add some code

scripts/p3_tools/README.md

@@ -0,0 +1,95 @@
+# P3音频格式转换与播放工具
+
+这个目录包含两个用于处理P3格式音频文件的Python脚本：
+
+## 1. 音频转换工具 (convert_audio_to_p3.py)
+
+将普通音频文件转换为P3格式（4字节header + Opus数据包的流式结构）并进行响度标准化。
+
+### 使用方法
+
+```bash
+python convert_audio_to_p3.py <输入音频文件> <输出P3文件> [-l LUFS] [-d]
+```
+
+其中，可选选项 `-l` 用于指定响度标准化的目标响度，默认为 -16 LUFS；可选选项 `-d` 可以禁用响度标准化。
+
+如果输入的音频文件符合下面的任一条件，建议使用 `-d` 禁用响度标准化：
+- 音频过短
+- 音频已经调整过响度
+- 音频来自默认 TTS （小智当前使用的 TTS 的默认响度已是 -16 LUFS）
+
+例如：
+```bash
+python convert_audio_to_p3.py input.mp3 output.p3
+```
+
+## 2. P3音频播放工具 (play_p3.py)
+
+播放P3格式的音频文件。
+
+### 特性
+
+- 解码并播放P3格式的音频文件
+- 在播放结束或用户中断时应用淡出效果，避免破音
+- 支持通过命令行参数指定要播放的文件
+
+### 使用方法
+
+```bash
+python play_p3.py <P3文件路径>
+```
+
+例如：
+```bash
+python play_p3.py output.p3
+```
+
+## 3. 音频转回工具 (convert_p3_to_audio.py)
+
+将P3格式转换回普通音频文件。
+
+### 使用方法
+
+```bash
+python convert_p3_to_audio.py <输入P3文件> <输出音频文件>
+```
+
+输出音频文件需要有扩展名。
+
+例如：
+```bash
+python convert_p3_to_audio.py input.p3 output.wav
+```
+## 4. 音频/P3批量转换工具
+
+一个图形化的工具，支持批量转换音频到P3，P3到音频
+
+![](./img/img.png)
+
+### 使用方法：
+```bash
+python batch_convert_gui.py
+```
+
+## 依赖安装
+
+在使用这些脚本前，请确保安装了所需的Python库：
+
+```bash
+pip install librosa opuslib numpy tqdm sounddevice pyloudnorm soundfile
+```
+
+或者使用提供的requirements.txt文件：
+
+```bash
+pip install -r requirements.txt
+```
+
+## P3格式说明
+
+P3格式是一种简单的流式音频格式，结构如下：
+- 每个音频帧由一个4字节的头部和一个Opus编码的数据包组成
+- 头部格式：[1字节类型, 1字节保留, 2字节长度]
+- 采样率固定为16000Hz，单声道
+- 每帧时长为60ms 

