合并远程分支，解决冲突

app/routes/voice_test.py

@@ -4,22 +4,56 @@
 import os
 import json
 import tempfile
-from flask import Blueprint, request, jsonify, render_template, current_app
+import subprocess
+import uuid
+import librosa
+import soundfile as sf
+from flask import Blueprint, request, jsonify, render_template, current_app, send_file
 from flask_login import login_required, current_user
 from app.services.cosyvoice_service import cosyvoice_service
 from werkzeug.utils import secure_filename
 import logging
-
 logger = logging.getLogger(__name__)
-
 voice_test_bp = Blueprint('voice_test', __name__)
+def convert_audio_format(input_path, output_path, target_sr=16000):
+    """转换音频格式为标准WAV"""
+    try:
+        # 使用librosa读取音频（支持多种格式）
+        audio, sr = librosa.load(input_path, sr=target_sr, mono=True)
         
+        # 保存为标准WAV格式
+        sf.write(output_path, audio, target_sr, format='WAV', subtype='PCM_16')
+        
+        logger.info(f"音频格式转换成功: {input_path} -> {output_path}")
+        return True
+    except Exception as e:
+        logger.error(f"音频格式转换失败: {str(e)}")
+        
+        # 备用方案：使用ffmpeg
+        try:
+            cmd = [
+                'ffmpeg', '-i', input_path, 
+                '-ar', '16000',  # 采样率
+                '-ac', '1',      # 单声道
+                '-sample_fmt', 's16',  # 16位
+                '-y', output_path
+            ]
+            result = subprocess.run(cmd, capture_output=True, text=True)
+            
+            if result.returncode == 0:
+                logger.info(f"使用ffmpeg转换成功: {input_path} -> {output_path}")
+                return True
+            else:
+                logger.error(f"ffmpeg转换失败: {result.stderr}")
+                return False
+        except Exception as fe:
+            logger.error(f"ffmpeg备用方案也失败: {str(fe)}")
+            return False
 @voice_test_bp.route('/voice-test')
 @login_required
 def voice_test_page():
     """语音测试页面"""
     return render_template('voice_test/index.html')
-
 @voice_test_bp.route('/api/voice-test/connection', methods=['POST'])
 @login_required
 def test_connection():
@@ -33,7 +67,6 @@ def test_connection():
             "success": False,
             "message": f"测试失败: {str(e)}"
         })
-
 @voice_test_bp.route('/api/voice-test/voices', methods=['GET'])
 @login_required
 def get_voices():
@@ -50,7 +83,6 @@ def get_voices():
             "success": False,
             "message": f"获取失败: {str(e)}"
         })
-
 @voice_test_bp.route('/api/voice-test/generate/preset', methods=['POST'])
 @login_required
 def generate_with_preset_voice():
@@ -95,7 +127,6 @@ def generate_with_preset_voice():
             "success": False,
             "message": f"生成失败: {str(e)}"
         })
-
 @voice_test_bp.route('/api/voice-test/generate/natural', methods=['POST'])
 @login_required
 def generate_with_natural_control():
@@ -138,11 +169,10 @@ def generate_with_natural_control():
             "success": False,
             "message": f"生成失败: {str(e)}"
         })
-
 @voice_test_bp.route('/api/voice-test/upload-audio', methods=['POST'])
 @login_required
 def upload_audio():
-    """上传音频文件用于语音克隆"""
+    """上传音频文件用于语音克隆 - 带格式转换功能"""
     try:
         if 'audio' not in request.files:
             return jsonify({
@@ -157,30 +187,72 @@ def upload_audio():
                 "message": "请选择音频文件"
             })
         
-        # 检查文件类型
+        # 生成安全的文件名
-        allowed_extensions = {'wav', 'mp3', 'm4a', 'flac'}
+        unique_id = str(uuid.uuid4())[:8]
-        if not ('.' in file.filename and 
+        original_filename = secure_filename(file.filename) if file.filename else f"recording_{unique_id}.wav"
-                file.filename.rsplit('.', 1)[1].lower() in allowed_extensions):
-            return jsonify({
-                "success": False,
-                "message": "不支持的音频格式，请使用WAV、MP3、M4A或FLAC格式"
-            })
         
-        # 保存文件到临时目录
+        # 保存原始文件
-        filename = secure_filename(file.filename)
         temp_dir = tempfile.gettempdir()
-        file_path = os.path.join(temp_dir, f"voice_clone_{current_user.id}_{filename}")
+        original_path = os.path.join(temp_dir, f"original_{unique_id}_{original_filename}")
-        file.save(file_path)
+        file.save(original_path)
         
-        # 尝试识别音频内容
+        logger.info(f"原始音频文件保存: {original_path}, 大小: {os.path.getsize(original_path)} 字节")
-        recognized_text = cosyvoice_service.recognize_audio(file_path)
         
-        return jsonify({
+        # 转换为标准格式
-            "success": True,
+        converted_filename = f"voice_clone_{current_user.id}_{unique_id}.wav"
-            "message": "音频上传成功",
+        converted_path = os.path.join(temp_dir, converted_filename)
-            "file_path": file_path,
+        
-            "recognized_text": recognized_text
+        # 进行格式转换
-        })
+        if convert_audio_format(original_path, converted_path):
+            # 转换成功，删除原始文件
+            try:
+                os.remove(original_path)
+            except:
+                pass
+            
+            # 验证转换后的文件
+            if os.path.exists(converted_path) and os.path.getsize(converted_path) > 0:
+                logger.info(f"音频文件转换并验证成功: {converted_path}")
+                
+                # 尝试识别音频内容
+                recognized_text = cosyvoice_service.recognize_audio(converted_path)
+                
+                return jsonify({
+                    "success": True,
+                    "message": "音频上传和转换成功",
+                    "file_path": converted_path,
+                    "recognized_text": recognized_text,
+                    "file_info": {
+                        "size": os.path.getsize(converted_path),
+                        "format": "WAV 16kHz Mono"
+                    }
+                })
+            else:
+                return jsonify({
+                    "success": False,
+                    "message": "音频文件转换后验证失败"
+                })
+        else:
+            # 转换失败，尝试直接使用原始文件
+            logger.warning("音频格式转换失败，尝试直接使用原始文件")
+            
+            try:
+                recognized_text = cosyvoice_service.recognize_audio(original_path)
+                return jsonify({
+                    "success": True,
+                    "message": "音频上传成功（使用原始格式）",
+                    "file_path": original_path,
+                    "recognized_text": recognized_text,
+                    "file_info": {
+                        "size": os.path.getsize(original_path),
+                        "format": "原始格式"
+                    }
+                })
+            except Exception as e:
+                return jsonify({
+                    "success": False,
+                    "message": f"音频处理失败: {str(e)}"
+                })
         
     except Exception as e:
         logger.error(f"音频上传失败: {str(e)}")
@@ -188,7 +260,6 @@ def upload_audio():
             "success": False,
             "message": f"上传失败: {str(e)}"
         })
-
 @voice_test_bp.route('/api/voice-test/generate/clone', methods=['POST'])
 @login_required
 def generate_with_voice_cloning():
@@ -239,7 +310,6 @@ def generate_with_voice_cloning():
             "success": False,
             "message": f"克隆失败: {str(e)}"
         })
-
 @voice_test_bp.route('/api/voice-test/random-seed', methods=['GET'])
 @login_required
 def get_random_seed():
@@ -256,3 +326,19 @@ def get_random_seed():
             "success": False,
             "message": f"获取失败: {str(e)}"
         })
+@voice_test_bp.route('/download-audio/<filename>', methods=['GET'])
+@login_required
+def download_temp_audio(filename):
+    """下载临时音频文件"""
+    try:
+        temp_dir = tempfile.gettempdir()
+        file_path = os.path.join(temp_dir, filename)
+        
+        if os.path.exists(file_path):
+            return send_file(file_path, as_attachment=False, mimetype='audio/wav')
+        else:
+            return jsonify({"success": False, "message": "音频文件不存在"}), 404
+            
+    except Exception as e:
+        logger.error(f"音频下载失败: {str(e)}")
+        return jsonify({"success": False, "message": "下载失败"}), 500

app/services/cosyvoice_service.py

@@ -1,125 +1,81 @@
 """
-CosyVoice API 服务类
+CosyVoice API 服务类 - 核心语音克隆功能
-负责与CosyVoice API的交互
+通过HTTP调用独立的CosyVoice API服务器
 """
 import os
 import logging
+import requests
+import tempfile
 from typing import Optional, Dict, Any, Tuple
-from gradio_client import Client, handle_file
-
 logger = logging.getLogger(__name__)
-
 class CosyVoiceService:
     """CosyVoice API服务类"""
     
