1. 初始化数字人 【所有的使用示例必须放到 script_main(self) 类函数下使用】
def init_metahuman(self, metahuman_mde_path: str, metahuman_scale_width: int, metahuman_scale_height: int, is_update_metahuman: bool = False, enable_random_image: bool = False ) -》 bool:
"""
初始化数字人，第一次初始化需要一些时间
Initializing digital people, it takes some time to initialize for the first time.
metahuman_mde_path: 数字人模型路径
metahuman_scale_width: 数字人宽度缩放倍数，1为原始大小。为2时放大一倍，0.5则缩小一半
metahuman_scale_height: 数字人高度缩放倍数，1为原始大小。为2时放大一倍，0.5则缩小一半
is_update_metahuman: 是否强制更新，默认fasle。为true时强制更新会拖慢初始化速度
enable_random_image: 是否启用随机对比度、亮度和形变 参数，默认Fasle
return: True或者False
metahuman_mde_path: the path of digital human model
metahuman_scale_width: the scaling factor of the digital person width, where 1 is the original size. When it is 2, it is doubled, and when it is 0.5, it is halved
metahuman_scale_height: the scaling factor of the digital person height, where 1 is the original size. When it is 2, it is doubled, and when it is 0.5, it is halved
is_update_metahuman: whether to force the update, the default is fasle. When true, forcing the update will slow down the initialization speed
enable_random_image: whether to enable random contrast, brightness and deformation parameters, the default is Fasle
return: True or False
"""
# 使用示例 [Demo]
result = self.init_metahuman("D:/AiboteMetahuman/metahumanMode", 0.5, 0.5, False, False)
print(result)
2. 数字人说话 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def metahuman_speech(self, save_voice_folder: str, text: str, language: str, voice_name: str, quality: int = 0,
wait_play_sound: bool = True, speech_rate: int = 0, voice_style: str = "General") -》 bool:
"""
数字人说话,
Digital people speak, this function needs to call initSpeechService to initialize voice service
save_voice_folder: 保存的发音文件目录，文件名以0开始依次增加，扩展为.mp3格式
text: 要转换语音的文本
language: 语言，参考语音服务最下方文档 语言和发音人
voice_name: 发音人，参考语音服务最下方文档 语言和发音人
quality: 音质，0低品质 1中品质 2高品质， 默认为0低品质
wait_play_sound: 等待音频播报完毕，默认为 true等待
speech_rate: 语速，默认为0，取值范围 -100 至 200
voice_style: 语音风格，默认General常规风格，其他风格参考开发文档 语言和发音人
return: True或者False
save_voice_folder: the directory of saved pronunciation files.
The file names are increased from 0 and expanded to. wav format
text: the text to be converted into speech
language: language, reference development document language and speaker
Voice_name: speaker, refer to the development document language and speaker
quality: sound quality, 0 low quality 1 medium quality 2 high quality, and 0 low quality by default
wait_play_sound: Wait until the audio broadcast is finished; the default value is true
speech_rate: speech speed, which is 0 by default, and the value range is -100 to 200
voice_style: voice style, the default General style,
and other styles refer to the development document language and speaker
return: True or False
"""
# 使用示例 [Demo]
result = self.metahuman_speech("D:/AiboteMetahuman/voice/1.mp3", "PyAibote is an excellent automation framework", "zh-cn", "zh-cn-XiaochenNeural", 0, True, 0, "General")
print(result)
3. 数字人说话缓存模式 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def metahuman_speech_cache(self, save_voice_folder: str, text: str, language: str, voice_name: str, quality: int = 0,
wait_play_sound: bool = True, speech_rate: int = 0, voice_style: str = "General") -》 bool:
"""
数字人说话缓存模式, 函数一般用于常用的话术播报，非常用话术切勿使用，否则内存泄漏
Digital people speak in cache mode, which requires calling initSpeechService to initialize voice service.