scripts/p3_tools/batch_convert_gui.py

@@ -0,0 +1,221 @@
+import tkinter as tk
+from tkinter import ttk, filedialog, messagebox
+import os
+import threading
+import sys
+from convert_audio_to_p3 import encode_audio_to_opus
+from convert_p3_to_audio import decode_p3_to_audio
+
+class AudioConverterApp:
+    def __init__(self, master):
+        self.master = master
+        master.title("音频/P3 批量转换工具")
+        master.geometry("680x600")  # 调整窗口高度
+
+        # 初始化变量
+        self.mode = tk.StringVar(value="audio_to_p3")
+        self.output_dir = tk.StringVar()
+        self.output_dir.set(os.path.abspath("output"))
+        self.enable_loudnorm = tk.BooleanVar(value=True)
+        self.target_lufs = tk.DoubleVar(value=-16.0)
+
+        # 创建UI组件
+        self.create_widgets()
+        self.redirect_output()
+
+    def create_widgets(self):
+        # 模式选择
+        mode_frame = ttk.LabelFrame(self.master, text="转换模式")
+        mode_frame.grid(row=0, column=0, padx=10, pady=5, sticky="ew")
+        
+        ttk.Radiobutton(mode_frame, text="音频转P3", variable=self.mode,
+                        value="audio_to_p3", command=self.toggle_settings,
+                        width=12).grid(row=0, column=0, padx=5)
+        ttk.Radiobutton(mode_frame, text="P3转音频", variable=self.mode,
+                        value="p3_to_audio", command=self.toggle_settings,
+                        width=12).grid(row=0, column=1, padx=5)
+
+        # 响度设置
+        self.loudnorm_frame = ttk.Frame(self.master)
+        self.loudnorm_frame.grid(row=1, column=0, padx=10, pady=5, sticky="ew")
+        
+        ttk.Checkbutton(self.loudnorm_frame, text="启用响度调整", 
+                       variable=self.enable_loudnorm, width=15
+                       ).grid(row=0, column=0, padx=2)
+        ttk.Entry(self.loudnorm_frame, textvariable=self.target_lufs, 
+                 width=6).grid(row=0, column=1, padx=2)
+        ttk.Label(self.loudnorm_frame, text="LUFS").grid(row=0, column=2, padx=2)
+
+        # 文件选择
+        file_frame = ttk.LabelFrame(self.master, text="输入文件")
+        file_frame.grid(row=2, column=0, padx=10, pady=5, sticky="nsew")
+        
+        # 文件操作按钮
+        ttk.Button(file_frame, text="选择文件", command=self.select_files,
+                  width=12).grid(row=0, column=0, padx=5, pady=2)
+        ttk.Button(file_frame, text="移除选中", command=self.remove_selected,
+                  width=12).grid(row=0, column=1, padx=5, pady=2)
+        ttk.Button(file_frame, text="清空列表", command=self.clear_files,
+                  width=12).grid(row=0, column=2, padx=5, pady=2)
+
+        # 文件列表（使用Treeview）
+        self.tree = ttk.Treeview(file_frame, columns=("selected", "filename"), 
+                               show="headings", height=8)
+        self.tree.heading("selected", text="选中", anchor=tk.W)
+        self.tree.heading("filename", text="文件名", anchor=tk.W)
+        self.tree.column("selected", width=60, anchor=tk.W)
+        self.tree.column("filename", width=600, anchor=tk.W)
+        self.tree.grid(row=1, column=0, columnspan=3, sticky="nsew", padx=5, pady=2)
+        self.tree.bind("<ButtonRelease-1>", self.on_tree_click)
+
+        # 输出目录
+        output_frame = ttk.LabelFrame(self.master, text="输出目录")
+        output_frame.grid(row=3, column=0, padx=10, pady=5, sticky="ew")
+        
+        ttk.Entry(output_frame, textvariable=self.output_dir, width=60
+                 ).grid(row=0, column=0, padx=5, sticky="ew")
+        ttk.Button(output_frame, text="浏览", command=self.select_output_dir,
+                  width=8).grid(row=0, column=1, padx=5)
+
+        # 转换按钮区域
+        button_frame = ttk.Frame(self.master)
+        button_frame.grid(row=4, column=0, padx=10, pady=10, sticky="ew")
+        
+        ttk.Button(button_frame, text="转换全部文件", command=lambda: self.start_conversion(True),
+                  width=15).pack(side=tk.LEFT, padx=5)
+        ttk.Button(button_frame, text="转换选中文件", command=lambda: self.start_conversion(False),
+                  width=15).pack(side=tk.LEFT, padx=5)
+
+        # 日志区域
+        log_frame = ttk.LabelFrame(self.master, text="日志")
+        log_frame.grid(row=5, column=0, padx=10, pady=5, sticky="nsew")
+        
+        self.log_text = tk.Text(log_frame, height=14, width=80)
+        self.log_text.pack(fill=tk.BOTH, expand=True)
+
+        # 配置布局权重
+        self.master.columnconfigure(0, weight=1)
+        self.master.rowconfigure(2, weight=1)
+        self.master.rowconfigure(5, weight=3)
+        file_frame.columnconfigure(0, weight=1)