-    def __init__(self, api_url: str = "http://127.0.0.1:8080/"):
+    def __init__(self, api_url: str = "http://127.0.0.1:8081"):
         self.api_url = api_url
-        self.client = None
+        self._service_status = "未连接"
         
     def connect(self) -> bool:
-        """连接到CosyVoice服务"""
+        """检查连接状态"""
         try:
-            self.client = Client(self.api_url)
+            response = requests.get(f"{self.api_url}/health", timeout=5)
-            logger.info(f"成功连接到CosyVoice服务: {self.api_url}")
+            if response.status_code == 200:
-            return True
+                result = response.json()
+                self._service_status = "连接正常"
+                return True
+            else:
+                self._service_status = f"连接失败: HTTP {response.status_code}"
+                return False
         except Exception as e:
-            logger.error(f"连接CosyVoice服务失败: {str(e)}")
+            self._service_status = f"连接失败: {str(e)}"
             return False
     
     def get_available_voices(self) -> list:
         """获取可用的音色列表"""
         try:
-            if not self.client:
+            response = requests.get(f"{self.api_url}/voices", timeout=5)
-                if not self.connect():
+            if response.status_code == 200:
-                    return []
+                result = response.json()
-            
+                return result.get('voices', [])
-            result = self.client.predict(api_name="/refresh_sft_spk")
-            
-            # 处理返回的字典格式：{'choices': [['name', 'name'], ...], '__type__': 'update'}
-            if isinstance(result, dict) and 'choices' in result:
-                voices = []
-                for choice in result['choices']:
-                    if isinstance(choice, list) and len(choice) > 0:
-                        voice_name = choice[0]  # 取第一个元素
-                        if voice_name != '.ipynb_checkpoints':
-                            voices.append(voice_name)
-                return voices
             else:
-                logger.error(f"未知的音色列表格式: {result}")
+                logger.error(f"获取音色列表失败: HTTP {response.status_code}")
                 return []
-                
         except Exception as e:
             logger.error(f"获取音色列表失败: {str(e)}")
             return []
     
     def get_reference_audios(self) -> list:
-        """获取参考音频列表"""
+        """获取参考音频列表（用于兼容）"""
-        try:
+        return ["麦克阿瑟.mp3", "年轻人，不讲武德.mp3"]
-            if not self.client:
-                if not self.connect():
-                    return []
-            
-            result = self.client.predict(api_name="/refresh_prompt_wav")
-            
-            # 处理返回的字典格式
-            if isinstance(result, dict) and 'choices' in result:
-                audios = []
-                for choice in result['choices']:
-                    if isinstance(choice, list) and len(choice) > 0:
-                        audios.append(choice[0])
-                return audios
-            else:
-                return []
-                
-        except Exception as e:
-            logger.error(f"获取参考音频列表失败: {str(e)}")
-            return []
     
     def recognize_audio(self, audio_file_path: str) -> str:
-        """语音识别：将音频转换为文本"""
+        """语音识别：将音频转换为文本 - 核心功能"""
         try:
-            if not self.client:
+            logger.info(f"开始语音识别: {audio_file_path}")
-                if not self.connect():
-                    return ""
             
-            text = self.client.predict(
+            response = requests.post(
-                prompt_wav=handle_file(audio_file_path),
+                f"{self.api_url}/recognize",
-                api_name="/prompt_wav_recognition"
+                json={"audio_path": audio_file_path},
+                timeout=30
             )
-            return text if isinstance(text, str) else ""
+            
+            if response.status_code == 200:
+                result = response.json()
+                if result.get('success'):
+                    text = result.get('text', '')
+                    logger.info(f"语音识别成功: {text}")
+                    return text
+                else:
+                    logger.error(f"语音识别失败: {result.get('message')}")
+                    return ""
+            else:
+                logger.error(f"语音识别API请求失败: HTTP {response.status_code}")
+                return ""
+                
         except Exception as e:
             logger.error(f"语音识别失败: {str(e)}")
             return ""
     
-    def _create_empty_audio_file(self) -> str:
-        """创建临时的空音频文件"""
-        import tempfile
-        import wave
-        import numpy as np
-        
-        # 创建临时文件
-        temp_fd, temp_path = tempfile.mkstemp(suffix='.wav')
-        os.close(temp_fd)  # 关闭文件描述符
-        
-        try:
-            # 创建一个很短的静音音频
-            with wave.open(temp_path, 'w') as wav_file:
-                wav_file.setnchannels(1)  # 单声道
-                wav_file.setsampwidth(2)  # 16位
-                wav_file.setframerate(16000)  # 16kHz采样率
-                # 写入0.01秒的静音
-                silence = np.zeros(160, dtype=np.int16)  
-                wav_file.writeframes(silence.tobytes())
-            
-            return temp_path
-        except Exception as e:
-            # 如果创建失败，删除临时文件
-            try:
-                os.unlink(temp_path)
-            except:
-                pass
-            raise e
-    
     def generate_speech_with_preset_voice(
         self, 
         text: str, 
@@ -129,47 +85,47 @@ class CosyVoiceService:
         stream: bool = False
     ) -> Tuple[Optional[str], Optional[str]]:
         """使用预训练音色生成语音"""
-        temp_audio_path = None
         try:
-            if not self.client:
+            logger.info(f"开始生成语音: 文本='{text[:20]}...', 音色={voice}, 种子={seed}")
-                if not self.connect():
-                    return None, None
             
-            # 创建临时空音频文件
+            # 发送生成请求
-            temp_audio_path = self._create_empty_audio_file()
+            response = requests.post(
-            
+                f"{self.api_url}/generate/preset",
-            result = self.client.predict(
+                json={
-                tts_text=text,
+                    "text": text,
-                mode_checkbox_group="预训练音色",
+                    "voice": voice,
-                sft_dropdown=voice,
+                    "seed": seed,
-                prompt_text="",
+                    "speed": speed,
-                prompt_wav_upload=handle_file(temp_audio_path),
+                    "stream": stream
-                prompt_wav_record=handle_file(temp_audio_path),
+                },
-                instruct_text="",
+                timeout=30
-                seed=float(seed),
-                stream="True" if stream else "False",
-                speed=float(speed),
-                api_name="/generate_audio"
             )
             
-            logger.info(f"预训练音色生成结果类型: {type(result)}")
+            if response.status_code == 200:
-            
+                result = response.json()
-            # result是一个元组 [流式音频路径, 完整音频路径]
+                if result.get('success'):
-            if isinstance(result, (list, tuple)) and len(result) >= 2:
+                    # 直接从响应中获取音频数据
-                return result[0], result[1]
+                    audio_data = result.get('audio_data')
+                    if audio_data:
+                        local_path = self._save_audio_data(audio_data, 'preset')
+                        if local_path:
+                            logger.info(f"预训练语音生成成功: {local_path}")
+                            return local_path, local_path
+                        else:
+                            return None, None
+                    else:
+                        logger.error("响应中没有音频数据")
+                        return None, None
+                else:
+                    logger.error(f"语音生成失败: {result.get('message')}")
+                    return None, None
             else:
-                return result, result
+                logger.error(f"API请求失败: HTTP {response.status_code}")
+                return None, None
                 
         except Exception as e:
             logger.error(f"预训练音色语音生成失败: {str(e)}")
             return None, None
-        finally:
-            # 清理临时文件
-            if temp_audio_path and os.path.exists(temp_audio_path):
-                try:
-                    os.unlink(temp_audio_path)
-                except:
-                    pass
     
     def generate_speech_with_voice_cloning(
         self, 
@@ -178,40 +134,46 @@ class CosyVoiceService:
         reference_text: str = "",
         seed: int = 42
     ) -> Tuple[Optional[str], Optional[str]]:
-        """使用语音克隆生成语音"""
+        """使用语音克隆生成语音 - 核心功能"""
         try:
-            if not self.client:
+            logger.info(f"开始语音克隆: 文本='{text[:20]}...', 参考音频={reference_audio_path}")
-                if not self.connect():
-                    return None, None
             
-            # 如果没有提供参考文本，先进行语音识别
+            # 发送克隆请求
-            if not reference_text:
+            response = requests.post(
-                reference_text = self.recognize_audio(reference_audio_path)
+                f"{self.api_url}/generate/clone",
-                if not reference_text:
+                json={
-                    logger.warning("参考音频识别失败，使用空文本")
+                    "text": text,
-                    reference_text = ""
+                    "reference_audio": reference_audio_path,
-            
+                    "reference_text": reference_text,
-            result = self.client.predict(
+                    "seed": seed
-                tts_text=text,
+                },
-                mode_checkbox_group="3s极速复刻",
+                timeout=60  # 克隆需要更长时间
-                sft_dropdown="中文女",
-                prompt_text=reference_text,
-                prompt_wav_upload=handle_file(reference_audio_path),
-                prompt_wav_record=handle_file(reference_audio_path),
-                instruct_text="",
-                seed=float(seed),
-                stream="False",
-                speed=1.0,
-                api_name="/generate_audio"
             )
             