Function is generally used for common speech broadcasting, and it should not be used if it is very popular, otherwise the memory will leak
save_voice_folder: 保存的发音文件目录，文件名以0开始依次增加，扩展为.wav格式
text: 要转换语音的文本
language: 语言，参考 语音服务 最下方文档 语言和发音人
voice_name: 语言，参考 语音服务 最下方文档 语言和发音人
quality: 音质，0低品质 1中品质 2高品质， 默认为0低品质
wait_play_sound: 等待音频播报完毕，默认为 true等待
speech_rate: 语速，默认为0，取值范围 -100 至 200
voice_style: 语音风格，默认General常规风格，其他风格参考开发文档 语言和发音人
return: True或者False
save_voice_folder: the directory of saved pronunciation files. The file names are increased from 0 and expanded to. wav format
text: the text to be converted into speech
language: language, reference development document language and speaker
voice_name: speaker, refer to the development document language and speaker
quality: sound quality, 0 low quality 1 medium quality 2 high quality, and 0 low quality by default
wait_play_sound: Wait until the audio broadcast is finished; the default value is true
speech_rate: speech speed, which is 0 by default, and the value range is -100 to 200
voice_style: voice style, the default General style, and other styles refer to the development document language and speaker
return: True or False
"""
# 使用示例 [Demo]
result = self.metahuman_speech_cache("D:/AiboteMetahuman/voice/1.mp3", "PyAibote is an excellent automation framework", "zh-cn", "zh-cn-XiaochenNeural", 0, True, 0, "General")
print(result)
4. 数字人插入视频 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def metahuman_insert_video(self, video_file_path: str, audio_file_path: str, wait_play_video: bool = True) -》 bool:
"""
数字人插入视频
Digital people insert videos.
video_file_path: 插入的视频文件路径
audio_file_path: 插入的音频文件路径
wait_play_video: 等待视频播放完毕，默认为 True等待
return: True或者False
video_file_path: the path of the inserted video file
audio_file_path: the path of the inserted audio file
wait_play_video: Wait until the video is finished; the default value is true
return: True or False
"""
# 使用示例 [Demo]
result = self.metahuman_insert_video("C:/Users/wenbaobao/Desktop/voide.mp4", "C:/Users/wenbaobao/Desktop/audio.wav", True)
print(result)
5. 替换数字人背景 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def replace_background(self, bg_file_path: str, replace_red: int = -1, replace_green: int = -1, replace_blue: int = -1, sim_value: int = 0) -》 bool:
"""
替换数字人背景
Replace the background of digital people
bg_file_path: 数字人背景 图片/视频 路径，默认不替换背景。仅替换绿幕背景的数字人模型
replace_red: 数字人背景的三通道之一的 R通道色值。默认-1 自动提取
replace_green: 数字人背景的三通道之一的 G通道色值。默认-1 自动提取
replace_blue: 数字人背景的三通道之一的 B通道色值。默认-1 自动提取
sim_value: 相似度。 默认为0，取值应当大于等于0
return: True或者False
bg_file_path: the background picture/video path of digital people, which does not replace the background by default.
Digital human model only replacing green screen background
replace_red: R channel color value of one of the three channels of digital human background. Default -1 automatic extraction
replace_green: the G-channel color value of one of the three channels of digital human background. Default -1 automatic extraction
replace_blue: B-channel color value of one of the three channels of digital human background. Default -1 automatic extraction
sim_value: similarity. The default value is 0 and the value should be greater than or equal to 0
return: True or False
"""
# 使用示例 [Demo]
result = self.replace_background("D:/1.jpg", -1, -1, -1, 50)
print(result)
6. 显示数字人说话的文本 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def show_speech_text(self, origin_y: int = 0, font_type: str = "Arial", font_size: int = 30, font_red: int = 128,
font_green: int = 255, font_blue: int = 0, italic: bool = False,underline: bool = False) -》 bool:
"""
显示数字人说话的文本
Displays the text spoken by a digital person
origin_y, 第一个字显示的起始Y坐标点。 默认0 自适应高度
font_type, 字体样式，支持操作系统已安装的字体。例如"Arial"、"微软雅黑"、"楷体"
