def init_new_metahuman(self, model_folder: str, scale: int, is_paly_media_audio: bool, enable_random_param: bool, push_stream_url: bool) -》 bool:
"""
初始化数字人
Initialize digital person
model_folder: 模型文件夹
scale: 缩放，1原始大小，0.5缩小一半
is_paly_media_audio: 是否播放素材中的音频
enable_random_param: 是否开启随机参数，随机内容：音频随机(不含数字人说话声音随机) 和 画面随机(包含平移、旋转、缩放、动作泛化)
push_stream_url: 推流地址，默认为 "" 不推流，仅支持rtmp，例如： rtmp://your_server_ip/live/stream
return: 成功返回True，失败返回错误信息
model_folder: model folder
scale: scaling, 1 original size, 0.5 reduced by half
is_paly_media_audio: whether to play the audio in the material
Enable_random_param: whether to turn on the random parameter, including audio random (excluding digital human voice random)
and picture random (including translation, rotation, scaling and motion generalization).
push_stream_url: push stream address; the default value is "",and only rtmp is supported; for example, RTMP://your _ server _ IP/live/stream
return: Returns True on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.init_new_metahuman(r"D:\Project\Aibote\NewHuman\model", 0.5, False, False, "")
print(result)
# Python拉流示例 [Demo]
import cv2
# 推流地址
rtmp_url = "rtmp://127.0.0.1/live/stream"
# 创建Video对象
cap = cv2.VideoCapture(rtmp_url)
# 检查是否成功打开流
if not cap.isOpened():
exit()
# 读取并显示视频帧
while True:
ret, frame = cap.read()
if not ret:
break
cv2.imshow('RTMP Stream', frame)
if cv2.waitKey(1) or 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
def new_metahuman_switch_action(self, figure_video_path: str, scale: int, is_paly_media_audio: bool, is_swap_color: bool) -》 bool:
"""
切换人物形象动作 (使用前需要先调用 init_new_metahuman 初始化数字人)
Switch character image action
figure_video_path: 人物视频路径
scale: 缩放，1原始大小，0.5缩小一半
is_paly_media_audio: 是否播放素材中的音频
is_swap_color: 是否更换基础嘴型颜色
return: 成功返回True，失败返回错误信息
figure_video_path: model folder
scale: scaling, 1 original size, 0.5 reduced by half
is_paly_media_audio: whether to play the audio in the material
is_swap_color: Do you want to change the basic mouth color
return: Returns True on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.new_metahuman_switch_action(r"D:\Project\666.mp4", 0.5, False, False)
print(result)
def new_metahuman_add_background(self, bg_path: str) -》 bool:
"""
添加视频/图片背景 (使用前需要先调用 init_new_metahuman 初始化数字人)
Add video/picture background
bg_path: 背景视频/图片路径
return: 成功返回True，失败返回错误信息
bg_path: background video/picture path
return: Returns True on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.new_metahuman_add_background(r"D:\Project\1.mp4")
print(result)
def new_metahuman_del_background(self) -》 bool:
"""
删除视频/图片背景
Delete video/picture background
return: 成功返回True，失败返回错误信息
return: Returns True on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.new_metahuman_del_background()
print(result)
def new_metahuman_generate_human_video(self, audio_path: str) -》 bool:
"""
生成数字人视频
Generate digital human video
audio_path: 音频路径，需要提前生成lab文件
return: 失败返回错误信息，成功返回true，并在audio_path同级目录下生成同名 .mp4 后缀的数字人视频文件
audio_path: audio path. lab files need to be generated in advance.