+        file_frame.rowconfigure(1, weight=1)
+
+    def toggle_settings(self):
+        if self.mode.get() == "audio_to_p3":
+            self.loudnorm_frame.grid()
+        else:
+            self.loudnorm_frame.grid_remove()
+
+    def select_files(self):
+        file_types = [
+            ("音频文件", "*.wav *.mp3 *.ogg *.flac") if self.mode.get() == "audio_to_p3" 
+            else ("P3文件", "*.p3")
+        ]
+        
+        files = filedialog.askopenfilenames(filetypes=file_types)
+        for f in files:
+            self.tree.insert("", tk.END, values=("[ ]", os.path.basename(f)), tags=(f,))
+
+    def on_tree_click(self, event):
+        """处理复选框点击事件"""
+        region = self.tree.identify("region", event.x, event.y)
+        if region == "cell":
+            col = self.tree.identify_column(event.x)
+            item = self.tree.identify_row(event.y)
+            if col == "#1":  # 点击的是选中列
+                current_val = self.tree.item(item, "values")[0]
+                new_val = "[√]" if current_val == "[ ]" else "[ ]"
+                self.tree.item(item, values=(new_val, self.tree.item(item, "values")[1]))
+
+    def remove_selected(self):
+        """移除选中的文件"""
+        to_remove = []
+        for item in self.tree.get_children():
+            if self.tree.item(item, "values")[0] == "[√]":
+                to_remove.append(item)
+        for item in reversed(to_remove):
+            self.tree.delete(item)
+
+    def clear_files(self):
+        """清空所有文件"""
+        for item in self.tree.get_children():
+            self.tree.delete(item)
+
+    def select_output_dir(self):
+        path = filedialog.askdirectory()
+        if path:
+            self.output_dir.set(path)
+
+    def redirect_output(self):
+        class StdoutRedirector:
+            def __init__(self, text_widget):
+                self.text_widget = text_widget
+                self.original_stdout = sys.stdout
+
+            def write(self, message):
+                self.text_widget.insert(tk.END, message)
+                self.text_widget.see(tk.END)
+                self.original_stdout.write(message)
+
+            def flush(self):
+                self.original_stdout.flush()
+
+        sys.stdout = StdoutRedirector(self.log_text)
+
+    def start_conversion(self, convert_all):
+        """开始转换"""
+        input_files = []
+        for item in self.tree.get_children():
+            if convert_all or self.tree.item(item, "values")[0] == "[√]":
+                input_files.append(self.tree.item(item, "tags")[0])
+        
+        if not input_files:
+            msg = "没有找到可转换的文件" if convert_all else "没有选中任何文件"
+            messagebox.showwarning("警告", msg)
+            return
+        
+        os.makedirs(self.output_dir.get(), exist_ok=True)
+
+        try:
+            if self.mode.get() == "audio_to_p3":
+                target_lufs = self.target_lufs.get() if self.enable_loudnorm.get() else None
+                thread = threading.Thread(target=self.convert_audio_to_p3, args=(target_lufs, input_files))
+            else:
+                thread = threading.Thread(target=self.convert_p3_to_audio, args=(input_files,))
+            
+            thread.start()
+        except Exception as e:
+            print(f"转换初始化失败: {str(e)}")
+
+    def convert_audio_to_p3(self, target_lufs, input_files):
+        """音频转P3转换逻辑"""
+        for input_path in input_files:
+            try:
+                filename = os.path.basename(input_path)
+                base_name = os.path.splitext(filename)[0]
+                output_path = os.path.join(self.output_dir.get(), f"{base_name}.p3")
+                
+                print(f"正在转换: {filename}")
+                encode_audio_to_opus(input_path, output_path, target_lufs)
+                print(f"转换成功: {filename}\n")
+            except Exception as e:
+                print(f"转换失败: {str(e)}\n")
+
+    def convert_p3_to_audio(self, input_files):
+        """P3转音频转换逻辑"""
+        for input_path in input_files:
+            try:
+                filename = os.path.basename(input_path)
+                base_name = os.path.splitext(filename)[0]
+                output_path = os.path.join(self.output_dir.get(), f"{base_name}.wav")
+                
+                print(f"正在转换: {filename}")
+                decode_p3_to_audio(input_path, output_path)
+                print(f"转换成功: {filename}\n")
+            except Exception as e:
+                print(f"转换失败: {str(e)}\n")
+
+if __name__ == "__main__":
+    root = tk.Tk()
+    app = AudioConverterApp(root)
+    root.mainloop()