font_size, 字体的大小。默认30
font_red, 字体颜色三通道之一的 R通道色值。默认128
font_green, 字体颜色三通道之一的 G通道色值。默认255
font_blue, 字体颜色三通道之一的 B通道色值。默认0
italic, 是否斜体,默认false
underline, 是否有下划线,默认false
return: True或者False
origin_y, the starting y coordinate point displayed by the first word. Default 0 Adaptive Height
font_type, font style, supports fonts installed in the operating system. For example, "Arial", "Microsoft yahei" and "italics"
font_size, the size of the font. Default 30
font_red, the R channel color value of one of the three channels of font color. Default 128
font_green, the color value of G channel, one of the three channels of font color. Default 255
font_blue, the color value of B channel, which is one of the three channels of font color. Default 0
italic, italic or not, the default is false
underline, whether it is underlined or not, the default is false
return: True or False
"""
# 使用示例 [Demo]
result = self.show_speech_text(0, "Arial", 30, 128, 255, 0, False, False)
print(result)
7. 生成数字人短视频 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def make_metahuman_video(self, save_video_folder: str, text: str, language: str, voice_name: str, bg_file_path: str,
sim_value: float = 0, voice_style: str = "General", quality: int = 0, speech_rate: int = 0,) -》 bool:
"""
生成数字人短视频
To generate short video of digital people, this function needs to call initSpeechService to initialize voice service
save_video_folder: 保存的视频目录
text: 要转换语音的文本
language: 语言，参考语音服务最下方文档 语言和发音人
voice_name: 发音人，参考语音服务最下方文档 语言和发音人
bg_file_path: 数字人背景 图片/视频 路径，扣除绿幕会自动获取绿幕的RGB值，null 则不替换背景。仅替换绿幕背景的数字人模型
sim_value: 相似度，默认为0。此处参数用作绿幕扣除微调RBG值。取值应当大于等于0
voice_style: 语音风格，默认General常规风格，其他风格参考开发文档 语言和发音人
quality: 音质，0低品质 1中品质 2高品质， 默认为0低品质
speech_rate: 语速，默认为0，取值范围 -100 至 200
return: True或者False
save_video_folder: the saved video directory
text: the text to be converted into speech
language: language, reference development document language and speaker
voice_name: speaker, refer to the development document language and speaker
bg_file_path: the background picture/video path of the digital person. If the green screen is subtracted,
the RGB value of the green screen will be automatically obtained; if it is null, the background will not be replaced.
Digital human model only replacing green screen background
sim_value: similarity, which defaults to 0. Here, the parameter is used as the RBG value of green screen subtraction fine tuning.
Value should be greater than or equal to 0
voice_style: voice style, the default General style, and other styles refer to the development document language and speaker
quality: sound quality, 0 low quality 1 medium quality 2 high quality, and 0 low quality by default
speech_rate: speech speed, which is 0 by default, and the value range is -100 to 200
return: True or False
"""
# 使用示例 [Demo]
result = self.make_metahuman_video("D://Video", "PyAibote Human", "zh-cn", "zh-cn-XiaochenNeural", "D://1.png", 0, "General", 0, 0)
print(result)
8. 数字人说话文件缓存模式 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def metahuman_speech_byFile(self, audio_path: str, wait_play_sound : bool = True) -》 bool:
"""
数字人说话文件缓存模式
Cache mode of digital human speech file
audio_path: 音频路径， 同名的 .lab文件需要和音频文件在同一目录下
wait_play_sound: 是否等待播报完毕，默认为true 等待
return: True或者False
audio_path: Audio path. lab file with the same name needs to be in the same directory as the audio file
wait_play_sound: whether to wait for the broadcast, the default value is true
return: True or False
"""
# 使用示例 [Demo]
result = self.metahuman_speech_byFile("D:/AiboteMetahuman/voice/1.mp3", True)
print(result)
9. 生成数字人说话文件 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def make_metahuman_speech_file(self, save_audio_path: str, text: str, language: str = "zh-cn", voice_name: str, quality: int = 0, speech_rate: int = 0, voice_style: str = "General") -》 bool:
"""
生成数字人说话文件，生成MP3文件和 lab文件，提供给 metahumanSpeechByFile 和使用
Generate digital human speech files, MP3 files and lab files, and provide them to metahumanSpeechByFile and use
save_audio_path: 保存的音频文件路径，扩展为.MP3格式。同名的 .lab文件需要和音频文件在同一目录下
text: 要转换语音的文本
language: 语言，参考开发文档 语言和发音人
voice_name: 发音人，参考开发文档 语言和发音人
quality: 音质，0低品质 1中品质 2高品质， 默认为0低品质
speech_rate: 语速，默认为0，取值范围 -100 至 200
voice_style: 语音风格，默认General常规风格，其他风格参考开发文档 语言和发音人
return: True或者False
save_audio_path: the path of the saved audio file, expanded to .mp3 format. The. lab file with the same name needs to be in the same directory as the audio file
text: the text to be converted into speech
language: language, reference development document language and speaker
voice_name: speaker, refer to the development document language and speaker
quality: sound quality, 0 low quality 1 medium quality 2 high quality, and 0 low quality by default
speech_rate: speech speed, which is 0 by default, and the value range is -100 to 200
voice_style: voice style, the default General style, and other styles refer to the development document language and speaker
return: True or False
"""
# 使用示例 [Demo]
result = self.make_metahuman_speech_file("D:/AiboteMetahuman/voice/1.mp3", "Aibote is an excellent automation framework", "zh-cn", "zh-cn-XiaochenNeural", 0, 0, "General")
print(result)
10. 打断数字人说话 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def metahuman_speech_break(self) -》 bool:
"""
打断数字人说话，一般用作人机对话场景。metahumanSpeech和metahumanSpeechCache的 waitPlaySound 参数 设置为false时，此函数才有意义
Interrupting a digital person's speech is generally used as a human-computer conversation scene.
This function only makes sense when the waitPlaySound parameter of metahumanSpeech and metahumanSpeechCache is set to false
return: 返回true打断正在说话， 返回false 则为未说话状态
return true to interrupt talking, return false to be silent
"""
# 使用示例 [Demo]
result = self.metahuman_speech_break()
print(result)
11. 数字人说话文件缓存模式 【使用此函数需要先调用 initSpeechService 初始化语音服务】
def metahuman_speech_byFile(self, audio_path: str, wait_play_sound : bool = True) -》 bool:
"""
数字人说话文件缓存模式
Cache mode of digital human speech file
audio_path: 音频路径， 同名的 .lab文件需要和音频文件在同一目录下
wait_play_sound: 是否等待播报完毕，默认为true 等待
return: True或者False
audio_path: Audio path. lab file with the same name needs to be in the same directory as the audio file
wait_play_sound: whether to wait for the broadcast, the default value is true
return: True or False
"""
# 使用示例 [Demo]
result = self.metahuman_speech_byFile("D:/AiboteMetahuman/voice/1.mp3", True)
print(result)
def init_speech_clone_service(self, api_key: str, voice_id: str) -》 bool:
"""
初始化数字人声音克隆服务(不支持win 7系统) 声音克隆服务卡密获取网站：elevenlabs.io
Initialize the digital human voice cloning service (win 7 system is not supported). The voice cloning service card is obtained from the website: elevenlabs.io
api_key: API密钥
voice_id: 声音ID
return: True或者False
api_key: API key
clone_val: Voice ID
return: True or False
"""
# 使用示例 [Demo]
result = self.init_speech_clone_service("178asdf325c95eafdaaasee3bbf64741", "tIdj8l8nPdqV86Ueasdf")
print(result)
13. 初始化数字人声音克隆服务 【使用此函数需要先调用 init_speech_clone_service 初始化克隆语音服务】
def metahuman_speech_clone(self, save_audio_path: str, text: str, language: str = "zh-cn", wait_play_sound: bool = True) -》 bool:
"""
数字人使用克隆声音说话，此函数需要调用 initSpeechCloneService 初始化语音服务
Generate digital human speech file (voice clone), and automatically generate MP3 file and lab file
save_audio_path: 保存的发音文件路径, 这里是文件路径，不是目录！
text: 要转换语音的文本
language: 语言，中文：zh-cn，其他语言：other-languages
wait_play_sound: 等待音频播报完毕，默认为 true等待
return: True或者False
save_audio_path: the path of the saved pronunciation file, here is the file path, not the directory!
