C# SpeechSynthesizer.SpeakSsml方法代码示例

        public void TestPhonemes()
        {
            EventWaitHandle waitHandle = new AutoResetEvent(false);

            using (MemoryStream stream = new MemoryStream())
            using (SpeechSynthesizer synth = new SpeechSynthesizer())
            {
                synth.SetOutputToWaveStream(stream);

                synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"UTF-8\"?><speak version=\"1.0\" xmlns=\"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>This is your <phoneme alphabet=\"ipa\" ph=\"leɪkɒn\">Lakon</phoneme>.</s></speak>");

                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"ˈdɛltə\">delta</phoneme> system.</s></speak>");
                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"bliːiː\">Bleae</phoneme> <phoneme alphabet=\"ipa\" ph=\"θuːə\">Thua</phoneme> system.</s></speak>");
                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the Amnemoi system.</s></speak>");
                //synth.Speak("You are travelling to the Barnard's Star system.");
                stream.Seek(0, SeekOrigin.Begin);

                IWaveSource source = new WaveFileReader(stream);

                var soundOut = new WasapiOut();
                soundOut.Stopped += (s, e) => waitHandle.Set();

                soundOut.Initialize(source);
                soundOut.Play();

                waitHandle.WaitOne();
                soundOut.Dispose();
                source.Dispose();
            }
        }

        public void Speak(string script, string voice, int echoDelay, int distortionLevel, int chorusLevel, int reverbLevel, int compressLevel, bool radio)
        {
            if (script == null) { return; }

            try
            {
                using (SpeechSynthesizer synth = new SpeechSynthesizer())
                using (MemoryStream stream = new MemoryStream())
                {
                    if (String.IsNullOrWhiteSpace(voice))
                    {
                        voice = configuration.StandardVoice;
                    }
                    if (voice != null)
                    {
                        try
                        {
                            synth.SelectVoice(voice);
                        }
                        catch { }
                    }

                    synth.Rate = configuration.Rate;

                    synth.SetOutputToWaveStream(stream);
                    string speech = SpeechFromScript(script);
                    if (speech.Contains("<phoneme"))
                    {
                        speech = "<?xml version=\"1.0\" encoding=\"UTF-8\"?><speak version=\"1.0\" xmlns=\"http://www.w3.org/2001/10/synthesis\" xml:lang=\"" + locale + "\"><s>" + speech + "</s></speak>";
                        synth.SpeakSsml(speech);
                    }
                    else
                    {
                        synth.Speak(speech);
                    }
                    stream.Seek(0, SeekOrigin.Begin);

                    using (System.IO.StreamWriter file = new System.IO.StreamWriter(Environment.GetEnvironmentVariable("AppData") + @"\EDDI\speech.log", true)) { file.WriteLine("" + System.Threading.Thread.CurrentThread.ManagedThreadId + ": Turned script " + script + " in to speech " + speech); }

                    IWaveSource source = new WaveFileReader(stream);

                    // We need to extend the duration of the wave source if we have any effects going on
                    if (chorusLevel != 0 || reverbLevel != 0 || echoDelay != 0)
                    {
                        // Add a base of 500ms plus 10ms per effect level over 50
                        source = source.AppendSource(x => new ExtendedDurationWaveSource(x, 500 + Math.Max(0, (configuration.EffectsLevel - 50) * 10)));
                    }

                    // Add various effects...