scripts/p3_tools/convert_audio_to_p3.py

@@ -0,0 +1,62 @@
+# convert audio files to protocol v3 stream
+import librosa
+import opuslib
+import struct
+import sys
+import tqdm
+import numpy as np
+import argparse
+import pyloudnorm as pyln
+
+def encode_audio_to_opus(input_file, output_file, target_lufs=None):
+    # Load audio file using librosa
+    audio, sample_rate = librosa.load(input_file, sr=None, mono=False, dtype=np.float32)
+    
+    # Convert to mono if stereo
+    if audio.ndim == 2:
+        audio = librosa.to_mono(audio)
+    
+    if target_lufs is not None:
+        print("Note: Automatic loudness adjustment is enabled, which may cause", file=sys.stderr) 
+        print("      audio distortion. If the input audio has already been ", file=sys.stderr)
+        print("      loudness-adjusted or if the input audio is TTS audio, ", file=sys.stderr)
+        print("      please use the `-d` parameter to disable loudness adjustment.", file=sys.stderr)
+        meter = pyln.Meter(sample_rate)
+        current_loudness = meter.integrated_loudness(audio)
+        audio = pyln.normalize.loudness(audio, current_loudness, target_lufs)
+        print(f"Adjusted loudness: {current_loudness:.1f} LUFS -> {target_lufs} LUFS")
+
+    # Convert sample rate to 16000Hz if necessary
+    target_sample_rate = 16000
+    if sample_rate != target_sample_rate:
+        audio = librosa.resample(audio, orig_sr=sample_rate, target_sr=target_sample_rate)
+        sample_rate = target_sample_rate
+    
+    # Convert audio data back to int16 after processing
+    audio = (audio * 32767).astype(np.int16)
+    
+    # Initialize Opus encoder
+    encoder = opuslib.Encoder(sample_rate, 1, opuslib.APPLICATION_AUDIO)
+
+    # Encode and save
+    with open(output_file, 'wb') as f:
+        duration = 60  # 60ms per frame
+        frame_size = int(sample_rate * duration / 1000)
+        for i in tqdm.tqdm(range(0, len(audio) - frame_size, frame_size)):
+            frame = audio[i:i + frame_size]
+            opus_data = encoder.encode(frame.tobytes(), frame_size=frame_size)
+            packet = struct.pack('>BBH', 0, 0, len(opus_data)) + opus_data
+            f.write(packet)
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='Convert audio to Opus with loudness normalization')
+    parser.add_argument('input_file', help='Input audio file')
+    parser.add_argument('output_file', help='Output .opus file')
+    parser.add_argument('-l', '--lufs', type=float, default=-16.0,
+                       help='Target loudness in LUFS (default: -16)')
+    parser.add_argument('-d', '--disable-loudnorm', action='store_true',
+                       help='Disable loudness normalization')
+    args = parser.parse_args()
+
+    target_lufs = None if args.disable_loudnorm else args.lufs
+    encode_audio_to_opus(args.input_file, args.output_file, target_lufs)