text: the text to be converted into speech
language: language, Chinese: zh-cn, other languages: other-languages
wait_play_sound: Wait until the audio broadcast is finished; the default value is true
return: True or False
"""
# 使用示例 [Demo]
result = self.metahuman_speech_clone("D:/AiboteMetahuman/voice/1.mp3", "Aibote is an excellent automation framework", "zh-cn", True)
print(result)
14. 使用克隆声音生成数字人短视频 【使用此函数需要先调用 init_speech_clone_service 初始化克隆语音服务】
def make_metahuman_video_clone(self, save_video_folder: str, text: str, language: str = "zh-cn", bg_file_path: str, sim_value: int = 0) -》 bool:
"""
使用克隆声音生成数字人短视频，此函数需要调用 initSpeechCloneService 初始化语音服务
Using cloned voice to generate short video of digital people, this function needs to call initSpeechCloneService to initialize voice service
save_video_folder: 保存的视频和音频文件目录
text: 要转换语音的文本
language: 语言，语言，中文：zh-cn，其他语言：other-languages
bg_file_path: 数字人背景 图片/视频 路径，扣除绿幕会自动获取绿幕的RGB值，null 则不替换背景。仅替换绿幕背景的数字人模型
sim_value: 相似度，默认为0。此处参数用作绿幕扣除微调RBG值。取值应当大于等于0
return: True或者False
save_video_folder: directory of saved video and audio files
text: the text to be converted into speech
language: language, language, Chinese: zh-cn, other languages: other-languages
bg_file_path: the background picture/video path of the digital person. If the green screen is subtracted, the RGB value of the green screen will be automatically obtained; if it is null,
the background will not be replaced. Digital human model only replacing green screen background
sim_value: similarity, which defaults to 0. Here, the parameter is used as the RBG value of green screen subtraction fine tuning. Value should be greater than or equal to 0
return: True or False
"""
# 使用示例 [Demo]
result = self.make_metahuman_video_clone("d:", "Aibote数字人生成短视频", "zh-cn", "", 0)
print(result)
15. 初始化数字人声音克隆服务 【使用此函数需要先调用 init_speech_clone_service 初始化克隆语音服务】
def make_metahuman_speech_file_clone(self, save_audio_path: str, text: str, language: str = "zh-cn") -》 bool:
"""
生成数字人说话文件(声音克隆)，生成MP3文件和 lab文件，提供给 metahumanSpeechByFile 和使用
Generate digital human speech file (voice clone), MP3 file and lab file, and provide them to metahumanSpeechByFile and use it
save_audio_path: 保存的发音文件路径。这里是路径，不是目录！
text: 要转换语音的文本
language: 语言，中文：zh-cn，其他语言：other-languages
return: True或者False
save_audio_path: the path of the saved pronunciation file. This is a path, not a directory!
text: the text to be converted into speech
language: language, Chinese: zh-cn, other languages: other-languages
return: True or False
"""
# 使用示例 [Demo]
result = self.make_metahuman_speech_file_clone("D:/AiboteMetahuman/voice/1.mp3", "Aibote is an excellent automation framework", "zh-cn")
print(result)
16. 切换新的人物形象动作 【使用此函数需要先调用 init_speech_clone_service 初始化克隆语音服务】
def switch_action(self, call_apiKey: str, action_video_or_image: str, user_simValue: bool) -> bool:
"""
切换新的人物形象动作，此函数无需训练数字人模型，直接切换各种人物形象动作和场景
Switch new character movements. This function can directly switch various character movements and scenes without training digital human models
call_apiKey: 调用函数的密钥
action_video_or_image: 闭嘴的人物视频或者图片
user_simValue: 是否使用近似值加速(影响精度)，默认不加速
return: True或者False
call_apiKey: String, the key of the calling function
action_video_or_image: a string, shut-up character video or picture
return: True or False
"""
# 使用示例 [Demo]
result = self.switch_action("dfjklDJFLJlfjkdljf", "d:/1.png", False)
print(result)
17. 训练数字人 【使用此函数需要先调用 init_speech_clone_service 初始化克隆语音服务】
def train_human_model(self, call_apiKey: str, train_video_or_image: str, src_metahuman_model_path: str, save_human_model_folder: str) -》 bool:
"""
训练数字人，训练时长为10-30分钟
Train digital people for 10-30 minutes
call_apiKey: 调用函数的密钥