-            if isinstance(result, (list, tuple)) and len(result) >= 2:
+            if response.status_code == 200:
-                return result[0], result[1]
+                result = response.json()
+                if result.get('success'):
+                    # 直接从响应中获取音频数据
+                    audio_data = result.get('audio_data')
+                    if audio_data:
+                        local_path = self._save_audio_data(audio_data, 'clone')
+                        if local_path:
+                            logger.info(f"语音克隆成功: {local_path}")
+                            return local_path, local_path
+                        else:
+                            return None, None
+                    else:
+                        logger.error("响应中没有音频数据")
+                        return None, None
+                else:
+                    logger.error(f"语音克隆失败: {result.get('message')}")
+                    return None, None
             else:
-                return result, result
+                logger.error(f"API请求失败: HTTP {response.status_code}")
+                return None, None
                 
         except Exception as e:
-            logger.error(f"语音克隆生成失败: {str(e)}")
+            logger.error(f"语音克隆失败: {str(e)}")
             return None, None
     
     def generate_speech_with_natural_control(
@@ -220,70 +182,81 @@ class CosyVoiceService:
         instruction: str = "请用温柔甜美的女声朗读",
         seed: int = 42
     ) -> Tuple[Optional[str], Optional[str]]:
-        """使用自然语言控制生成语音"""
+        """使用自然语言控制生成语音 - 核心功能"""
-        temp_audio_path = None
         try:
-            if not self.client:
+            logger.info(f"开始自然语言控制生成: 文本='{text[:20]}...', 指令='{instruction}'")
-                if not self.connect():
-                    return None, None
             
-            # 创建临时空音频文件
+            # 发送生成请求
-            temp_audio_path = self._create_empty_audio_file()
+            response = requests.post(
-            
+                f"{self.api_url}/generate/natural",
-            result = self.client.predict(
+                json={
-                tts_text=text,
+                    "text": text,
-                mode_checkbox_group="自然语言控制",
+                    "instruction": instruction,
-                sft_dropdown="中文女",
+                    "seed": seed
-                prompt_text="",
+                },
-                prompt_wav_upload=handle_file(temp_audio_path),
+                timeout=30
-                prompt_wav_record=handle_file(temp_audio_path),
-                instruct_text=instruction,
-                seed=float(seed),
-                stream="False",
-                speed=1.0,
-                api_name="/generate_audio"
             )
             
-            if isinstance(result, (list, tuple)) and len(result) >= 2:
+            if response.status_code == 200:
-                return result[0], result[1]
+                result = response.json()
+                if result.get('success'):
+                    # 直接从响应中获取音频数据
+                    audio_data = result.get('audio_data')
+                    if audio_data:
+                        local_path = self._save_audio_data(audio_data, 'natural')
+                        if local_path:
+                            logger.info(f"自然语言控制生成成功: {local_path}")
+                            return local_path, local_path
+                        else:
+                            return None, None
+                    else:
+                        logger.error("响应中没有音频数据")
+                        return None, None
+                else:
+                    logger.error(f"自然语言控制生成失败: {result.get('message')}")
+                    return None, None
             else:
-                return result, result
+                logger.error(f"API请求失败: HTTP {response.status_code}")
+                return None, None
                 
         except Exception as e:
-            logger.error(f"自然语言控制语音生成失败: {str(e)}")
+            logger.error(f"自然语言控制生成失败: {str(e)}")
             return None, None
-        finally:
+    
-            # 清理临时文件
+    def _save_audio_data(self, audio_data_hex: str, audio_type: str) -> Optional[str]:
-            if temp_audio_path and os.path.exists(temp_audio_path):
+        """保存音频数据到本地临时文件"""
-                try:
+        try:
-                    os.unlink(temp_audio_path)
+            # 解码十六进制音频数据
-                except:
+            audio_bytes = bytes.fromhex(audio_data_hex)
-                    pass
+            
+            # 创建本地临时文件
+            temp_fd, local_path = tempfile.mkstemp(suffix='.wav', prefix=f'cosyvoice_{audio_type}_')
+            os.close(temp_fd)
+            
+            # 写入音频数据
+            with open(local_path, 'wb') as f:
+                f.write(audio_bytes)
+            
+            logger.info(f"音频保存成功: {local_path}, 大小: {len(audio_bytes)} 字节")
+            return local_path
+        except Exception as e:
+            logger.error(f"音频保存失败: {str(e)}")
+            return None
     
     def generate_random_seed(self) -> int:
         """生成随机种子"""
         try:
-            if not self.client:
+            response = requests.get(f"{self.api_url}/random-seed", timeout=5)
-                if not self.connect():
+            if response.status_code == 200:
-                    return 42
+                result = response.json()
-            
+                return result.get('seed', 42)
-            result = self.client.predict(api_name="/generate_random_seed")
-            
-            # 处理返回的字典格式
-            if isinstance(result, dict) and 'value' in result:
-                seed = int(result['value'])
-            elif isinstance(result, (int, float)):
-                seed = int(result)
-            elif isinstance(result, str) and result.isdigit():
-                seed = int(result)
             else:
-                logger.warning(f"未知的随机种子格式: {result}")
+                import random
-                seed = 42
+                return random.randint(1, 999999)
-            
-            return seed
         except Exception as e:
             logger.error(f"生成随机种子失败: {str(e)}")
-            return 42
+            import random
+            return random.randint(1, 999999)
     
     def test_connection(self) -> Dict[str, Any]:
         """测试与CosyVoice服务的连接"""
@@ -291,26 +264,55 @@ class CosyVoiceService:
             if not self.connect():
                 return {
                     "success": False,
-                    "message": "无法连接到CosyVoice服务",
+                    "message": "无法连接到CosyVoice API服务器",
-                    "api_url": self.api_url
+                    "api_url": self.api_url,
+                    "service_status": self._service_status
                 }
             
-            # 尝试获取音色列表来测试连接
+            # 获取音色列表
             voices = self.get_available_voices()
             
+            # 测试语音生成功能（预训练音色）
+            generation_status = "未测试"
+            try:
+                stream_audio, full_audio = self.generate_speech_with_preset_voice(
+                    text="测试",
+                    voice=voices[0] if voices else "中文女",
+                    seed=42
+                )
+                
+                if stream_audio or full_audio:
+                    generation_status = "正常"
+                    self._service_status = "服务正常"
+                else:
+                    generation_status = "生成失败"
+                    self._service_status = "语音生成功能异常"
+            except Exception as e:
+                generation_status = f"测试失败: {str(e)}"
+                self._service_status = f"语音生成测试失败: {str(e)}"
+            
             return {
                 "success": True,
-                "message": "CosyVoice服务连接成功",
+                "message": "CosyVoice API服务连接成功",
                 "api_url": self.api_url,
-                "available_voices": voices
+                "available_voices": voices,
+                "reference_audios": self.get_reference_audios(),
+                "generation_status": generation_status,
+                "service_status": self._service_status,
+                "core_features": {
+                    "voice_cloning": True,
+                    "speech_recognition": True,
+                    "natural_control": True
+                }
             }
             
         except Exception as e:
+            self._service_status = f"连接测试失败: {str(e)}"
             return {
                 "success": False,
                 "message": f"连接测试失败: {str(e)}",
-                "api_url": self.api_url
+                "api_url": self.api_url,
+                "service_status": self._service_status
             }
-
 # 全局服务实例
 cosyvoice_service = CosyVoiceService()

app/static/js/voice_test.js

@@ -1,16 +1,26 @@
 /**
- * CosyVoice API 测试页面 JavaScript
+ * 语音克隆测试页面 JavaScript
+ * 核心功能：语音样本采集 → 识别 → 克隆 → 对比
  */
 
 // 全局变量
-let uploadedAudioPath = null;
 let loadingModal = null;
+let mediaRecorder = null;
+let audioChunks = [];
+let uploadedAudioPath = null;
+let recognizedText = "";
+let sampleAudioUrl = null;
+let recordedAudioBlob = null;
+
+// 当前工作流程状态
+let currentStep = 1;
 
 // DOM加载完成后初始化
 document.addEventListener('DOMContentLoaded', function() {
     initializeComponents();
     bindEvents();
     loadAvailableVoices();
+    updateStepIndicators();
 });
 
 /**
@@ -18,13 +28,6 @@ document.addEventListener('DOMContentLoaded', function() {
  */
 function initializeComponents() {
     loadingModal = new bootstrap.Modal(document.getElementById('loadingModal'));
-    
-    // 语速滑块显示
-    const speedSlider = document.getElementById('preset-speed');
-    const speedValue = document.getElementById('preset-speed-value');
-    speedSlider.addEventListener('input', function() {
-        speedValue.textContent = this.value;
-    });
 }
 
 /**
@@ -34,28 +37,98 @@ function bindEvents() {
     // 连接测试
     document.getElementById('test-connection-btn').addEventListener('click', testConnection);
     
-    // 预训练音色测试
+    // 语音样本采集
-    document.getElementById('preset-voice-form').addEventListener('submit', generatePresetVoice);
+    document.getElementById('voice-sample-form').addEventListener('submit', uploadVoiceSample);
-    document.getElementById('preset-random-seed').addEventListener('click', () => getRandomSeed('preset-seed'));
+    document.getElementById('voice-sample-upload').addEventListener('change', handleFileSelect);
+    document.getElementById('start-recording').addEventListener('click', startRecording);
+    document.getElementById('stop-recording').addEventListener('click', stopRecording);
     