return: an error message is returned in case of failure, and true is returned in case of success, and a digital human video file with the same name and .mp4 suffix is generated in the same level directory of audio_path
"""
# 使用示例 [Demo]
result = self.new_metahuman_generate_human_video(r"D:\Project\888.wav")
print(result)
def new_metahuman_audio_to_lab(self, server_ip: str, audio_path: str) -》 bool:
"""
生成lab文件
Generate lab file
server_ip: lab服务端IP
audio_path: 音频文件路径
return: 失败返回错误信息，成功返回true，并在audioPath同级目录下生成同名 .lab 后缀的lab文件
server_ip: lab server IP
audio_path: audio file path
return: failure returns an error message, success returns true, and a lab file with the same name. lab suffix is generated in the same directory of audioPath
"""
# 使用示例 [Demo]
result = self.new_metahuman_audio_to_lab("127.0.0.1", r"D:\Project\888.wav")
print(result)
def new_metahuman_text_to_audio(self, server_ip: str, save_audio_path: str, refer_audio_path: str, refer_text: str, text : str, speed_factor : int) -》 bool:
"""
文本转音频
Text to audio
server_ip: 声音克隆服务端IP
save_audio_path: 保存的音频文件路径
refer_audio_path: 参考音频路径
refer_text: 参考文本内容(内容必须与参考音频播报的内容一致)
text: 要合成的文本内容
speed_factor: 语速，默认为1正常语速
return: 失败返回错误信息，成功返回true
server_ip: voice cloning server IP
save_audio_path: the path of the saved audio file
reference _ audio _ path: the refer_audio_path
refer_text: Reference text content (the content must be consistent with the content broadcast by reference audio)
text: the text content to be synthesized
speed_factor: speech speed, which defaults to 1 normal speech speed
return: failure returns an error message, and success returns true
"""
# 使用示例 [Demo]
result = self.new_metahuman_text_to_audio("127.0.0.1", r"D:\Project\888.wav",r"D:\Project\666.wav", "我是参考音频中的文本内容", "我是一个实际生成的文本内容", 1)
print(result)
def new_metahuman_audio_to_text(self, server_ip: str, audio_path: str) -》 bool:
"""
语音转文本
Speech to text
server_ip: 语音识别服务端IP
audio_path: 音频文件路径
return: 失败返回"None"，成功返回语音识别的文本内容
server_ip: IP of speech recognition server
audio_path: audio file path
return: "None" is returned in case of failure, and the text content of speech recognition is returned successfully
"""
# 使用示例 [Demo]
result = self.new_metahuman_audio_to_text("127.0.0.1", r"D:\Project\888.wav")
print(result)
def new_metahuman_human_speak(self, audio_path: str, wait_play_sound: bool) -》 bool:
"""
数字人说话
Digital people speak
audio_path: 音频路径，需要提前生成lab文件
wait_play_sound: 是否等待播报完毕
return: 成功返回true，失败返回错误信息
audio_path: audio path. lab files need to be generated in advance
wait_play_sound: Do you want to wait for the broadcast to finish
return: Returns true on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.new_metahuman_human_speak(r"D:\Project\888.wav", True)
print(result)
def new_metahuman_stop_speak(self) -》 bool:
"""
打断说话
Interrupt speech
return: 返回true
return: Returns true
"""
# 使用示例 [Demo]
result = self.new_metahuman_stop_speak()
print(result)
def new_metahuman_start_record(self, save_audio_path) -》 bool:
"""
录制麦克风
Recording microphone
save_audio_path: 录制保存的音频路径，需要调用 stopRecord 结束录制并保存
return: 返回true
save_audio_path: record the saved audio path. You need to call stopRecord to end the recording and save it
return: returns true
"""
# 使用示例 [Demo]
result = self.new_metahuman_start_record(r"D:\Project\888.wav")
print(result)
def new_metahuman_stop_record(self) -》 bool:
"""
停止录制
Stop recording
return: 返回true
return: returns true
"""
# 使用示例 [Demo]
result = self.new_metahuman_stop_record()
print(result)
def new_get_face_data(self, server_ip: str, call_api_key: str, video_path: str) -》 bool:
"""
获取脸部数据
Get face data
server_ip: 服务端IP
call_api_key: 调用密钥, 密钥和IP获取联系QQ：2766463939
video_path: 人脸视频路径
return: 失败返回错误信息，成功返回true。并在videoPath同级目录下生成.pt 后缀的人脸数据文件
server_ip: server IP
call_api_key: call key
video_path: face video path
return: failure returns an error message, and success returns true. And generate a face data file with. pt suffix in the same level directory of videoPath
"""
# 使用示例 [Demo]
result = self.new_get_face_data("82.68.89.156", "yNjEtNDM5Mi04OGZmLTU2NzBmYjgwYTAxMAACKlszLCIwZDYyNTc0MC0yN", r"D:\Project\888.mp4")
print(result)
def new_metahuman_generate_human_video_ex(self, server_ip: str, call_api_key: str, audio_path: str, video_path: str, save_video_path: str) -》 bool:
"""
云端算力生成短视频
Cloud computing generates short videos
server_ip: 服务端IP
call_api_key: 调用密钥, 密钥和IP获取联系QQ：2766463939