train_video_or_image: 闭嘴的人物视频或者图片 素材
src_metahuman_model_path: 预训练数字人模型路径
save_human_model_folder: 保存训练完成的模型目录
return: True或者False
call_apiKey: the key of the calling function
train_video_or_image: shut up the character video or picture material
src_metahuman_model_path: the path of pre-training digital human model
save_human_model_folder: save the model directory after training
return: True or False
"""
# 使用示例 [Demo]
result = self.train_human_model("dfjklDJFLJlfjkdljf", "d:/1.png", "d:/humanModel", "d:/newHumanModel")
print(result)
18. 通过音频文件生成数字人短视频 【使用此函数依赖 initMetahuman 函数运行, 否则程序会崩溃
def make_metahuman_video_by_file(self, audio_path: str, bg_file_path: str, sim_value: str = 0) -》 bool:
"""
通过音频文件生成数字人短视频， 此函数依赖 initMetahuman 函数运行，否则程序会崩溃
Switch new character movements. This function can directly switch various character movements and scenes without training digital human models
audio_path: 字符串型，音频路径， 同名的 .lab文件需要和音频文件在同一目录下
bg_file_path: 字符串型，数字人背景 图片/视频 路径，扣除绿幕会自动获取绿幕的RGB值，null 则不替换背景。仅替换绿幕背景的数字人模型
sim_value: 整型，相似度，默认为0。此处参数用作绿幕扣除微调RBG值。取值应当大于等于0
return: True or False
audio_path: String type, audio path, and. lab file with the same name need to be in the same directory as the audio file
bg_file_path: String type, background picture/video path of digital person. If the green screen is subtracted, the RGB value of the green screen will be automatically obtained,
and null will not replace the background. Digital human model only replacing green screen background
sim_value: Integer, similarity, default is 0. Here, the parameter is used as the RBG value of green screen subtraction fine tuning
Value should be greater than or equal to 0
return: True or False
"""
# 使用示例 [Demo]
result = self.make_metahuman_video_by_file("D:/AiboteMetahuman/voice/1.mp3", "", 0)
print(result)
def get_switch_action_state(self) -》 bool:
"""
获取切换人物形象动作状态
Obtain that action state of switching the character image
return: True或者False
return: True or False
"""
# 使用示例 [Demo]
result = self.get_switch_action_state()
print(result)
def make_clone_audio(self, clone_server_ip: str, save_audio_path: str, refer_audio_path: str, refer_text: str, clone_text: str, speed_factor: float) -》 bool:
"""
克隆声音，需要部署服务端
To clone sound, you need to deploy the server.
clone_server_ip: 克隆声音服务端IP
save_audio_path: 保存克隆声音的路径
refer_audio_path: 参考音频路径，10-40秒，推荐25秒左右的参考音频
refer_text: 参考音频对应的文本
clone_text: 要克隆的文本
speed_factor: 语速（0.5为半速，1.0为正常速度，1.5为1.5倍速，以此类推）。默认为1.0 正常语速
return: True或者False
clone_server_ip: clone voice server IP
save_audio_path: the path to save the cloned sound
refer_audio_path: Reference audio path, 10-40 seconds, 25 seconds is recommended
refer_text: the text corresponding to the reference audio
clone_text: the text to be cloned
speed_factor: speech speed (0.5 is half speed, 1.0 is normal speed, 1.5 is 1.5 times speed, and so on). The default is 1.0 normal speech speed
return: True or False
"""
# 使用示例 [Demo]
result = self.make_clone_audio("127.0.0.1", "d:/save.wav", "d:/refer.wav", "人生苦短我用PyAibote", "生成的音频文本内容", 1.0)
print(result)
def make_clone_lab(self, lab_server_ip: str, audio_path: str) -》 bool:
"""
生成lab文件，需要部署服务端
To generate lab files, you need to deploy the server.
lab_server_ip: 保存的发音文件路径。这里是路径，不是目录！
audio_path: 要转换语音的文本
return: True或者False
lab_server_ip: saved pronunciation file path. This is a path, not a directory
audio_path: the text to be converted into speech
return: True or False
"""
# 使用示例 [Demo]
result = self.make_clone_lab("127.0.0.1", "d:/save.wav")
print(result)
def clone_audio_to_text(self, lab_server_ip: str, audio_path: str) -》 bool:
"""
语音识别，需要部署服务端
Speech recognition requires the deployment of the server.