-    // 自然语言控制测试
+    // 语音克隆生成
-    document.getElementById('natural-control-form').addEventListener('submit', generateNaturalControl);
+    document.getElementById('clone-generation-form').addEventListener('submit', generateClonedVoice);
-    document.getElementById('natural-random-seed').addEventListener('click', () => getRandomSeed('natural-seed'));
-    
-    // 语音克隆测试
-    document.getElementById('audio-upload-form').addEventListener('submit', uploadReferenceAudio);
-    document.getElementById('voice-clone-form').addEventListener('submit', generateVoiceClone);
     document.getElementById('clone-random-seed').addEventListener('click', () => getRandomSeed('clone-seed'));
     
-    // 清空日志
+    // 高级功能
+    document.getElementById('preset-voice-form').addEventListener('submit', generatePresetVoice);
+    document.getElementById('natural-control-form').addEventListener('submit', generateNaturalControl);
+    
+    // 其他
     document.getElementById('clear-log').addEventListener('click', clearLog);
 }
 
+/**
+ * 更新步骤指示器
+ */
+function updateStepIndicators() {
+    for (let i = 1; i <= 4; i++) {
+        const indicator = document.getElementById(`step-${i}-indicator`);
+        indicator.classList.remove('active', 'completed');
+        
+        if (i < currentStep) {
+            indicator.classList.add('completed');
+        } else if (i === currentStep) {
+            indicator.classList.add('active');
+        }
+    }
+    
+    // 更新连接线
+    document.querySelectorAll('.step-line').forEach((line, index) => {
+        if (index + 1 < currentStep) {
+            line.classList.add('completed');
+        } else {
+            line.classList.remove('completed');
+        }
+    });
+}
+
+/**
+ * 跳转到指定步骤
+ */
+function goToStep(step) {
+    currentStep = step;
+    updateStepIndicators();
+    
+    // 启用/禁用相应按钮
+    if (step >= 3) {
+        document.getElementById('generate-clone-btn').disabled = false;
+    }
+}
+
+/**
+ * 重置工作流程
+ */
+function resetWorkflow() {
+    currentStep = 1;
+    updateStepIndicators();
+    
+    // 清空数据
+    uploadedAudioPath = null;
+    recognizedText = "";
+    sampleAudioUrl = null;
+    recordedAudioBlob = null;
+    
+    // 重置界面
+    document.getElementById('sample-player').style.display = 'none';
+    document.getElementById('recognition-result').style.display = 'none';
+    document.getElementById('recognition-waiting').style.display = 'block';
+    document.getElementById('comparison-result').style.display = 'none';
+    document.getElementById('comparison-waiting').style.display = 'block';
+    
+    // 重置按钮状态
+    document.getElementById('upload-sample-btn').disabled = true;
+    document.getElementById('generate-clone-btn').disabled = true;
+    
+    // 清空文件选择
+    document.getElementById('voice-sample-upload').value = '';
+    
+    addLog('工作流程已重置，可以重新开始', 'info');
+}
+
 /**
  * 显示加载状态
  */
-function showLoading(message = '正在处理中...') {
+function showLoading(message = '正在处理中...', detail = '请稍候...') {
     document.getElementById('loading-message').textContent = message;
+    document.getElementById('loading-detail').textContent = detail;
     loadingModal.show();
 }
 
@@ -93,7 +166,7 @@ function addLog(message, type = 'info') {
  */
 function clearLog() {
     const logContainer = document.getElementById('test-log');
-    logContainer.innerHTML = '<p class="text-muted">测试记录将显示在这里...</p>';
+    logContainer.innerHTML = '<p class="text-muted">操作记录将显示在这里...</p>';
 }
 
 /**
@@ -116,7 +189,6 @@ function showError(message) {
     const bsToast = new bootstrap.Toast(toast);
     bsToast.show();
     
-    // 自动移除
     setTimeout(() => {
         if (toast.parentNode) {
             toast.parentNode.removeChild(toast);
@@ -144,7 +216,6 @@ function showSuccess(message) {
     const bsToast = new bootstrap.Toast(toast);
     bsToast.show();
     
-    // 自动移除
     setTimeout(() => {
         if (toast.parentNode) {
             toast.parentNode.removeChild(toast);
@@ -152,6 +223,14 @@ function showSuccess(message) {
     }, 3000);
 }
 
+/**
+ * 创建音频播放URL
+ */
+function createAudioUrl(audioPath) {
+    const filename = audioPath.split('/').pop();
+    return `/voice-test/download-audio/${filename}`;
+}
+
 /**
  * 测试连接
  */
@@ -179,10 +258,10 @@ async function testConnection() {
             statusDiv.innerHTML = `
                 <span class="text-success">
                     <i class="fas fa-check-circle me-2"></i>连接成功
-                    <small class="text-muted">（${result.api_url}）</small>
+                    <small class="text-muted">支持语音克隆、识别、自然控制</small>
                 </span>
             `;
-            addLog(`连接成功！可用音色数量: ${result.available_voices ? result.available_voices.length : 0}`, 'success');
+            addLog(`连接成功！核心功能可用：语音克隆、识别、自然控制`, 'success');
             
             // 更新音色列表
             if (result.available_voices) {
@@ -259,147 +338,155 @@ async function getRandomSeed(inputId) {
 }
 
 /**
- * 预训练音色语音生成
+ * 处理文件选择
  */
-async function generatePresetVoice(e) {
+function handleFileSelect(e) {
-    e.preventDefault();
+    const file = e.target.files[0];
+    if (file) {
+        // 重置录音状态
+        recordedAudioBlob = null;
+        sampleAudioUrl = null;
         
-    const text = document.getElementById('preset-text').value.trim();
+        addLog(`选择了音频文件: ${file.name} (${(file.size/1024/1024).toFixed(2)} MB)`);
-    const voice = document.getElementById('preset-voice').value;
+        document.getElementById('upload-sample-btn').disabled = false;
-    const seed = parseInt(document.getElementById('preset-seed').value);
+        document.getElementById('upload-sample-btn').innerHTML = '<i class="fas fa-upload me-2"></i>上传并识别语音';
-    const speed = parseFloat(document.getElementById('preset-speed').value);
-    
-    if (!text) {
-        showError('请输入要合成的文本');
-        return;
-    }
-    
-    showLoading('正在生成语音...');
-    addLog(`开始预训练音色生成 - 音色: ${voice}, 种子: ${seed}, 语速: ${speed}x`);
-    
-    try {
-        const response = await fetch('/voice-test/api/voice-test/generate/preset', {
-            method: 'POST',
-            headers: {
-                'Content-Type': 'application/json',
-            },
-            body: JSON.stringify({
-                text: text,
-                voice: voice,
-                seed: seed,
-                speed: speed
-            })
-        });
-        
-        const result = await response.json();
-        
-        if (result.success) {
-            // 显示音频播放器
-            const audioSource = document.getElementById('preset-audio-source');
-            const resultDiv = document.getElementById('preset-result');
-            
-            audioSource.src = result.audio_url;
-            audioSource.parentElement.load();
-            resultDiv.style.display = 'block';
-            
-            addLog(`预训练音色生成成功！音频地址: ${result.audio_url}`, 'success');
-            showSuccess('语音生成成功！');
-        } else {
-            addLog(`预训练音色生成失败: ${result.message}`, 'error');
-            showError(result.message);
-        }
-        
-    } catch (error) {
-        addLog(`预训练音色生成出错: ${error.message}`, 'error');
-        showError('生成失败，请检查网络连接');
-    } finally {
-        hideLoading();
     }
 }
 
 /**
- * 自然语言控制语音生成
+ * 开始录音
  */
-async function generateNaturalControl(e) {
+async function startRecording() {
-    e.preventDefault();
-    
-    const text = document.getElementById('natural-text').value.trim();
-    const instruction = document.getElementById('natural-instruction').value.trim();
-    const seed = parseInt(document.getElementById('natural-seed').value);
-    
-    if (!text) {
-        showError('请输入要合成的文本');
-        return;
-    }
-    
-    if (!instruction) {
-        showError('请输入语音指令');
-        return;
-    }
-    
-    showLoading('正在生成语音...');
-    addLog(`开始自然语言控制生成 - 指令: ${instruction}, 种子: ${seed}`);
-    
     try {
-        const response = await fetch('/voice-test/api/voice-test/generate/natural', {
+        // 重置文件选择
-            method: 'POST',
+        document.getElementById('voice-sample-upload').value = '';
-            headers: {
+        
-                'Content-Type': 'application/json',
+        const stream = await navigator.mediaDevices.getUserMedia({ 
-            },
+            audio: {
-            body: JSON.stringify({
+                sampleRate: 16000,  // 设置采样率为16kHz
-                text: text,
+                channelCount: 1,    // 单声道
-                instruction: instruction,
+                echoCancellation: true,
-                seed: seed
+                noiseSuppression: true
-            })
+            }
         });
         
-        const result = await response.json();
+        // 创建MediaRecorder，明确指定格式
+        const options = {
+            mimeType: 'audio/webm;codecs=opus'  // 使用webm格式
+        };
         
