Alibaba Cloud Model Studio:Real-time speech synthesis - CosyVoice

Python

# coding=utf-8

import dashscope
from dashscope.audio.tts_v2 import *

# If you have not configured the API Key in an environment variable, replace "your-api-key" with your API Key.
# dashscope.api_key = "your-api-key"

# Model
# Different model versions require corresponding voice versions:
# cosyvoice-v3-flash/cosyvoice-v3-plus: Use voices such as longanyang.
# cosyvoice-v2: Use voices such as longxiaochun_v2.
model = "cosyvoice-v3-flash"
# Voice
voice = "longanyang"

# Instantiate SpeechSynthesizer and pass request parameters such as model and voice in the constructor.
synthesizer = SpeechSynthesizer(model=model, voice=voice)
# Send the text to be synthesized to get the binary audio.
audio = synthesizer.call("What is the weather like today?")
# The first time you send text, a WebSocket connection must be established, so the first-packet latency includes the connection setup time.
print('[Metric] Request ID: {}, First-packet latency: {} ms'.format(
    synthesizer.get_last_request_id(),
    synthesizer.get_first_package_delay()))

# Save the audio to a local file.
with open('output.mp3', 'wb') as f:
    f.write(audio)

Java

import com.alibaba.dashscope.audio.ttsv2.SpeechSynthesisParam;
import com.alibaba.dashscope.audio.ttsv2.SpeechSynthesizer;

import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.nio.ByteBuffer;

public class Main {
    // Model
    // Different model versions require corresponding voice versions:
    // cosyvoice-v3-flash/cosyvoice-v3-plus: Use voices such as longanyang.
    // cosyvoice-v2: Use voices such as longxiaochun_v2.
    private static String model = "cosyvoice-v3-flash";
    // Voice
    private static String voice = "longanyang";

    public static void streamAudioDataToSpeaker() {
        // Request parameters
        SpeechSynthesisParam param =
                SpeechSynthesisParam.builder()
                        // If you have not configured the API Key in an environment variable, uncomment the following line and replace "your-api-key" with your API Key.
                        // .apiKey("your-api-key")
                        .model(model) // Model
                        .voice(voice) // Voice
                        .build();

        // Synchronous mode: Disable callback (the second parameter is null).
        SpeechSynthesizer synthesizer = new SpeechSynthesizer(param, null);
        ByteBuffer audio = null;
        try {
            // Block until the audio is returned.
            audio = synthesizer.call("What is the weather like today?");
        } catch (Exception e) {
            throw new RuntimeException(e);
        } finally {
            // Close the WebSocket connection when the task is finished.
            synthesizer.getDuplexApi().close(1000, "bye");
        }
        if (audio != null) {
            // Save the audio data to the local file "output.mp3".
            File file = new File("output.mp3");
            // The first time you send text, a WebSocket connection must be established, so the first-packet latency includes the connection setup time.
            System.out.println(
                    "[Metric] Request ID: "
                            + synthesizer.getLastRequestId()
                            + " First-packet latency (ms): "
                            + synthesizer.getFirstPackageDelay());
            try (FileOutputStream fos = new FileOutputStream(file)) {
                fos.write(audio.array());
            } catch (IOException e) {
                throw new RuntimeException(e);
            }
        }
    }

    public static void main(String[] args) {
        streamAudioDataToSpeaker();
        System.exit(0);
    }
}

Python

Before running the Python example, you need to install a third-party audio playback library using pip.

# coding=utf-8
# Installation instructions for pyaudio:
# APPLE Mac OS X
#   brew install portaudio
#   pip install pyaudio
# Debian/Ubuntu
#   sudo apt-get install python-pyaudio python3-pyaudio
#   or
#   pip install pyaudio
# CentOS
#   sudo yum install -y portaudio portaudio-devel && pip install pyaudio
# Microsoft Windows
#   python -m pip install pyaudio

import pyaudio
import dashscope
from dashscope.audio.tts_v2 import *


from http import HTTPStatus
from dashscope import Generation

# If you have not configured the API Key in an environment variable, uncomment the following line and replace "apiKey" with your API Key.
# dashscope.api_key = "apiKey"