lab_server_ip: lab服务端IP
audio_path: 音频文件
return: 失败返回None, 成功返回识别到的内容
lab_server_ip: lab server IP
audio_path: audio file
return: None is returned in case of failure, and the recognized content is returned successfully
"""
# 使用示例 [Demo]
result = self.clone_audio_to_text("127.0.0.1", "d:/save.wav")
print(result)
def text_to_audio_and_lab_file(self, save_audio_path: str, text: str, language: str = "zh-cn", voice_name: str = "", quality: int = 0, speech_rate: int = 0, voice_style: str = "General") -》 bool:
"""
文本合成语音保存到本地音频文件
Text synthesis speech is saved to a local audio file
save_audio_path: 保存的音频文件路径，扩展为.MP3格式。同名的 .lab文件需要和音频文件在同一目录下
text: 要转换语音的文本
language: 语言，参考开发文档 语言和发音人
voice_name: 发音人，参考开发文档 语言和发音人
quality: 音质，0低品质 1中品质 2高品质， 默认为0低品质
speech_rate: 语速，默认为0，取值范围 -100 至 200
voice_style: 语音风格，默认General常规风格，其他风格参考开发文档 语言和发音人
return: True或者False
save_audio_path: the path of the saved audio file, expanded to .mp3 format. The. lab file with the same name needs to be in the same directory as the audio file
text: the text to be converted into speech
language: language, reference development document language and speaker
voice_name: speaker, refer to the development document language and speaker
quality: sound quality, 0 low quality 1 medium quality 2 high quality, and 0 low quality by default
speech_rate: speech speed, which is 0 by default, and the value range is -100 to 200
voice_style: voice style, the default General style, and other styles refer to the development document language and speaker
return: True or False
"""
# 使用示例 [Demo]
result = self.text_to_audio_and_lab_file("D:/AiboteMetahuman/voice/1.mp3", "PyAibote is an excellent automation framework", "zh-cn", "zh-cn-XiaochenNeural", 0, 0, "General")
print(result)
def switch_clone_audio_model(self, clone_server_ip: str, gpt_weights_path: str, sovits_weights_path: str) -》 bool:
"""
切换声音克隆模型，需要部署服务端(切换到与原模型无关音色的模型，切记更换参考音频和文本)
To switch the voice cloning model, you need to deploy the server
clone_server_ip: 克隆声音服务端
gpt_weights_path: gpt 模型权重路径。指克隆服务所在的电脑/服务器 路径
sovits_weights_path: sovits 模型权重路径。指克隆服务所在的电脑/服务器 路径
return: 成功返回True，失败返回False
clone_server_ip: clone the sound server
gpt _ weights _ path: the weight path of GPT model. Refers to the computer/server path where the cloning service is located
sovits _ weights _ path: the weight path of Sovits model. Refers to the computer/server path where the cloning service is located
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.switch_clone_audio_model("127.0.0.1", "d:/gptWeightsPath.ckpt", "d:/sovitsWeightsPath.pth")
print(result)
def restart_clone_audio_server(self, clone_server_ip: str) -》 bool:
"""
重启声音克隆服务，需要部署服务端
To restart the voice cloning service, the server needs to be deployed
clone_server_ip: 克隆声音服务端
return: 成功返回True，失败返回False
clone_server_ip: clone the sound server
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.restart_clone_audio_server("127.0.0.1")
print(result)
def play_audio(self, audio_path: str, is_wait: bool = True) -》 bool:
"""
播报音频文件 不需要lab文件
Broadcast audio file
audio_path: 音频文件路径
is_wait: 是否等待，为True时，等待播放完毕
return: 成功返回True，失败返回False
audio_path: audio file path
is_wait: whether to wait, if True, wait for the playback to finish
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.play_audio("D:/Save.wav", True)
print(result)
def play_media(self, video_path: str, video_sacle: float = 1.0, isLoop_play: bool = False, enable_random_param: bool = False, is_wait: bool = True) -》 bool:
"""
播报视频文件 (多个视频切换播放 视频和音频编码必须一致)
Broadcast video frequency files (video and audio coding must be consistent when playing multiple videos)
video_path: 视频文件路径
video_sacle: 视频缩放（0.5缩小一半，1.0为原始大小）
isLoop_play: 是否循环播放
enable_random_param: 是否启用随机去重参数
is_wait: 是否等待播报完毕。值为false时，不等待播放结束。未播报结束前再次调用此函数 会终止前面的播报内容
return: 成功返回True，失败返回False
video_path: video file path
video_sacle: video scaling (0.5 is reduced by half, 1.0 is the original size)
isLoop_play: whether to loop or not
enable_random_param: whether to enable the random deduplication parameter
is_wait: whether to wait for the broadcast to finish. When the value is false, do not wait for the end of playback.