audio_path: 驱动口型的音频路径
video_path: 原视频路径
save_video_path: 保存的合成结果路径
return: 失败返回错误信息，成功返回true
server_ip: server IP
call_api_key: call key
audio_path: the audio path of the driving mouth
video_path: original video path
save_video_path: the saved composition result path
return: failure returns an error message, and success returns true
"""
# 使用示例 [Demo]
result = self.new_metahuman_generate_human_video_ex("82.68.89.156", "yNjEtNDM5Mi04OGZmLTU2NzBmYjgwYTAxMAACKlszLCIwZDYyNTc0MC0yN", r"D:\Project\888.wav", r"D:\Project\888.mp4", r"D:\Project")
print(result)
def new_metahuman_train_voice_ex(self, appid: str, token: str, spk_id: str, refer_audio_path: str) -》 bool:
"""
云端算力训练声音
Cloud computing power training voice
云端算力密钥获取地址： https://console.volcengine.com/speech/service/9999?AppID=1330699505
Cloud computing key acquisition address: https://console.volcengine.com/speech/service/9999?AppID=1330699505
appid: APP ID
token: Access Token
spk_id: 声音ID
refer_audio_path: 参考音频
return: 训练成功返回true，失败返回错误信息
appid: APP ID
token: Access Token
spk_ID: sound id
reference_audio_path: reference audio
return: training success returns true, failure returns an error message
"""
# 使用示例 [Demo]
result = self.new_metahuman_train_voice_ex("1330149505", "jorvx09WGQDYCJFIOalT42TeYlkSEY7A", "S_HRboGJG2r1", r"D:\Project\888.mp4")
print(result)
def new_metahuman_get_train_status_ex(self, appid: str, token: str, spk_id: str) -》 str:
"""
获取 trainVoiceEx 训练状态
Get the trainVoiceEx training status
云端算力密钥获取地址： https://console.volcengine.com/speech/service/9999?AppID=1330699505
Cloud computing key acquisition address: https://console.volcengine.com/speech/service/9999?AppID=1330699505
appid: APP ID
token: Access Token
spk_id: 声音ID
return: 返回训练状态"Train Success"、"NotFound"、"Training"、"Failed"、unknow"
appid: APP ID
token: Access Token
spk_ID: sound id
return: returns to the Training status of "Train Success", "NotFound", "training", "Failed" and "unknow"
"""
# 使用示例 [Demo]
result = self.new_metahuman_get_train_status_ex("1330149505", "jorvx09WGQDYCJFIOalT42TeYlkSEY7A", "S_HRboGJG2r1")
print(result)
def new_metahuman_text_to_audio_ex(self, appid: str, token: str, spk_id: str, cluster: str, text: str, speed_ratio: str, save_audio_path: str) -》 bool:
"""
云端算力文本转语音
Cloud computing power text-to-speech
云端算力密钥获取地址： https://console.volcengine.com/speech/service/9999?AppID=1330699505
Cloud computing key acquisition address: https://console.volcengine.com/speech/service/9999?AppID=1330699505
appid: APP ID
token: Access Token
spk_id: 声音ID
cluster: Cluster ID 声音复刻大模型："volcano_icl" 语音合成大模型："volcano_tts"
text: 合成的文本
speed_ratio: 语速，正常为1
save_audio_path: 保存的音频路径， 同时会在音频同目录下生成lab文件
return: 成功返回true，失败返回错误信息
appid: APP ID
token: Access Token
spk_ID: sound id
cluster: Cluster ID sound reproduction model: "volcano_icl" speech synthesis model: "volcano_tts"
text: synthesized text
speed_ratio: speech speed, which is normally 1
save_audio_path: the saved audio path, and a lab file will be generated in the same audio directory
return: Returns true on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.new_metahuman_text_to_audio_ex("1330149505", "jorvx09WGQDYCJFIOalT42TeYlkSEY7A", "S_HRboGJG2r1", "volcano_icl", "Aibote数字人CPU实时推理嘴型全球第一", 1, r"D:\Project\888.wav")
print(result)
def new_metahuman_get_extend_param(self) -》 str:
"""
获取驱动程序命令行参数(不包含ip和port)
Get the driver command line parameters (excluding ip and port)
return: 成功返回参数，失败返回None
return: parameter is returned successfully, and None is returned on failure.
"""
# 使用示例 [Demo]
result = self.new_metahuman_get_extend_param()
print(result)
def new_metahuman_get_driver_folder(self) -》 str:
"""
获取驱动程序 所在的文件夹路径
Gets the folder path where the driver is located
return: 返回AiDriver.exe所在的文件夹路径
return: Returns the folder path where AiDriver.exe is located
"""
# 使用示例 [Demo]
result = self.new_metahuman_get_driver_folder()
print(result)
def new_metahuman_close_driver(self) -》 str:
"""
关闭驱动
Turn off drive
return: 返回true
return: Return true
"""
# 使用示例 [Demo]
result = self.new_metahuman_close_driver()
print(result)
def new_metahuman_insert_video(self, video_path: str) -》 bool:
"""
插入特写视频
Insert close-up video
video_path: 要播放的视频
return: 成功返回true，失败返回错误信息
video_path: the video to play
return: Returns true on success, and returns an error message on failure
"""
# 使用示例 [Demo]
result = self.new_metahuman_insert_video(r"D:\Project\888.mp4")
print(result)