-        if (result.success) {
+        // 检查浏览器支持的格式
-            // 显示音频播放器
+        if (!MediaRecorder.isTypeSupported(options.mimeType)) {
-            const audioSource = document.getElementById('natural-audio-source');
+            if (MediaRecorder.isTypeSupported('audio/webm')) {
-            const resultDiv = document.getElementById('natural-result');
+                options.mimeType = 'audio/webm';
-            
+            } else if (MediaRecorder.isTypeSupported('audio/wav')) {
-            audioSource.src = result.audio_url;
+                options.mimeType = 'audio/wav';
-            audioSource.parentElement.load();
+            } else {
-            resultDiv.style.display = 'block';
+                // 使用默认格式
-            
+                delete options.mimeType;
-            addLog(`自然语言控制生成成功！音频地址: ${result.audio_url}`, 'success');
+            }
-            showSuccess('语音生成成功！');
-        } else {
-            addLog(`自然语言控制生成失败: ${result.message}`, 'error');
-            showError(result.message);
         }
         
+        mediaRecorder = new MediaRecorder(stream, options);
+        audioChunks = [];
+        
+        mediaRecorder.ondataavailable = function(event) {
+            if (event.data.size > 0) {
+                audioChunks.push(event.data);
+            }
+        };
+        
+        mediaRecorder.onstop = function() {
+            // 创建音频Blob
+            recordedAudioBlob = new Blob(audioChunks, { 
+                type: mediaRecorder.mimeType || 'audio/webm' 
+            });
+            
+            const audioUrl = URL.createObjectURL(recordedAudioBlob);
+            
+            // 显示录音预览
+            const sampleAudio = document.getElementById('sample-audio');
+            const sampleSource = document.getElementById('sample-audio-source');
+            sampleSource.src = audioUrl;
+            sampleAudio.load();
+            document.getElementById('sample-player').style.display = 'block';
+            
+            // 启用上传按钮
+            document.getElementById('upload-sample-btn').disabled = false;
+            document.getElementById('upload-sample-btn').innerHTML = '<i class="fas fa-upload me-2"></i>上传并识别语音';
+            
+            // 保存录音数据
+            sampleAudioUrl = audioUrl;
+            
+            addLog(`录音完成，格式: ${mediaRecorder.mimeType || 'default'}, 大小: ${(recordedAudioBlob.size/1024).toFixed(1)} KB`, 'success');
+        };
+        
+        mediaRecorder.start(100); // 每100ms收集一次数据
+        
+        // 更新UI
+        document.getElementById('start-recording').disabled = true;
+        document.getElementById('stop-recording').disabled = false;
+        document.getElementById('recording-status').textContent = '正在录音...';
+        document.getElementById('recording-status').className = 'text-danger';
+        
+        addLog('开始录音...', 'info');
+        
     } catch (error) {
-        addLog(`自然语言控制生成出错: ${error.message}`, 'error');
+        addLog(`录音失败: ${error.message}`, 'error');
-        showError('生成失败，请检查网络连接');
+        showError('录音失败，请检查麦克风权限');
-    } finally {
-        hideLoading();
     }
 }
 
 /**
- * 上传参考音频
+ * 停止录音
  */
-async function uploadReferenceAudio(e) {
+function stopRecording() {
-    e.preventDefault();
+    if (mediaRecorder && mediaRecorder.state !== 'inactive') {
-    
+        mediaRecorder.stop();
-    const fileInput = document.getElementById('reference-audio');
+        mediaRecorder.stream.getTracks().forEach(track => track.stop());
-    const file = fileInput.files[0];
-    
-    if (!file) {
-        showError('请选择音频文件');
-        return;
     }
     
-    showLoading('正在上传并识别音频...');
+    // 更新UI
-    addLog(`开始上传音频文件: ${file.name} (${(file.size/1024/1024).toFixed(2)} MB)`);
+    document.getElementById('start-recording').disabled = false;
+    document.getElementById('stop-recording').disabled = true;
+    document.getElementById('recording-status').textContent = '录音已完成';
+    document.getElementById('recording-status').className = 'text-success';
     
-    const formData = new FormData();
+    addLog('录音停止', 'info');
-    formData.append('audio', file);
+}
+
+/**
+ * 上传语音样本并进行识别
+ */
+async function uploadVoiceSample(e) {
+    e.preventDefault();
+    
+    showLoading('正在上传和识别语音...', '包括格式转换和语音识别，请稍候');
+    addLog('开始上传语音样本进行识别...');
     
     try {
+        const fileInput = document.getElementById('voice-sample-upload');
+        const file = fileInput.files[0];
+        
+        let formData = new FormData();
+        
+        if (file) {
+            // 上传文件
+            formData.append('audio', file);
+            addLog(`上传文件: ${file.name}`);
+        } else if (recordedAudioBlob) {
+            // 上传录音 - 使用正确的文件名和类型
+            const timestamp = new Date().toISOString().replace(/[:.]/g, '-');
+            const filename = `recording_${timestamp}.webm`;
+            formData.append('audio', recordedAudioBlob, filename);
+            addLog(`上传录音: ${filename}, 大小: ${(recordedAudioBlob.size/1024).toFixed(1)} KB`);
+        } else {
+            showError('请选择音频文件或先录音');
+            hideLoading();
+            return;
+        }
+        
+        // 上传并识别
         const response = await fetch('/voice-test/api/voice-test/upload-audio', {
             method: 'POST',
             body: formData
@@ -410,27 +497,34 @@ async function uploadReferenceAudio(e) {
         if (result.success) {
             // 保存音频路径
             uploadedAudioPath = result.file_path;
+            recognizedText = result.recognized_text || '';
             
             // 显示识别结果
-            const resultDiv = document.getElementById('upload-result');
+            document.getElementById('recognized-text').value = recognizedText;
-            const recognizedText = document.getElementById('recognized-text');
+            document.getElementById('recognition-result').style.display = 'block';
+            document.getElementById('recognition-waiting').style.display = 'none';
             
-            recognizedText.value = result.recognized_text || '';
+            // 更新步骤
-            resultDiv.style.display = 'block';
+            goToStep(2);
+            setTimeout(() => goToStep(3), 1000);
             
-            // 启用克隆按钮
+            addLog(`语音识别成功: "${recognizedText}"`, 'success');
-            const cloneBtn = document.querySelector('#voice-clone-form button[type="submit"]');
+            addLog(`音频处理信息: ${result.file_info?.format || '已转换格式'}`, 'info');
-            cloneBtn.disabled = false;
+            showSuccess('语音样本上传成功！AI已识别出内容');
+            
+            // 保存原始音频用于对比
+            const originalAudio = document.getElementById('original-audio');
+            const originalSource = document.getElementById('original-audio-source');
+            originalSource.src = createAudioUrl(uploadedAudioPath);
+            originalAudio.load();
             
-            addLog(`音频上传成功！识别文本: ${result.recognized_text || '(无内容)'}`, 'success');
-            showSuccess('音频上传成功！');
         } else {
-            addLog(`音频上传失败: ${result.message}`, 'error');
+            addLog(`语音识别失败: ${result.message}`, 'error');
             showError(result.message);
         }
         
     } catch (error) {
-        addLog(`音频上传出错: ${error.message}`, 'error');
+        addLog(`上传出错: ${error.message}`, 'error');
         showError('上传失败，请检查网络连接');
     } finally {
         hideLoading();
@@ -438,27 +532,28 @@ async function uploadReferenceAudio(e) {
 }
 
 /**
- * 语音克隆生成
+ * 生成克隆语音
  */
-async function generateVoiceClone(e) {
+async function generateClonedVoice(e) {
     e.preventDefault();
     
     if (!uploadedAudioPath) {
-        showError('请先上传参考音频');
+        showError('请先上传语音样本');
         return;
     }
     
     const text = document.getElementById('clone-text').value.trim();
-    const referenceText = document.getElementById('recognized-text').value.trim();
     const seed = parseInt(document.getElementById('clone-seed').value);
+    const referenceText = document.getElementById('recognized-text').value.trim();
     
     if (!text) {
         showError('请输入要合成的文本');
         return;
     }
     
-    showLoading('正在进行语音克隆...');
+    showLoading('正在克隆你的声音...', '这是最复杂的步骤，请耐心等待');
-    addLog(`开始语音克隆 - 种子: ${seed}`);
+    addLog(`开始语音克隆 - 目标文本: "${text.substring(0, 20)}..."`);
+    addLog(`使用音频文件: ${uploadedAudioPath}`);
     
     try {
         const response = await fetch('/voice-test/api/voice-test/generate/clone', {
@@ -477,25 +572,143 @@ async function generateVoiceClone(e) {
         const result = await response.json();
         
         if (result.success) {
-            // 显示音频播放器
+            // 显示克隆语音
-            const audioSource = document.getElementById('clone-audio-source');
+            const clonedAudio = document.getElementById('cloned-audio');
-            const resultDiv = document.getElementById('clone-result');
+            const clonedSource = document.getElementById('cloned-audio-source');
+            clonedSource.src = createAudioUrl(result.audio_url);
+            clonedAudio.load();
             