                    // We always have chorus
                    if (chorusLevel != 0)
                    {
                        source = source.AppendSource(x => new DmoChorusEffect(x) { Depth = chorusLevel, WetDryMix = Math.Min(100, (int)(180 * ((decimal)configuration.EffectsLevel) / ((decimal)100))), Delay = 16, Frequency = 2, Feedback = 25 });
                    }

                    // We only have reverb and echo if we're not transmitting or receiving
                    if (!radio)
                    {
                        if (reverbLevel != 0)
                        {
                            // We tone down the reverb level with the distortion level, as the combination is nasty
                            source = source.AppendSource(x => new DmoWavesReverbEffect(x) { ReverbTime = (int)(1 + 999 * ((decimal)configuration.EffectsLevel) / ((decimal)100)), ReverbMix = Math.Max(-96, -96 + (96 * reverbLevel / 100) - distortionLevel) });
                        }

                        if (echoDelay != 0)
                        {
                            // We tone down the echo level with the distortion level, as the combination is nasty
                            source = source.AppendSource(x => new DmoEchoEffect(x) { LeftDelay = echoDelay, RightDelay = echoDelay, WetDryMix = Math.Max(5, (int)(10 * ((decimal)configuration.EffectsLevel) / ((decimal)100)) - distortionLevel), Feedback = Math.Max(0, 10 - distortionLevel / 2) });
                        }
                    }

                    if (configuration.EffectsLevel > 0 && distortionLevel > 0)
                    {
                        source = source.AppendSource(x => new DmoDistortionEffect(x) { Edge = distortionLevel, Gain = -6 - (distortionLevel / 2), PostEQBandwidth = 4000, PostEQCenterFrequency = 4000 });
                    }

                    if (radio)
                    {
                        source = source.AppendSource(x => new DmoDistortionEffect(x) { Edge = 7, Gain = -4 - distortionLevel / 2, PostEQBandwidth = 2000, PostEQCenterFrequency = 6000 });
                        source = source.AppendSource(x => new DmoCompressorEffect(x) { Attack = 1, Ratio = 3, Threshold = -10 });
                    }

                    EventWaitHandle waitHandle = new EventWaitHandle(false, EventResetMode.AutoReset);
                    var soundOut = new WasapiOut();
                    soundOut.Initialize(source);
                    soundOut.Stopped += (s, e) => waitHandle.Set();

                    activeSpeeches.Add(soundOut);
                    soundOut.Play();

                    // Add a timeout, in case it doesn't come back
                    waitHandle.WaitOne(source.GetTime(source.Length));

                    // It's possible that this has been disposed of, so ensure that it's still there before we try to finish it
                    lock (activeSpeeches)
                    {
                        if (activeSpeeches.Contains(soundOut))
                        {
                            activeSpeeches.Remove(soundOut);
//.........这里部分代码省略.........

        public FileInfo MakeAudioFile(string ssmlToSpeak)
        {
            string wavFileName = DateTime.Now.ToString().Replace(" ", "");
            //AddInHost.Current.MediaCenterEnvironment.Dialog("Doing the playback", "", DialogButtons.Ok, 5, true);
            wavFileName = System.Environment.GetEnvironmentVariable("TEMP") + "\\" + wavFileName.Replace("-", "").Replace("/", "").Replace(":", "") + ".wav";
            //EventLog.WriteEntry("VmcController Client AddIn", "Generating speech file to " + wavFileName);

            ssmlToSpeak = "<speak version='1.0' xml:lang='en-US'>" + ssmlToSpeak + "<break time='2000ms'/></speak>";

            using (SpeechSynthesizer synth = new SpeechSynthesizer())
            {

                synth.SetOutputToWaveFile(wavFileName);

                synth.SpeakSsml(ssmlToSpeak);

            }

            //EventLog.WriteEntry("VmcController Client AddIn", "Speech file written to " + wavFileName);

            FileInfo fi = new FileInfo(wavFileName);

            return fi;
        }

        public void TestPhonemes()
        {
            EventWaitHandle waitHandle = new AutoResetEvent(false);

            using (MemoryStream stream = new MemoryStream())
            using (SpeechSynthesizer synth = new SpeechSynthesizer())
            {
                synth.SetOutputToWaveStream(stream);