# Different model versions require corresponding voice versions:
# cosyvoice-v3-flash/cosyvoice-v3-plus: Use voices such as longanyang.
# cosyvoice-v2: Use voices such as longxiaochun_v2.
model = "cosyvoice-v3-flash"
voice = "longanyang"


class Callback(ResultCallback):
    _player = None
    _stream = None

    def on_open(self):
        print("websocket is open.")
        self._player = pyaudio.PyAudio()
        self._stream = self._player.open(
            format=pyaudio.paInt16, channels=1, rate=22050, output=True
        )

    def on_complete(self):
        print("speech synthesis task complete successfully.")

    def on_error(self, message: str):
        print(f"speech synthesis task failed, {message}")

    def on_close(self):
        print("websocket is closed.")
        # stop player
        self._stream.stop_stream()
        self._stream.close()
        self._player.terminate()

    def on_event(self, message):
        print(f"recv speech synthsis message {message}")

    def on_data(self, data: bytes) -> None:
        print("audio result length:", len(data))
        self._stream.write(data)


def synthesizer_with_llm():
    callback = Callback()
    synthesizer = SpeechSynthesizer(
        model=model,
        voice=voice,
        format=AudioFormat.PCM_22050HZ_MONO_16BIT,
        callback=callback,
    )

    messages = [{"role": "user", "content": "Please introduce yourself."}]
    responses = Generation.call(
        model="qwen-turbo",
        messages=messages,
        result_format="message",  # set result format as 'message'
        stream=True,  # enable stream output
        incremental_output=True,  # enable incremental output 
    )
    for response in responses:
        if response.status_code == HTTPStatus.OK:
            print(response.output.choices[0]["message"]["content"], end="")
            synthesizer.streaming_call(response.output.choices[0]["message"]["content"])
        else:
            print(
                "Request id: %s, Status code: %s, error code: %s, error message: %s"
                % (
                    response.request_id,
                    response.status_code,
                    response.code,
                    response.message,
                )
            )
    synthesizer.streaming_complete()
    print('Request ID: ', synthesizer.get_last_request_id())


if __name__ == "__main__":
    synthesizer_with_llm()

Java

import com.alibaba.dashscope.aigc.generation.Generation;
import com.alibaba.dashscope.aigc.generation.GenerationParam;
import com.alibaba.dashscope.aigc.generation.GenerationResult;
import com.alibaba.dashscope.audio.tts.SpeechSynthesisResult;
import com.alibaba.dashscope.audio.ttsv2.SpeechSynthesisAudioFormat;
import com.alibaba.dashscope.audio.ttsv2.SpeechSynthesisParam;
import com.alibaba.dashscope.audio.ttsv2.SpeechSynthesizer;
import com.alibaba.dashscope.common.Message;
import com.alibaba.dashscope.common.ResultCallback;
import com.alibaba.dashscope.common.Role;
import com.alibaba.dashscope.exception.InputRequiredException;
import com.alibaba.dashscope.exception.NoApiKeyException;
import io.reactivex.Flowable;
import java.nio.ByteBuffer;
import java.util.Arrays;
import java.util.concurrent.ConcurrentLinkedQueue;
import java.util.concurrent.atomic.AtomicBoolean;
import javax.sound.sampled.*;

public class Main {
    // Different model versions require corresponding voice versions:
    // cosyvoice-v3-flash/cosyvoice-v3-plus: Use voices such as longanyang.
    // cosyvoice-v2: Use voices such as longxiaochun_v2.
    private static String model = "cosyvoice-v3-flash";
    private static String voice = "longanyang";
    public static void process() throws NoApiKeyException, InputRequiredException {
        // Playback thread
        class PlaybackRunnable implements Runnable {
            // Set the audio format. Configure this based on your device,
            // synthesized audio parameters, and platform. Here, it is set to
            // 22050 Hz 16-bit single channel. We recommend that you choose other
            // sample rates and formats based on the model sample rate and device
            // compatibility.
            private AudioFormat af = new AudioFormat(22050, 16, 1, true, false);
            private DataLine.Info info = new DataLine.Info(SourceDataLine.class, af);
            private SourceDataLine targetSource = null;
            private AtomicBoolean runFlag = new AtomicBoolean(true);
            private ConcurrentLinkedQueue<ByteBuffer> queue =
                    new ConcurrentLinkedQueue<>();

            // Prepare the player.
            public void prepare() throws LineUnavailableException {
                targetSource = (SourceDataLine) AudioSystem.getLine(info);
                targetSource.open(af, 4096);
                targetSource.start();
            }

            public void put(ByteBuffer buffer) {
                queue.add(buffer);
            }