-            audioSource.src = result.audio_url;
+            // 显示对比界面
-            audioSource.parentElement.load();
+            document.getElementById('comparison-result').style.display = 'block';
-            resultDiv.style.display = 'block';
+            document.getElementById('comparison-waiting').style.display = 'none';
+            
+            // 更新到最后步骤
+            goToStep(4);
+            
+            addLog(`🎉 语音克隆成功！请对比原声和克隆效果`, 'success');
+            showSuccess('语音克隆完成！请播放音频对比效果');
             
-            addLog(`语音克隆成功！音频地址: ${result.audio_url}`, 'success');
-            showSuccess('语音克隆成功！');
         } else {
             addLog(`语音克隆失败: ${result.message}`, 'error');
-            showError(result.message);
+            showError(result.message || '语音克隆失败，请重试');
         }
         
     } catch (error) {
-        addLog(`语音克隆出错: ${error.message}`, 'error');
+        addLog(`克隆出错: ${error.message}`, 'error');
         showError('克隆失败，请检查网络连接');
     } finally {
         hideLoading();
     }
 }
+
+/**
+ * 预训练音色语音生成（高级功能）
+ */
+async function generatePresetVoice(e) {
+    e.preventDefault();
+    
+    const text = document.getElementById('preset-text').value.trim();
+    const voice = document.getElementById('preset-voice').value;
+    
+    if (!text) {
+        showError('请输入要合成的文本');
+        return;
+    }
+    
+    showLoading('正在生成预训练音色语音...');
+    
+    try {
+        const response = await fetch('/voice-test/api/voice-test/generate/preset', {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/json',
+            },
+            body: JSON.stringify({
+                text: text,
+                voice: voice,
+                seed: 42,
+                speed: 1.0
+            })
+        });
+        
+        const result = await response.json();
+        
+        if (result.success) {
+            const audioSource = document.getElementById('preset-audio-source');
+            const resultDiv = document.getElementById('preset-result');
+            
+            audioSource.src = createAudioUrl(result.audio_url);
+            audioSource.parentElement.load();
+            resultDiv.style.display = 'block';
+            
+            addLog(`预训练音色生成成功 - ${voice}`, 'success');
+        } else {
+            addLog(`预训练音色生成失败: ${result.message}`, 'error');
+            showError(result.message);
+        }
+        
+    } catch (error) {
+        addLog(`生成出错: ${error.message}`, 'error');
+        showError('生成失败，请检查网络连接');
+    } finally {
+        hideLoading();
+    }
+}
+
+/**
+ * 自然语言控制语音生成（高级功能）
+ */
+async function generateNaturalControl(e) {
+    e.preventDefault();
+    
+    const text = document.getElementById('natural-text').value.trim();
+    const instruction = document.getElementById('natural-instruction').value.trim();
+    
+    if (!text) {
+        showError('请输入要合成的文本');
+        return;
+    }
+    
+    if (!instruction) {
+        showError('请输入语音指令');
+        return;
+    }
+    
+    showLoading('正在生成自然语言控制语音...');
+    
+    try {
+        const response = await fetch('/voice-test/api/voice-test/generate/natural', {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/json',
+            },
+            body: JSON.stringify({
+                text: text,
+                instruction: instruction,
+                seed: 42
+            })
+        });
+        
+        const result = await response.json();
+        
+        if (result.success) {
+            const audioSource = document.getElementById('natural-audio-source');
+            const resultDiv = document.getElementById('natural-result');
+            
+            audioSource.src = createAudioUrl(result.audio_url);
+            audioSource.parentElement.load();
+            resultDiv.style.display = 'block';
+            
+            addLog(`自然语言控制生成成功`, 'success');
+        } else {
+            addLog(`自然语言控制生成失败: ${result.message}`, 'error');
+            showError(result.message);
+        }
+        
+    } catch (error) {
+        addLog(`生成出错: ${error.message}`, 'error');
+        showError('生成失败，请检查网络连接');
+    } finally {
+        hideLoading();
+    }
+}

app/templates/voice_test/index.html

@@ -1,6 +1,6 @@
 {% extends "base.html" %}
 
-{% block title %}CosyVoice API 测试 - 儿童语言学习系统{% endblock %}
+{% block title %}语音克隆测试 - 儿童语言学习系统{% endblock %}
 
 {% block content %}
 <div class="container py-4">
@@ -13,10 +13,10 @@
                 </a>
                 <div>
                     <h2 class="fw-bold mb-1">
-                        <i class="fas fa-microphone-alt text-primary me-2"></i>
+                        <i class="fas fa-magic text-primary me-2"></i>
-                        CosyVoice API 测试
+                        语音克隆技术测试
                     </h2>
-                    <p class="text-muted mb-0">测试语音合成的各种功能</p>
+                    <p class="text-muted mb-0">体验"听自己说"的神奇技术 - 让AI学会你的声音</p>
                 </div>
             </div>
         </div>
@@ -43,168 +43,290 @@
         </div>
     </div>
 
-    <!-- 功能测试区域 -->
+    <!-- 核心功能：语音克隆工作流程 -->
-    <div class="row">
+    <div class="row mb-4">
-        <!-- 预训练音色测试 -->
+        <div class="col-12">
-        <div class="col-lg-6 mb-4">
+            <div class="card border-0 shadow-sm border-primary" style="border-width: 2px !important;">
-            <div class="card border-0 shadow-sm h-100">
                 <div class="card-body">
-                    <h5 class="card-title text-primary">
+                    <h4 class="card-title text-primary">
-                        <i class="fas fa-user-tie me-2"></i>预训练音色测试
+                        <i class="fas fa-clone me-2"></i>语音克隆工作流程
-                    </h5>
+                        <span class="badge bg-primary ms-2">核心功能</span>
-                    <form id="preset-voice-form">
+                    </h4>
-                        <div class="mb-3">
+                    <p class="text-muted mb-4">按照步骤操作，体验完整的语音克隆过程</p>
-                            <label class="form-label">要合成的文本</label>
+
-                            <textarea class="form-control" id="preset-text" rows="3" 
+                    <!-- 步骤指示器 -->
-                                placeholder="输入要转换成语音的文字...">今天天气真不错，适合出门散步。</textarea>
+                    <div class="row mb-4">
-                        </div>
+                        <div class="col-12">
-                        <div class="mb-3">
+                            <div class="d-flex justify-content-between position-relative">
-                            <label class="form-label">选择音色</label>
+                                <div class="step-indicator active" id="step-1-indicator">
-                            <select class="form-select" id="preset-voice">
+                                    <div class="step-circle">1</div>
-                                <option value="中文女">中文女</option>
+                                    <div class="step-label">录制声音</div>
-                                <option value="中文男">中文男</option>
+                                </div>
-                                <option value="播报女">播报女</option>
+                                <div class="step-line"></div>
-                                <option value="新闻播报男">新闻播报男</option>
+                                <div class="step-indicator" id="step-2-indicator">
-                                <option value="英文女">英文女</option>
+                                    <div class="step-circle">2</div>
-                                <option value="英文男">英文男</option>
+                                    <div class="step-label">识别文字</div>
-                            </select>
+                                </div>
-                        </div>
+                                <div class="step-line"></div>
-                        <div class="row mb-3">
+                                <div class="step-indicator" id="step-3-indicator">
-                            <div class="col-6">
+                                    <div class="step-circle">3</div>
-                                <label class="form-label">随机种子</label>
+                                    <div class="step-label">克隆生成</div>
-                                <div class="input-group">
+                                </div>
-                                    <input type="number" class="form-control" id="preset-seed" value="42">
+                                <div class="step-line"></div>
-                                    <button type="button" class="btn btn-outline-secondary" id="preset-random-seed">
+                                <div class="step-indicator" id="step-4-indicator">
-                                        <i class="fas fa-dice"></i>
+                                    <div class="step-circle">4</div>
-                                    </button>
+                                    <div class="step-label">效果对比</div>
                                 </div>
                             </div>
-                            <div class="col-6">
-                                <label class="form-label">语速</label>
-                                <input type="range" class="form-range" id="preset-speed" min="0.5" max="2" step="0.1" value="1">
-                                <small class="text-muted">当前: <span id="preset-speed-value">1.0</span>x</small>
-                            </div>
-                        </div>
-                        <button type="submit" class="btn btn-primary w-100">
-                            <i class="fas fa-play me-2"></i>生成语音
-                        </button>
-                    </form>
-                    <div id="preset-result" class="mt-3" style="display: none;">
-                        <div class="alert alert-success">
-                            <i class="fas fa-check-circle me-2"></i>生成成功！
-                            <audio controls class="d-block mt-2 w-100">
-                                <source id="preset-audio-source" type="audio/wav">
-                                您的浏览器不支持音频播放。
-                            </audio>
                         </div>
                     </div>
-                </div>
-            </div>
-        </div>
 