scripts/p3_tools/convert_p3_to_audio.py

@@ -0,0 +1,51 @@
+import struct
+import sys
+import opuslib
+import numpy as np
+from tqdm import tqdm
+import soundfile as sf
+
+
+def decode_p3_to_audio(input_file, output_file):
+    sample_rate = 16000
+    channels = 1
+    decoder = opuslib.Decoder(sample_rate, channels)
+
+    pcm_frames = []
+    frame_size = int(sample_rate * 60 / 1000)
+
+    with open(input_file, "rb") as f:
+        f.seek(0, 2)
+        total_size = f.tell()
+        f.seek(0)
+
+        with tqdm(total=total_size, unit="B", unit_scale=True) as pbar:
+            while True:
+                header = f.read(4)
+                if not header or len(header) < 4:
+                    break
+
+                pkt_type, reserved, opus_len = struct.unpack(">BBH", header)
+                opus_data = f.read(opus_len)
+                if len(opus_data) != opus_len:
+                    break
+
+                pcm = decoder.decode(opus_data, frame_size)
+                pcm_frames.append(np.frombuffer(pcm, dtype=np.int16))
+
+                pbar.update(4 + opus_len)
+
+    if not pcm_frames:
+        raise ValueError("No valid audio data found")
+
+    pcm_data = np.concatenate(pcm_frames)
+
+    sf.write(output_file, pcm_data, sample_rate, subtype="PCM_16")
+
+
+if __name__ == "__main__":
+    if len(sys.argv) != 3:
+        print("Usage: python convert_p3_to_audio.py <input.p3> <output.wav>")
+        sys.exit(1)
+
+    decode_p3_to_audio(sys.argv[1], sys.argv[2])