Calling this function again before the broadcast ends will terminate the previous broadcast content
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.play_media("D:/test.mp4", 1.0, False, False, True)
print(result)
def extract_audio(self, video_path: str) -》 bool:
"""
复制视频中的音频为mp3文件
Copy the audio in the video as an mp3 file.
video_path: 视频文件路径 音频文件保存为mp3和视频在一个目录
return: 成功返回True，失败返回False
video_path: video file path. Audio files are saved as mp3 and videos are in one directory
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.extract_audio("D:/test.mp4")
print(result)
def set_media_volume_scale(self, volume_scale: float) -》 bool:
"""
调节 playMedia 音量大小(底层用的内存共享，支持多进程控制)
Adjust the volume of playMedia (memory sharing for the bottom layer, supporting multi-process control)
volume_scale: 浮点型，音量缩放（0.5调低一半，1.0为原始音量大小）。默认为原始大小
return: 总是返回True
volume_scale: floating-point, volume scaling (0.5 is reduced by half, 1.0 is the original volume). Default is the original size
return: always returns True
"""
# 使用示例 [Demo]
result = self.set_media_volume_scale(1.0)
print(result)
def metahuman_speech_by_file_ex(self, audio_path: str, enable_random_param: bool , wait_play_sound: bool = True) -》 bool:
"""
数字人说话文件缓存模式(Ex)，不能与 PlayAudioEx 同步执行
Digital human speech file cache mode (Ex) cannot be synchronized with PlayAudioEx
audio_path: 音频路径， 同名的 .lab文件需要和音频文件在同一目录下。若.lab文件不存在，则自动生成.lab文件。生成.lab文件产生的费用，请联系管理员
enable_random_param: 是否启用随机去重参数
wait_play_sound : 是否等待播报完毕，默认为True 等待。为False时 多次调用此函数会添加到队列按顺序播报
return: 成功返回True，失败返回False
audio_path: the audio path. The. lab file with the same name needs to be in the same directory as the audio file. If the. lab file does not exist, the. lab file is automatically generated. Please contact the administrator for the expenses incurred by generating the. lab file
enable_random_param: whether to enable the random deduplication parameter
wait_play_sound: whether to wait for the broadcast, the default value is True. When False, calling this function multiple times will be added to the queue to broadcast in sequence
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.metahuman_speech_by_file_ex("D:/AiboteMetahuman/voice/1.mp3", True, False)
print(result)
def play_audio_ex(self, audio_path, enable_random_param, is_wait) -》 bool:
"""
播报音频文件(EX)，不能与 metahumanSpeechByFileEx 同步执行
Broadcast audio file (EX), which can't be executed synchronously with metahumanSpeechByFileEx
audio_path: 音频文件路径
enable_random_param: 是否启用随机去重参数
is_wait : 是否等待。为true时,等待播放完毕
return: 总是返回true，函数仅添加播放音频文件到队列不处理返回
audio_path: the audio path. The. lab file with the same name needs to be in the same directory as the audio file. If the. lab file does not exist, the. lab file is automatically generated. Please contact the administrator for the expenses incurred by generating the. lab file
enable_random_param: whether to enable the random deduplication parameter
wait_play_sound: whether to wait for the broadcast, the default value is True. When False, calling this function multiple times will be added to the queue to broadcast in sequence
return: Returns True on success and False on failure
"""
# 使用示例 [Demo]
result = self.play_audio_ex("D:/save.wav", True, False)
print(result)