-        <!-- 自然语言控制测试 -->
-        <div class="col-lg-6 mb-4">
-            <div class="card border-0 shadow-sm h-100">
-                <div class="card-body">
-                    <h5 class="card-title text-success">
-                        <i class="fas fa-comments me-2"></i>自然语言控制测试
-                    </h5>
-                    <form id="natural-control-form">
-                        <div class="mb-3">
-                            <label class="form-label">要合成的文本</label>
-                            <textarea class="form-control" id="natural-text" rows="3" 
-                                placeholder="输入要转换成语音的文字...">这是一段测试文本，用于验证自然语言控制功能。</textarea>
-                        </div>
-                        <div class="mb-3">
-                            <label class="form-label">语音指令</label>
-                            <textarea class="form-control" id="natural-instruction" rows="2" 
-                                placeholder="描述你想要的语音风格...">请用温柔甜美的女声朗读</textarea>
-                            <div class="form-text">
-                                示例：请用活泼开朗的语调、请用严肃正式的男声、请用轻柔的语气等
-                            </div>
-                        </div>
-                        <div class="mb-3">
-                            <label class="form-label">随机种子</label>
-                            <div class="input-group">
-                                <input type="number" class="form-control" id="natural-seed" value="42">
-                                <button type="button" class="btn btn-outline-secondary" id="natural-random-seed">
-                                    <i class="fas fa-dice"></i>
-                                </button>
-                            </div>
-                        </div>
-                        <button type="submit" class="btn btn-success w-100">
-                            <i class="fas fa-magic me-2"></i>生成语音
-                        </button>
-                    </form>
-                    <div id="natural-result" class="mt-3" style="display: none;">
-                        <div class="alert alert-success">
-                            <i class="fas fa-check-circle me-2"></i>生成成功！
-                            <audio controls class="d-block mt-2 w-100">
-                                <source id="natural-audio-source" type="audio/wav">
-                                您的浏览器不支持音频播放。
-                            </audio>
-                        </div>
-                    </div>
-                </div>
-            </div>
-        </div>
-
-        <!-- 语音克隆测试 -->
-        <div class="col-12 mb-4">
-            <div class="card border-0 shadow-sm">
-                <div class="card-body">
-                    <h5 class="card-title text-warning">
-                        <i class="fas fa-clone me-2"></i>语音克隆测试
-                    </h5>
                     <div class="row">
-                        <div class="col-lg-6">
+                        <!-- 步骤1：语音样本采集 -->
-                            <h6 class="fw-bold mb-3">1. 上传参考音频</h6>
+                        <div class="col-lg-6 mb-4">
-                            <form id="audio-upload-form" enctype="multipart/form-data">
+                            <div class="step-content" id="step-1-content">
-                                <div class="mb-3">
+                                <h5 class="text-success">
-                                    <input type="file" class="form-control" id="reference-audio" 
+                                    <i class="fas fa-microphone me-2"></i>步骤1：录制你的声音样本
-                                           accept=".wav,.mp3,.m4a,.flac" required>
+                                </h5>
-                                    <div class="form-text">
+                                <p class="text-muted">录制一段3-10秒的清晰语音，作为语音克隆的模板</p>
-                                        支持格式：WAV、MP3、M4A、FLAC，建议3-10秒
+                                
+                                <form id="voice-sample-form" enctype="multipart/form-data">
+                                    <div class="mb-3">
+                                        <label class="form-label">方式一：上传音频文件</label>
+                                        <input type="file" class="form-control" id="voice-sample-upload" 
+                                               accept=".wav,.mp3,.m4a,.flac">
+                                        <div class="form-text">建议：WAV格式，16kHz采样率，3-10秒时长</div>
+                                    </div>
+                                    
+                                    <div class="mb-3 text-center">
+                                        <span class="text-muted">或</span>
+                                    </div>
+                                    
+                                    <div class="mb-3">
+                                        <label class="form-label">方式二：在线录音</label>
+                                        <div class="text-center">
+                                            <button type="button" id="start-recording" class="btn btn-danger me-2">
+                                                <i class="fas fa-record-vinyl me-2"></i>开始录音
+                                            </button>
+                                            <button type="button" id="stop-recording" class="btn btn-secondary" disabled>
+                                                <i class="fas fa-stop me-2"></i>停止录音
+                                            </button>
+                                            <div class="mt-2">
+                                                <small id="recording-status" class="text-muted">点击开始录音</small>
+                                            </div>
+                                        </div>
+                                    </div>
+                                    
+                                    <button type="submit" class="btn btn-success w-100" disabled id="upload-sample-btn">
+                                        <i class="fas fa-upload me-2"></i>上传语音样本
+                                    </button>
+                                </form>
+
+                                <!-- 样本播放区域 -->
+                                <div id="sample-player" class="mt-3" style="display: none;">
+                                    <div class="alert alert-info">
+                                        <h6><i class="fas fa-headphones me-2"></i>你的语音样本：</h6>
+                                        <audio controls class="w-100 mt-2" id="sample-audio">
+                                            <source id="sample-audio-source" type="audio/wav">
+                                        </audio>
                                     </div>
                                 </div>
-                                <button type="submit" class="btn btn-outline-warning">
-                                    <i class="fas fa-upload me-2"></i>上传并识别
-                                </button>
-                            </form>
-                            <div id="upload-result" class="mt-3" style="display: none;">
-                                <div class="alert alert-info">
-                                    <h6><i class="fas fa-text-width me-2"></i>识别的文本内容：</h6>
-                                    <textarea class="form-control mt-2" id="recognized-text" rows="3"></textarea>
-                                </div>
                             </div>
                         </div>
-                        <div class="col-lg-6">
+
-                            <h6 class="fw-bold mb-3">2. 生成克隆语音</h6>
+                        <!-- 步骤2：语音识别 -->
-                            <form id="voice-clone-form">
+                        <div class="col-lg-6 mb-4">
-                                <div class="mb-3">
+                            <div class="step-content" id="step-2-content">
-                                    <label class="form-label">要合成的文本</label>
+                                <h5 class="text-info">
-                                    <textarea class="form-control" id="clone-text" rows="3" 
+                                    <i class="fas fa-text-width me-2"></i>步骤2：语音识别结果
-                                        placeholder="输入要用克隆声音说的话...">这是使用克隆声音合成的新内容。</textarea>
+                                </h5>
-                                </div>
+                                <p class="text-muted">AI识别出你说的内容，这将用于语音克隆训练</p>
-                                <div class="mb-3">
+                                
-                                    <label class="form-label">随机种子</label>
+                                <div id="recognition-result" style="display: none;">
-                                    <div class="input-group">
+                                    <div class="mb-3">
-                                        <input type="number" class="form-control" id="clone-seed" value="42">
+                                        <label class="form-label">识别的文本内容：</label>
-                                        <button type="button" class="btn btn-outline-secondary" id="clone-random-seed">
+                                        <textarea class="form-control" id="recognized-text" rows="3" 
-                                            <i class="fas fa-dice"></i>
+                                                  placeholder="语音识别结果将显示在这里..."></textarea>
+                                        <div class="form-text">你可以手动修正识别错误的文字</div>
+                                    </div>
+                                    
+                                    <div class="d-flex justify-content-between">
+                                        <small class="text-success">
+                                            <i class="fas fa-check-circle me-1"></i>识别完成
+                                        </small>
+                                        <button class="btn btn-outline-primary btn-sm" onclick="goToStep(3)">
+                                            下一步：克隆生成 <i class="fas fa-arrow-right ms-1"></i>
                                         </button>
                                     </div>
                                 </div>
-                                <button type="submit" class="btn btn-warning w-100" disabled>
+                                
-                                    <i class="fas fa-clone me-2"></i>生成克隆语音
+                                <div id="recognition-waiting" class="text-center text-muted">
-                                </button>
+                                    <i class="fas fa-clock fa-2x mb-2"></i>
-                            </form>
+                                    <p>等待上传语音样本...</p>
-                            <div id="clone-result" class="mt-3" style="display: none;">
+                                </div>
-                                <div class="alert alert-success">
+                            </div>
-                                    <i class="fas fa-check-circle me-2"></i>克隆成功！
+                        </div>
-                                    <audio controls class="d-block mt-2 w-100">
+
-                                        <source id="clone-audio-source" type="audio/wav">
+                        <!-- 步骤3：语音克隆生成 -->
-                                        您的浏览器不支持音频播放。
+                        <div class="col-lg-6 mb-4">
+                            <div class="step-content" id="step-3-content">
+                                <h5 class="text-warning">
+                                    <i class="fas fa-magic me-2"></i>步骤3：克隆你的声音
+                                </h5>
+                                <p class="text-muted">让AI用你的声音说新的话</p>
+                                
+                                <form id="clone-generation-form">
+                                    <div class="mb-3">
+                                        <label class="form-label">想让AI用你的声音说什么？</label>
+                                        <textarea class="form-control" id="clone-text" rows="3" 
+                                                  placeholder="例如：你好，我是AI克隆的声音，听起来像真的一样！">你好，我是AI克隆的声音，听起来像真的一样！</textarea>
+                                    </div>
+                                    
+                                    <div class="row mb-3">
+                                        <div class="col-6">
+                                            <label class="form-label">随机种子</label>
+                                            <div class="input-group">
+                                                <input type="number" class="form-control" id="clone-seed" value="42">
+                                                <button type="button" class="btn btn-outline-secondary" id="clone-random-seed">
+                                                    <i class="fas fa-dice"></i>
+                                                </button>
+                                            </div>
+                                        </div>
+                                        <div class="col-6">
+                                            <label class="form-label">克隆质量</label>
+                                            <select class="form-select">
+                                                <option>高质量（推荐）</option>
+                                            </select>
+                                        </div>
+                                    </div>
+                                    
+                                    <button type="submit" class="btn btn-warning w-100" disabled id="generate-clone-btn">
+                                        <i class="fas fa-magic me-2"></i>生成克隆语音
+                                    </button>
+                                </form>
+                            </div>
+                        </div>
+
+                        <!-- 步骤4：效果对比 -->
+                        <div class="col-lg-6 mb-4">
+                            <div class="step-content" id="step-4-content">
+                                <h5 class="text-success">
+                                    <i class="fas fa-balance-scale me-2"></i>步骤4：效果对比
+                                </h5>
+                                <p class="text-muted">对比原声和克隆声音的效果</p>
+                                
+                                <div id="comparison-result" style="display: none;">
+                                    <div class="row">
+                                        <div class="col-6">
+                                            <h6><i class="fas fa-user me-2"></i>原声样本</h6>
+                                            <audio controls class="w-100" id="original-audio">
+                                                <source id="original-audio-source" type="audio/wav">
+                                            </audio>
+                                            <small class="text-muted">你录制的原始声音</small>
+                                        </div>
+                                        <div class="col-6">
+                                            <h6><i class="fas fa-robot me-2"></i>克隆声音</h6>
+                                            <audio controls class="w-100" id="cloned-audio">
+                                                <source id="cloned-audio-source" type="audio/wav">
+                                            </audio>
+                                            <small class="text-muted">AI克隆的声音</small>
+                                        </div>
+                                    </div>
+                                    
+                                    <div class="mt-3 p-3 bg-light rounded">
+                                        <h6><i class="fas fa-star me-2"></i>给克隆效果打分：</h6>
+                                        <div class="btn-group" role="group">
+                                            <input type="radio" class="btn-check" name="rating" id="rating1">
+                                            <label class="btn btn-outline-warning" for="rating1">1★</label>
+                                            <input type="radio" class="btn-check" name="rating" id="rating2">
+                                            <label class="btn btn-outline-warning" for="rating2">2★</label>
+                                            <input type="radio" class="btn-check" name="rating" id="rating3">
+                                            <label class="btn btn-outline-warning" for="rating3">3★</label>
+                                            <input type="radio" class="btn-check" name="rating" id="rating4">
+                                            <label class="btn btn-outline-warning" for="rating4">4★</label>
+                                            <input type="radio" class="btn-check" name="rating" id="rating5">
+                                            <label class="btn btn-outline-warning" for="rating5">5★</label>
+                                        </div>
+                                    </div>
+                                </div>
+                                
+                                <div id="comparison-waiting" class="text-center text-muted">
+                                    <i class="fas fa-hourglass-half fa-2x mb-2"></i>
+                                    <p>等待生成克隆语音...</p>
+                                </div>
+                            </div>
+                        </div>
+                    </div>
+
+                    <!-- 重新开始按钮 -->
+                    <div class="text-center mt-4">
+                        <button class="btn btn-outline-primary" onclick="resetWorkflow()">
+                            <i class="fas fa-redo me-2"></i>重新开始克隆流程
+                        </button>
+                    </div>
+                </div>
+            </div>
+        </div>
+    </div>
+
+    <!-- 高级功能区域 -->
+    <div class="row mb-4">
+        <div class="col-12">
+            <div class="card border-0 shadow-sm">
+                <div class="card-body">
+                    <h5 class="card-title">
+                        <i class="fas fa-cogs me-2"></i>高级功能
+                        <button class="btn btn-outline-secondary btn-sm float-end" type="button" data-bs-toggle="collapse" data-bs-target="#advanced-features">
+                            <i class="fas fa-chevron-down"></i>
+                        </button>
+                    </h5>
+                    
+                    <div class="collapse" id="advanced-features">
+                        <div class="row">
+                            <!-- 预训练音色测试 -->
+                            <div class="col-lg-6 mb-4">
+                                <h6 class="text-secondary">预训练音色测试</h6>
+                                <form id="preset-voice-form">
+                                    <div class="mb-3">
+                                        <textarea class="form-control" id="preset-text" rows="2" 
+                                            placeholder="输入要转换成语音的文字...">今天天气真不错，适合出门散步。</textarea>
+                                    </div>
+                                    <div class="mb-3">
+                                        <select class="form-select" id="preset-voice">
+                                            <option value="中文女">中文女</option>
+                                        </select>
+                                    </div>
+                                    <button type="submit" class="btn btn-secondary btn-sm">
+                                        <i class="fas fa-play me-2"></i>生成语音
+                                    </button>
+                                </form>
+                                <div id="preset-result" class="mt-2" style="display: none;">
+                                    <audio controls class="w-100">
+                                        <source id="preset-audio-source" type="audio/wav">
+                                    </audio>
+                                </div>
+                            </div>
+
+                            <!-- 自然语言控制测试 -->
+                            <div class="col-lg-6 mb-4">
+                                <h6 class="text-secondary">自然语言控制</h6>
+                                <form id="natural-control-form">
+                                    <div class="mb-3">
+                                        <textarea class="form-control" id="natural-text" rows="2" 
+                                            placeholder="输入要合成的文字...">这是自然语言控制测试。</textarea>
+                                    </div>
+                                    <div class="mb-3">
+                                        <input class="form-control" id="natural-instruction" 
+                                               placeholder="语音指令，如：请用温柔甜美的女声朗读" 
+                                               value="请用温柔甜美的女声朗读">
+                                    </div>
+                                    <button type="submit" class="btn btn-secondary btn-sm">
+                                        <i class="fas fa-magic me-2"></i>生成语音
+                                    </button>
+                                </form>
+                                <div id="natural-result" class="mt-2" style="display: none;">
+                                    <audio controls class="w-100">
+                                        <source id="natural-audio-source" type="audio/wav">
                                     </audio>
                                 </div>
                             </div>
@@ -215,19 +337,19 @@
         </div>
     </div>
 
