lanhao
59 lines
1.8 KiB
Python
59 lines
1.8 KiB
Python
import os
|
||
import uuid
|
||
import requests
|
||
import base64
|
||
|
||
|
||
def synthesize_doubao(text, speed=1.0, pitch=0, volume=1.0, emotion='neutral'):
|
||
"""调用火山引擎豆包 Seed-ICL 2.0 语音合成 API"""
|
||
api_key = os.getenv('DOUBAO_API_KEY', '')
|
||
if not api_key:
|
||
raise ValueError('DOUBAO_API_KEY未设置')
|
||
|
||
url = 'https://openspeech.bytedance.com/api/v1/tts'
|
||
headers = {
|
||
'x-api-key': api_key,
|
||
'Content-Type': 'application/json',
|
||
'X-Api-Resource-Id': 'seed-icl-2.0'
|
||
}
|
||
|
||
payload = {
|
||
'app': {
|
||
'cluster': 'volcano_icl'
|
||
},
|
||
'user': {
|
||
'uid': 'military_tech_voice'
|
||
},
|
||
'audio': {
|
||
'voice_type': 'S_LK0j18H72',
|
||
'encoding': 'mp3',
|
||
'speed_ratio': speed,
|
||
'pitch_ratio': max(0.1, min(3.0, 1.0 + pitch / 12)), # 前端-12~12映射到0.1~3.0
|
||
'volume_ratio': max(0.1, min(3.0, volume)), # 前端0~10，API 0.1~3.0
|
||
'bitrate': 128000,
|
||
},
|
||
'request': {
|
||
'reqid': str(uuid.uuid4()),
|
||
'text': text,
|
||
'operation': 'query',
|
||
'model_type': 4, # ICL 2.0 模型
|
||
}
|
||
}
|
||
|
||
response = requests.post(url, json=payload, headers=headers, timeout=60)
|
||
|
||
if response.status_code != 200:
|
||
raise Exception(f'豆包API错误: {response.status_code} - {response.text}')
|
||
|
||
result = response.json()
|
||
|
||
# 豆包返回格式: {"code": 3000, "data": "base64编码的音频"}
|
||
if result.get('code') == 3000 and result.get('data'):
|
||
audio_bytes = base64.b64decode(result['data'])
|
||
return audio_bytes
|
||
|
||
# 错误处理
|
||
error_msg = result.get('message', '未知错误')
|
||
error_code = result.get('code', 'unknown')
|
||
raise Exception(f'豆包返回错误 [{error_code}]: {error_msg}')
|