Spaces:
Running
Running
| import gradio as gr | |
| import edge_tts | |
| import asyncio | |
| import tempfile | |
| from moviepy.editor import AudioFileClip | |
| import re | |
| async def get_voices(): | |
| voices = await edge_tts.list_voices() | |
| return {f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v['ShortName'] for v in voices} | |
| async def text_to_speech(text, voice, rate, pitch): | |
| if not text.strip(): | |
| return None, gr.Warning("Please enter the text to convert.") | |
| if not voice: | |
| return None, gr.Warning("Please select a voice.") | |
| voice_short_name = voice.split(" - ")[0] | |
| rate_str = f"{rate:+d}%" | |
| pitch_str = f"{pitch:+d}Hz" | |
| communicate = edge_tts.Communicate(text, voice_short_name, rate=rate_str, pitch=pitch_str) | |
| with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file: | |
| tmp_path = tmp_file.name | |
| await communicate.save(tmp_path) | |
| return tmp_path, None | |
| def generate_srt(text, audio_duration, max_words_per_line): | |
| # Eliminate extra spaces and split into words | |
| text = re.sub(r'\s+', ' ', text.strip()) | |
| words = text.split() | |
| srt_lines = [] | |
| current_line = [] | |
| total_words = len(words) | |
| for i, word in enumerate(words): | |
| current_line.append(word) | |
| # Create a line if we reach the max words per line or at the end of the text | |
| if len(current_line) >= max_words_per_line or i == total_words - 1: | |
| line_text = ' '.join(current_line) | |
| # Adjust duration proportionally | |
| duration = audio_duration * (len(current_line) / total_words) | |
| start_time = (sum(len(' '.join(srt_lines[j].split()[2:])) for j in range(len(srt_lines))) / total_words) * audio_duration if srt_lines else 0 | |
| end_time = start_time + duration | |
| # Formatting time for SRT | |
| start_time_str = f"{int(start_time // 3600):02}:{int((start_time % 3600) // 60):02}:{int(start_time % 60):02},{int((start_time % 1) * 1000):03}" | |
| end_time_str = f"{int(end_time // 3600):02}:{int((end_time % 3600) // 60):02}:{int(end_time % 60):02},{int((end_time % 1) * 1000):03}" | |
| srt_lines.append(f"{len(srt_lines) + 1}\n{start_time_str} --> {end_time_str}\n{line_text}\n") | |
| current_line = [] | |
| return ''.join(srt_lines) | |
| def tts_interface(text, voice, rate, pitch, max_words_per_line): | |
| audio_path, warning = asyncio.run(text_to_speech(text, voice, rate, pitch)) | |
| if warning: | |
| return None, None, warning | |
| # Calculate audio duration | |
| audio_duration = AudioFileClip(audio_path).duration # Get duration in seconds | |
| # Generate SRT file | |
| srt_content = generate_srt(text, audio_duration, max_words_per_line) | |
| srt_path = audio_path.replace('.mp3', '_subtitle.srt') | |
| with open(srt_path, 'w') as f: | |
| f.write(srt_content) | |
| return audio_path, srt_path, None | |
| async def create_demo(): | |
| voices = await get_voices() | |
| demo = gr.Interface( | |
| fn=tts_interface, | |
| inputs=[ | |
| gr.Textbox(label="Input Text", lines=5), | |
| gr.Dropdown(choices=[""] + list(voices.keys()), label="Select Voice", value=""), | |
| gr.Slider(minimum=-50, maximum=50, value=0, label="Rate Adjustment (%)", step=1), | |
| gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1), | |
| gr.Slider(minimum=3, maximum=8, value=5, label="Max Words per Line", step=1), | |
| ], | |
| outputs=[ | |
| gr.Audio(label="Generated Audio", type="filepath"), | |
| gr.File(label="Generated Subtitle (.srt)"), | |
| gr.Markdown(label="Warning", visible=False) | |
| ], | |
| title="Edge TTS Text to Speech with SRT", | |
| description="Convert text to speech and generate synchronized subtitles based on speech rate.", | |
| analytics_enabled=False, | |
| allow_flagging=False, | |
| ) | |
| return demo | |
| if __name__ == "__main__": | |
| demo = asyncio.run(create_demo()) | |
| demo.launch() | |