-    <!-- 测试记录区域 -->
+    <!-- 操作日志 -->
     <div class="row">
         <div class="col-12">
             <div class="card border-0 shadow-sm">
                 <div class="card-body">
                     <h5 class="card-title">
-                        <i class="fas fa-history me-2"></i>测试记录
+                        <i class="fas fa-history me-2"></i>操作日志
                         <button class="btn btn-outline-secondary btn-sm float-end" id="clear-log">
                             <i class="fas fa-trash me-1"></i>清空
                         </button>
                     </h5>
                     <div id="test-log" class="border rounded p-3" style="height: 300px; overflow-y: auto; background-color: #f8f9fa;">
-                        <p class="text-muted">测试记录将显示在这里...</p>
+                        <p class="text-muted">操作记录将显示在这里...</p>
                     </div>
                 </div>
             </div>
@@ -242,10 +364,88 @@
             <div class="modal-body text-center py-4">
                 <div class="spinner-border text-primary mb-3" role="status"></div>
                 <h5 id="loading-message">正在处理中...</h5>
+                <p id="loading-detail" class="text-muted">请稍候...</p>
             </div>
         </div>
     </div>
 </div>
+
+<!-- 添加步骤指示器样式 -->
+<style>
+.step-indicator {
+    display: flex;
+    flex-direction: column;
+    align-items: center;
+    position: relative;
+    z-index: 2;
+}
+
+.step-circle {
+    width: 40px;
+    height: 40px;
+    border-radius: 50%;
+    background-color: #e9ecef;
+    color: #6c757d;
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    font-weight: bold;
+    margin-bottom: 8px;
+    transition: all 0.3s ease;
+}
+
+.step-indicator.active .step-circle {
+    background-color: #0d6efd;
+    color: white;
+}
+
+.step-indicator.completed .step-circle {
+    background-color: #198754;
+    color: white;
+}
+
+.step-label {
+    font-size: 0.9rem;
+    text-align: center;
+    color: #6c757d;
+}
+
+.step-indicator.active .step-label {
+    color: #0d6efd;
+    font-weight: 600;
+}
+
+.step-indicator.completed .step-label {
+    color: #198754;
+    font-weight: 600;
+}
+
+.step-line {
+    flex: 1;
+    height: 2px;
+    background-color: #e9ecef;
+    margin: 20px 0;
+    position: relative;
+    z-index: 1;
+}
+
+.step-line.completed {
+    background-color: #198754;
+}
+
+@media (max-width: 768px) {
+    .step-indicator {
+        flex-direction: row;
+        align-items: center;
+        margin-bottom: 10px;
+    }
+    
+    .step-circle {
+        margin-bottom: 0;
+        margin-right: 10px;
+    }
+}
+</style>
 {% endblock %}
 
 {% block scripts %}

requirements.txt

@@ -10,3 +10,5 @@ email-validator==2.0.0
 cryptography==41.0.4
 Werkzeug==2.3.7
 gradio_client==0.8.1
+librosa 
+soundfile