                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"viˈga\">Vega</phoneme> system.</s></speak>");
                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"ækɜˈnɑ\">Achenar</phoneme> system.</s></speak>");
                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"ˈsɪɡni\">Cygni</phoneme> system.</s></speak>");
                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"ˈsɪɡnəs\">Cygnus</phoneme> system.</s></speak>");
                // synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"ʃɪnˈrɑːrtə\">Shinrarta</phoneme> <phoneme alphabet=\"ipa\" ph=\"ˈdezɦrə\">Dezhra</phoneme> system.</s></speak>");
                //synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the <phoneme alphabet=\"ipa\" ph=\"ˈnjuːənɛts\">Reorte</phoneme> system.</s></speak>");
                synth.SpeakSsml("<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?><speak version = \"1.0\" xmlns = \"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-GB\"><s>You are travelling to the Eravate system.</s></speak>");

                stream.Seek(0, SeekOrigin.Begin);

                IWaveSource source = new WaveFileReader(stream);

                var soundOut = new WasapiOut();
                soundOut.Stopped += (s, e) => waitHandle.Set();

                soundOut.Initialize(source);
                soundOut.Play();

                waitHandle.WaitOne();
                soundOut.Dispose();
                source.Dispose();
            }
        }

        public override void DownloadAndPlayItem(PhraseItem item, string folder)
        {
            String SSMLText = String.Format(@"<?xml version=""1.0""?>
            <speak xmlns=""http://www.w3.org/2001/10/synthesis"" version=""1.0"" xml:lang=""{1}"">
            <p>
            <s>{0}</s>
            </p>
            </speak>
            ", item.Phrase.Replace("&", "&amp;"), SelectedVoice.Language);

            try
            {
                new Task(() =>
                {
                    using (MemoryStream ms = new MemoryStream())
                    {
                        ms.Seek(0, SeekOrigin.Begin);
                        using (var synth = new SpeechSynthesizer())
                        {
                            synth.SelectVoice(SelectedVoice.Name);
                            synth.Volume = Int32.Parse(SelectedDiscreteVolume);
                            synth.Rate = Int32.Parse(SelectedDiscreteSpeed);
                            synth.SetOutputToWaveStream(ms);//.SetOutputToWaveFile(String.Format("{0}\\wav22050\\{1}\\{2}.wav", folder, item.Folder, item.FileName));
                            synth.SpeakSsml(SSMLText);
                        }
                        ms.Seek(0, SeekOrigin.Begin);
                        using (WaveFileReader wav = new WaveFileReader(ms))// String.Format("{0}\\wav22050\\{1}\\{2}.wav", folder, item.Folder, item.FileName)))
                        {
                            using (FileStream fs = new FileStream(String.Format("{0}\\mp3\\{1}\\{2}.mp3", folder, item.Folder, item.FileName), FileMode.Create))
                            {
                                using (var writer = new LameMP3FileWriter(fs, wav.WaveFormat, 128))
                                {
                                    wav.CopyTo(writer);
                                }
                            }
                        }
                        ms.Seek(0, SeekOrigin.Begin);
                        using (WaveFileReader wav = new WaveFileReader(ms))// String.Format("{0}\\wav22050\\{1}\\{2}.wav", folder, item.Folder, item.FileName)))
                        {
                            ConvertToWav(item, folder, true, new String[] { Name, SelectedVoice.Name, SelectedDiscreteSpeed, SelectedDiscreteVolume });
                        }
                    }
                }).Start();
            }
            catch (Exception Ex)
            {
                Logger.Log(Ex.ToString());
            }
        }

        public void Speak(string speech, string voice, int echoDelay, int distortionLevel, int chorusLevel, int reverbLevel, int compressLevel, bool wait = true, int priority = 3)
        {
            if (speech == null) { return; }

            Thread speechThread = new Thread(() =>
            {
                string finalSpeech = null;
                try
                {
                    using (SpeechSynthesizer synth = new SpeechSynthesizer())
                    using (MemoryStream stream = new MemoryStream())
                    {
                        if (string.IsNullOrWhiteSpace(voice))
                        {
                            voice = configuration.StandardVoice;
                        }
                        if (voice != null && !voice.Contains("Microsoft Server Speech Text to Speech Voice"))
                        {
                            try
                            {
                                Logging.Debug("Selecting voice " + voice);
                                synth.SelectVoice(voice);
                                Logging.Debug("Selected voice " + synth.Voice.Name);
                            }
                            catch (Exception ex)
                            {
                                Logging.Error("Failed to select voice " + voice, ex);
                            }
                        }