            // Stop playback.
            public void stop() {
                runFlag.set(false);
            }

            @Override
            public void run() {
                if (targetSource == null) {
                    return;
                }

                while (runFlag.get()) {
                    if (queue.isEmpty()) {
                        try {
                            Thread.sleep(100);
                        } catch (InterruptedException e) {
                        }
                        continue;
                    }

                    ByteBuffer buffer = queue.poll();
                    if (buffer == null) {
                        continue;
                    }

                    byte[] data = buffer.array();
                    targetSource.write(data, 0, data.length);
                }

                // Play all remaining cached audio.
                if (!queue.isEmpty()) {
                    ByteBuffer buffer = null;
                    while ((buffer = queue.poll()) != null) {
                        byte[] data = buffer.array();
                        targetSource.write(data, 0, data.length);
                    }
                }
                // Release the player.
                targetSource.drain();
                targetSource.stop();
                targetSource.close();
            }
        }

        // Create a subclass that inherits from ResultCallback<SpeechSynthesisResult>
        // to implement the callback interface.
        class ReactCallback extends ResultCallback<SpeechSynthesisResult> {
            private PlaybackRunnable playbackRunnable = null;

            public ReactCallback(PlaybackRunnable playbackRunnable) {
                this.playbackRunnable = playbackRunnable;
            }

            // Callback for when the service returns the streaming synthesis result.
            @Override
            public void onEvent(SpeechSynthesisResult result) {
                // Get the binary data of the streaming result using getAudioFrame.
                if (result.getAudioFrame() != null) {
                    // Stream the data to the player.
                    playbackRunnable.put(result.getAudioFrame());
                }
            }

            // Callback for when the service completes the synthesis.
            @Override
            public void onComplete() {
                // Notify the playback thread to end.
                playbackRunnable.stop();
            }

            // Callback for when an error occurs.
            @Override
            public void onError(Exception e) {
                // Tell the playback thread to end.
                System.out.println(e);
                playbackRunnable.stop();
            }
        }

        PlaybackRunnable playbackRunnable = new PlaybackRunnable();
        try {
            playbackRunnable.prepare();
        } catch (LineUnavailableException e) {
            throw new RuntimeException(e);
        }
        Thread playbackThread = new Thread(playbackRunnable);
        // Start the playback thread.
        playbackThread.start();
        /*******  Call the generative AI model to get streaming text *******/
        // Prepare for the LLM call.
        Generation gen = new Generation();
        Message userMsg = Message.builder()
                .role(Role.USER.getValue())
                .content("Please introduce yourself.")
                .build();
        GenerationParam genParam =
                GenerationParam.builder()
                        // If you have not configured the API Key in an environment variable, uncomment the following line and replace "apiKey" with your API Key.
                        // .apiKey("apikey")
                        .model("qwen-turbo")
                        .messages(Arrays.asList(userMsg))
                        .resultFormat(GenerationParam.ResultFormat.MESSAGE)
                        .topP(0.8)
                        .incrementalOutput(true)
                        .build();
        // Prepare the speech synthesis task.
        SpeechSynthesisParam param =
                SpeechSynthesisParam.builder()
                        // If you have not configured the API Key in an environment variable, uncomment the following line and replace "apiKey" with your API Key.
                        // .apiKey("apikey")
                        .model(model)
                        .voice(voice)
                        .format(SpeechSynthesisAudioFormat
                                .PCM_22050HZ_MONO_16BIT)
                        .build();
        SpeechSynthesizer synthesizer =
                new SpeechSynthesizer(param, new ReactCallback(playbackRunnable));
        Flowable<GenerationResult> result = gen.streamCall(genParam);
        result.blockingForEach(message -> {
            String text =
                    message.getOutput().getChoices().get(0).getMessage().getContent().trim();
            if (text != null && !text.isEmpty()) {
                System.out.println("LLM output: " + text);
                synthesizer.streamingCall(text);
            }
        });
        synthesizer.streamingComplete();
        System.out.print("Request ID: " + synthesizer.getLastRequestId());
        try {
            // Wait for the playback thread to finish playing all audio.
            playbackThread.join();
        } catch (InterruptedException e) {
            throw new RuntimeException(e);
        }
    }

    public static void main(String[] args) throws NoApiKeyException, InputRequiredException {
        process();
        System.exit(0);
    }
}