scripts/p3_tools/p3_gui_player.py

@@ -0,0 +1,241 @@
+import tkinter as tk
+from tkinter import filedialog, messagebox
+import threading
+import time
+import opuslib
+import struct
+import numpy as np
+import sounddevice as sd
+import os
+
+
+def play_p3_file(input_file, stop_event=None, pause_event=None):
+    """
+    播放p3格式的音频文件
+    p3格式: [1字节类型, 1字节保留, 2字节长度, Opus数据]
+    """
+    # 初始化Opus解码器
+    sample_rate = 16000  # 采样率固定为16000Hz
+    channels = 1  # 单声道
+    decoder = opuslib.Decoder(sample_rate, channels)
+    
+    # 帧大小 (60ms)
+    frame_size = int(sample_rate * 60 / 1000)
+    
+    # 打开音频流
+    stream = sd.OutputStream(
+        samplerate=sample_rate,
+        channels=channels,
+        dtype='int16'
+    )
+    stream.start()
+    
+    try:
+        with open(input_file, 'rb') as f:
+            print(f"正在播放: {input_file}")
+            
+            while True:
+                if stop_event and stop_event.is_set():
+                    break
+
+                if pause_event and pause_event.is_set():
+                    time.sleep(0.1)
+                    continue
+
+                # 读取头部 (4字节)
+                header = f.read(4)
+                if not header or len(header) < 4:
+                    break
+                
+                # 解析头部
+                packet_type, reserved, data_len = struct.unpack('>BBH', header)
+                
+                # 读取Opus数据
+                opus_data = f.read(data_len)
+                if not opus_data or len(opus_data) < data_len:
+                    break
+                
+                # 解码Opus数据
+                pcm_data = decoder.decode(opus_data, frame_size)
+                
+                # 将字节转换为numpy数组
+                audio_array = np.frombuffer(pcm_data, dtype=np.int16)
+                
+                # 播放音频
+                stream.write(audio_array)
+                
+    except KeyboardInterrupt:
+        print("\n播放已停止")
+    finally:
+        stream.stop()
+        stream.close()
+        print("播放完成")
+
+
+class P3PlayerApp:
+    def __init__(self, root):
+        self.root = root
+        self.root.title("P3 文件简易播放器")
+        self.root.geometry("500x400")
+
+        self.playlist = []
+        self.current_index = 0
+        self.is_playing = False
+        self.is_paused = False
+        self.stop_event = threading.Event()
+        self.pause_event = threading.Event()
+        self.loop_playback = tk.BooleanVar(value=False)  # 循环播放复选框的状态
+
+        # 创建界面组件
+        self.create_widgets()
+
+    def create_widgets(self):
+        # 播放列表
+        self.playlist_label = tk.Label(self.root, text="播放列表:")
+        self.playlist_label.pack(pady=5)
+
+        self.playlist_frame = tk.Frame(self.root)
+        self.playlist_frame.pack(fill=tk.BOTH, expand=True, padx=10, pady=5)
+
+        self.playlist_listbox = tk.Listbox(self.playlist_frame, selectmode=tk.SINGLE)
+        self.playlist_listbox.pack(fill=tk.BOTH, expand=True)
+
+        # 复选框和移除按钮
+        self.checkbox_frame = tk.Frame(self.root)
+        self.checkbox_frame.pack(pady=5)
+
+        self.remove_button = tk.Button(self.checkbox_frame, text="移除文件", command=self.remove_files)
+        self.remove_button.pack(side=tk.LEFT, padx=5)
+
+        # 循环播放复选框
+        self.loop_checkbox = tk.Checkbutton(self.checkbox_frame, text="循环播放", variable=self.loop_playback)
+        self.loop_checkbox.pack(side=tk.LEFT, padx=5)
+
+        # 控制按钮
+        self.control_frame = tk.Frame(self.root)
+        self.control_frame.pack(pady=10)
+
+        self.add_button = tk.Button(self.control_frame, text="添加文件", command=self.add_file)
+        self.add_button.grid(row=0, column=0, padx=5)
+
+        self.play_button = tk.Button(self.control_frame, text="播放", command=self.play)
+        self.play_button.grid(row=0, column=1, padx=5)
+
+        self.pause_button = tk.Button(self.control_frame, text="暂停", command=self.pause)
+        self.pause_button.grid(row=0, column=2, padx=5)
+
+        self.stop_button = tk.Button(self.control_frame, text="停止", command=self.stop)
+        self.stop_button.grid(row=0, column=3, padx=5)
+
+        # 状态标签
+        self.status_label = tk.Label(self.root, text="未在播放", fg="blue")
+        self.status_label.pack(pady=10)
+
+    def add_file(self):
+        files = filedialog.askopenfilenames(filetypes=[("P3 文件", "*.p3")])
+        if files:
+            self.playlist.extend(files)
+            self.update_playlist()
+
+    def update_playlist(self):
+        self.playlist_listbox.delete(0, tk.END)
+        for file in self.playlist:
+            self.playlist_listbox.insert(tk.END, os.path.basename(file))  # 仅显示文件名
+
+    def update_status(self, status_text, color="blue"):
+        """更新状态标签的内容"""
+        self.status_label.config(text=status_text, fg=color)
+
+    def play(self):
+        if not self.playlist:
+            messagebox.showwarning("警告", "播放列表为空！")
+            return
+
+        if self.is_paused:
+            self.is_paused = False
+            self.pause_event.clear()
+            self.update_status(f"正在播放：{os.path.basename(self.playlist[self.current_index])}", "green")
+            return
+
+        if self.is_playing:
+            return
+
+        self.is_playing = True
+        self.stop_event.clear()
+        self.pause_event.clear()
+        self.current_index = self.playlist_listbox.curselection()[0] if self.playlist_listbox.curselection() else 0
+        self.play_thread = threading.Thread(target=self.play_audio, daemon=True)
+        self.play_thread.start()
+        self.update_status(f"正在播放：{os.path.basename(self.playlist[self.current_index])}", "green")
+
+    def play_audio(self):
+        while True:
+            if self.stop_event.is_set():
+                break
+
+            if self.pause_event.is_set():
+                time.sleep(0.1)
+                continue
+
+            # 检查当前索引是否有效
+            if self.current_index >= len(self.playlist):
+                if self.loop_playback.get():  # 如果勾选了循环播放
+                    self.current_index = 0  # 回到第一首
+                else:
+                    break  # 否则停止播放
+
+            file = self.playlist[self.current_index]
+            self.playlist_listbox.selection_clear(0, tk.END)
+            self.playlist_listbox.selection_set(self.current_index)
+            self.playlist_listbox.activate(self.current_index)
+            self.update_status(f"正在播放：{os.path.basename(self.playlist[self.current_index])}", "green")
+            play_p3_file(file, self.stop_event, self.pause_event)
+
+            if self.stop_event.is_set():
+                break
+
+            if not self.loop_playback.get():  # 如果没有勾选循环播放
+                break  # 播放完当前文件后停止
+
+            self.current_index += 1
+            if self.current_index >= len(self.playlist):
+                if self.loop_playback.get():  # 如果勾选了循环播放
+                    self.current_index = 0  # 回到第一首
+
+        self.is_playing = False
+        self.is_paused = False
+        self.update_status("播放已停止", "red")
+
+    def pause(self):
+        if self.is_playing:
+            self.is_paused = not self.is_paused
+            if self.is_paused:
+                self.pause_event.set()
+                self.update_status("播放已暂停", "orange")
+            else:
+                self.pause_event.clear()
+                self.update_status(f"正在播放：{os.path.basename(self.playlist[self.current_index])}", "green")
+
+    def stop(self):
+        if self.is_playing or self.is_paused:
+            self.is_playing = False
+            self.is_paused = False
+            self.stop_event.set()
+            self.pause_event.clear()
+            self.update_status("播放已停止", "red")
+
+    def remove_files(self):
+        selected_indices = self.playlist_listbox.curselection()
+        if not selected_indices:
+            messagebox.showwarning("警告", "请先选择要移除的文件！")
+            return
+
+        for index in reversed(selected_indices):
+            self.playlist.pop(index)
+        self.update_playlist()
+
+
+if __name__ == "__main__":
+    root = tk.Tk()
+    app = P3PlayerApp(root)
+    root.mainloop()