                        Logging.Debug("Post-selection");
                        Logging.Debug("Configuration is " + configuration == null ? "<null>" : JsonConvert.SerializeObject(configuration));
                        synth.Rate = configuration.Rate;
                        Logging.Debug("Rate is " + synth.Rate);
                        synth.Volume = configuration.Volume;
                        Logging.Debug("Volume is " + synth.Volume);

                        synth.StateChanged += new EventHandler<StateChangedEventArgs>(synth_StateChanged);
                        Logging.Debug("Tracking state changes");
                        synth.SetOutputToWaveStream(stream);
                        Logging.Debug("Output set to stream");
                        if (speech.Contains("<phoneme") || speech.Contains("<break"))
                        {
                            Logging.Debug("Speech is SSML");
                            if (configuration.DisableSsml)
                            {
                                Logging.Debug("Disabling SSML at user request");
                                // User has disabled SSML so remove it
                                finalSpeech = Regex.Replace(speech, "<.*?>", string.Empty);
                                synth.Speak(finalSpeech);
                            }
                            else
                            {
                                Logging.Debug("Obtaining best guess culture");
                                string culture = bestGuessCulture(synth);
                                Logging.Debug("Best guess culture is " + culture);
                                finalSpeech = "<?xml version=\"1.0\" encoding=\"UTF-8\"?><speak version=\"1.0\" xmlns=\"http://www.w3.org/2001/10/synthesis\" xml:lang=\"" + bestGuessCulture(synth) + "\"><s>" + speech + "</s></speak>";
                                Logging.Debug("SSML speech: " + finalSpeech);
                                try
                                {
                                    Logging.Debug("Speaking SSML");
                                    synth.SpeakSsml(finalSpeech);
                                    Logging.Debug("Finished speaking SSML");
                                }
                                catch (Exception ex)
                                {
                                    Logging.Error("Best guess culture of " + bestGuessCulture(synth) + " for voice " + synth.Voice.Name + " was incorrect", ex);
                                    Logging.Info("SSML does not work for the chosen voice; falling back to normal speech");
                                    // Try again without Ssml
                                    finalSpeech = Regex.Replace(speech, "<.*?>", string.Empty);
                                    synth.Speak(finalSpeech);
                                }
                            }
                        }
                        else
                        {
                            Logging.Debug("Speech does not contain SSML");
                            Logging.Debug("Speech: " + speech);
                            finalSpeech = speech;
                            Logging.Debug("Speaking normal speech");
                            synth.Speak(finalSpeech);
                            Logging.Debug("Finished speaking normal speech");
                        }
                        Logging.Debug("Seeking back to the beginning of the stream");
                        stream.Seek(0, SeekOrigin.Begin);

                        Logging.Debug("Setting up source from stream");
                        IWaveSource source = new WaveFileReader(stream);

                        // We need to extend the duration of the wave source if we have any effects going on
                        if (chorusLevel != 0 || reverbLevel != 0 || echoDelay != 0)
                        {
                            // Add a base of 500ms plus 10ms per effect level over 50
                            Logging.Debug("Extending duration by " + 500 + Math.Max(0, (configuration.EffectsLevel - 50) * 10) + "ms");
                            source = source.AppendSource(x => new ExtendedDurationWaveSource(x, 500 + Math.Max(0, (configuration.EffectsLevel - 50) * 10)));
                        }

                        // Add various effects...
                        Logging.Debug("Effects level is " + configuration.EffectsLevel + ", chorus level is " + chorusLevel + ", reverb level is " + reverbLevel + ", echo delay is " + echoDelay);
                        // We always have chorus
//.........这里部分代码省略.........