckBox)
 		self.trimLeadingSilenceCheckBox.SetValue(config.conf["speech"]["trimLeadingSilence"])
 		self.trimLeadingSilenceCheckBox.defaultValue = self._getDefaultValue(["speech", "trimLeadingSilence"])
 
@@ -3985,6 +3985,14 @@ def restoreToDefaults(self):
 
 	def onSave(self):
 		log.debug("Saving advanced config")
+
+		if config.conf["speech"]["trimLeadingSilence"] != self.trimLeadingSilenceCheckBox.IsChecked():
+			# Reload the synthesizer if "trimLeadingSilence" changes
+			config.conf["speech"]["trimLeadingSilence"] = self.trimLeadingSilenceCheckBox.IsChecked()
+			currentSynth = getSynth()
+			if not setSynth(currentSynth.name):
+				_synthWarningDialog(currentSynth.name)
+
 		config.conf["development"]["enableScratchpadDir"] = self.scratchpadCheckBox.IsChecked()
 		selectiveUIAEventRegistrationChoice = self.selectiveUIAEventRegistrationCombo.GetSelection()
 		config.conf["UIA"]["eventRegistration"] = self.selectiveUIAEventRegistrationVals[
@@ -3997,7 +4005,6 @@ def onSave(self):
 		config.conf["featureFlag"]["cancelExpiredFocusSpeech"] = (
 			self.cancelExpiredFocusSpeechCombo.GetSelection()
 		)
-		config.conf["speech"]["trimLeadingSilence"] = self.trimLeadingSilenceCheckBox.IsChecked()
 		config.conf["UIA"]["allowInChromium"] = self.UIAInChromiumCombo.GetSelection()
 		self.enhancedEventProcessingComboBox.saveCurrentValueToConf()
 		config.conf["terminals"]["speakPasswords"] = self.winConsoleSpeakPasswordsCheckBox.IsChecked()

source/nvwave.py

@@ -294,7 +294,9 @@ def __init__(
 				NVDAHelper.localLib.wasSilence_init(outputDevice)
 				WasapiWavePlayer._silenceDevice = outputDevice
 		# Enable trimming by default for speech only
-		self.enableTrimmingLeadingSilence(purpose is AudioPurpose.SPEECH)
+		self.enableTrimmingLeadingSilence(
+			purpose is AudioPurpose.SPEECH and config.conf["speech"]["trimLeadingSilence"],
+		)
 		if self._enableTrimmingLeadingSilence:
 			self.startTrimmingLeadingSilence()
 
@@ -481,8 +483,6 @@ def enableTrimmingLeadingSilence(self, enable: bool) -> None:
 
 	def startTrimmingLeadingSilence(self, start: bool = True) -> None:
 		"""Start or stop trimming the leading silence from the next audio chunk."""
-		if not config.conf["speech"]["trimLeadingSilence"]:
-			return  ## disabled by user
 		NVDAHelper.localLib.wasPlay_startTrimmingLeadingSilence(self._player, start)
 
 	def _setVolumeFromConfig(self):

user_docs/en/userGuide.md

@@ -3321,10 +3321,11 @@ This option enables behaviour which attempts to cancel speech for expired focus
 In particular moving quickly through messages in Gmail with Chrome can cause NVDA to speak outdated information.
 This functionality is enabled by default as of NVDA 2021.1.
 
-##### Trim leading silence in speech audio {#AdvancedSettingsTrimLeadingSilence}
+##### Trim leading silence in speech audio {#TrimLeadingSilence}
 
-This option enables NVDA to trim the silence at the beginning of the speech audio, making the voices respond faster.
-This option is enabled by default, and should only affect the silence at the beginning. But if you find that some necessary silence periods are also missing (e.g. pause between two sentences) when using a speech synthesizer add-on, you may try turning off this feature.
+When enabled, NVDA will remove silence from the start of speech audio, which may improve the responsiveness of some speech synthesizers.
+This option is enabled by default, and should only affect the silence at the beginning of speech.
+If you find that some necessary silence periods are also missing (e.g. pause between two sentences) when using a speech synthesizer add-on, you may turn this feature off entirely to resolve the issue.
 
 ##### Caret move timeout (in MS) {#AdvancedSettingsCaretMoveTimeout}