scripts/p3_tools/play_p3.py

@@ -0,0 +1,71 @@
+# 播放p3格式的音频文件
+import opuslib
+import struct
+import numpy as np
+import sounddevice as sd
+import argparse
+
+def play_p3_file(input_file):
+    """
+    播放p3格式的音频文件
+    p3格式: [1字节类型, 1字节保留, 2字节长度, Opus数据]
+    """
+    # 初始化Opus解码器
+    sample_rate = 16000  # 采样率固定为16000Hz
+    channels = 1  # 单声道
+    decoder = opuslib.Decoder(sample_rate, channels)
+    
+    # 帧大小 (60ms)
+    frame_size = int(sample_rate * 60 / 1000)
+    
+    # 打开音频流
+    stream = sd.OutputStream(
+        samplerate=sample_rate,
+        channels=channels,
+        dtype='int16'
+    )
+    stream.start()
+    
+    try:
+        with open(input_file, 'rb') as f:
+            print(f"正在播放: {input_file}")
+            
+            while True:
+                # 读取头部 (4字节)
+                header = f.read(4)
+                if not header or len(header) < 4:
+                    break
+                
+                # 解析头部
+                packet_type, reserved, data_len = struct.unpack('>BBH', header)
+                
+                # 读取Opus数据
+                opus_data = f.read(data_len)
+                if not opus_data or len(opus_data) < data_len:
+                    break
+                
+                # 解码Opus数据
+                pcm_data = decoder.decode(opus_data, frame_size)
+                
+                # 将字节转换为numpy数组
+                audio_array = np.frombuffer(pcm_data, dtype=np.int16)
+                
+                # 播放音频
+                stream.write(audio_array)
+                
+    except KeyboardInterrupt:
+        print("\n播放已停止")
+    finally:
+        stream.stop()
+        stream.close()
+        print("播放完成")
+
+def main():
+    parser = argparse.ArgumentParser(description='播放p3格式的音频文件')
+    parser.add_argument('input_file', help='输入的p3文件路径')
+    args = parser.parse_args()
+    
+    play_p3_file(args.input_file)
+
+if __name__ == "__main__":
+    main() 

scripts/p3_tools/requirements.txt

@@ -0,0 +1,7 @@
+librosa>=0.9.2
+opuslib>=3.0.1
+numpy>=1.20.0
+tqdm>=4.62.0
+sounddevice>=0.4.4 
+pyloudnorm>=0.1.1
+soundfile>=0.13.1