Skip to content

利用 AI 大模型和自动化工作流,根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.

feat: add no voice synthesis mode

harry070376d7e097bc71
diff --git a/app/services/video.py b/app/services/video.py
index d30bc2c3..eeece3d0 100644
--- a/app/services/video.py
+++ b/app/services/video.py
@@ -124,26 +124,14 @@ def _prioritize_unique_source_clips(
def get_ffmpeg_binary():
- # 优先复用用户在 config.toml / 环境变量里显式指定的 ffmpeg,可避免
- # Windows 便携包、Docker、自定义安装目录等场景下 PATH 不一致。
- configured_ffmpeg = os.environ.get("IMAGEIO_FFMPEG_EXE")
- if configured_ffmpeg:
- return configured_ffmpeg
-
- system_ffmpeg = shutil.which("ffmpeg")
- if system_ffmpeg:
- return system_ffmpeg
-
- try:
- import imageio_ffmpeg
-
- bundled_ffmpeg = imageio_ffmpeg.get_ffmpeg_exe()
- if bundled_ffmpeg:
- return bundled_ffmpeg
- except Exception as exc:
- logger.warning(f"failed to resolve bundled ffmpeg binary: {str(exc)}")
+ """
+ 兼容历史上直接从 video 服务读取 FFmpeg 路径的调用方。
- return "ffmpeg"
+ 真正的解析逻辑已经抽到 `app.utils.utils.get_ffmpeg_binary()`,视频、语音
+ 和后续新增链路都应复用同一套优先级;这里保留薄包装,避免外部脚本或
+ 旧测试直接导入 `app.services.video.get_ffmpeg_binary` 时出现 AttributeError。
+ """
+ return utils.get_ffmpeg_binary()
def _get_configured_video_codec() -> str:
@@ -216,7 +204,7 @@ def _get_effective_video_codec(preferred_codec: str | None = None) -> str:
)
return _DEFAULT_VIDEO_CODEC
- ffmpeg_binary = get_ffmpeg_binary()
+ ffmpeg_binary = utils.get_ffmpeg_binary()
if not _ffmpeg_encoder_exists(ffmpeg_binary, selected_codec):
logger.warning(
f"ffmpeg encoder {selected_codec} is not available, "
@@ -300,7 +288,7 @@ def concat_video_clips_with_ffmpeg(
def build_command(codec: str) -> list[str]:
return [
- get_ffmpeg_binary(),
+ utils.get_ffmpeg_binary(),
"-y",
"-f",
"concat",
diff --git a/app/services/voice.py b/app/services/voice.py
index 11d14fc6..be9bcb62 100644
--- a/app/services/voice.py
+++ b/app/services/voice.py
@@ -7,9 +7,10 @@ import math
import os
import queue
import re
-import shutil
+import subprocess
import threading
import time
+import unicodedata
from datetime import datetime
from typing import Union
from xml.sax.saxutils import unescape
@@ -28,18 +29,15 @@ from app.utils import utils
_DEFAULT_EDGE_TTS_TIMEOUT_SECONDS = 30.0
_MIMO_DEFAULT_BASE_URL = "https://api.xiaomimimo.com/v1"
_MIMO_DEFAULT_TTS_MODEL = "mimo-v2.5-tts"
+NO_VOICE_NAME = "no-voice"
+# `none` 是 PR #981 里曾使用过的无配音标识。这里短期兼容这个值,避免
+# 已经手动调用过该分支的 API 用户升级后立即失效;WebUI 和新代码统一使用
+# 更明确的 `no-voice`。
+_NO_VOICE_ALIASES = {NO_VOICE_NAME, "none"}
def _configure_pydub_ffmpeg(audio_segment_cls):
- configured_ffmpeg = os.environ.get("IMAGEIO_FFMPEG_EXE") or shutil.which("ffmpeg")
- if not configured_ffmpeg:
- try:
- import imageio_ffmpeg
-
- configured_ffmpeg = imageio_ffmpeg.get_ffmpeg_exe()
- except Exception as exc:
- logger.warning(f"failed to resolve bundled ffmpeg binary: {str(exc)}")
-
+ configured_ffmpeg = utils.get_ffmpeg_binary()
if configured_ffmpeg:
audio_segment_cls.converter = configured_ffmpeg
@@ -202,6 +200,104 @@ def is_mimo_voice(voice_name: str):
return voice_name.startswith("mimo:")
+def is_no_voice(voice_name: str | None) -> bool:
+ """
+ 判断用户是否明确选择了“无配音”模式。
+
+ 这里刻意不把空字符串当成无配音:空 voice 更可能是配置损坏、旧版本
+ WebUI 状态丢失或接口参数缺失。只有明确的 sentinel 才进入静音分支,
+ 这样可以避免把真实错误伪装成正常生成。
+ """
+ return str(voice_name or "").strip().lower() in _NO_VOICE_ALIASES
+
+
+def estimate_no_voice_duration(text: str) -> float:
+ """
+ 为无配音模式估算一个稳定的视频时间轴长度。
+
+ 无配音仍需要一个音频占位来驱动现有素材裁剪、字幕时间轴和最终合成。
+ 估算策略尽量简单:
+ 1. 中文等 CJK 字符按约 4.2 字/秒估算;
+ 2. 英文/数字按约 2.7 词/秒估算;
+ 3. 其他语种文字按约 4.0 字符/秒兜底估算,覆盖俄语、阿拉伯语、
+ 日文假名、韩文等非 ASCII 文本;
+ 4. 每个断句补一点停顿,让字幕切换不至于过于紧凑;
+ 5. 最少 3 秒,避免极短脚本生成 0 秒音频。
+ """
+ normalized_text = (text or "").strip()
+ if not normalized_text:
+ return 3.0
+
+ cjk_chars = len(re.findall(r"[\u4e00-\u9fff]", normalized_text))
+ words = len(re.findall(r"[A-Za-z0-9]+", normalized_text))
+ ascii_word_chars = sum(len(word) for word in re.findall(r"[A-Za-z0-9]+", normalized_text))
+ other_text_chars = 0
+ for char in normalized_text:
+ # Unicode category 以 L 开头表示各语种字母,N 表示数字。前面已经单独
+ # 统计了 CJK 和 ASCII 单词,这里只统计剩余文字,避免英文被重复计时。
+ category = unicodedata.category(char)
+ if category.startswith(("L", "N")):
+ other_text_chars += 1
+ other_text_chars = max(other_text_chars - cjk_chars - ascii_word_chars, 0)
+ sentence_count = max(len(utils.split_string_by_punctuations(normalized_text)), 1)
+
+ cjk_duration = cjk_chars / 4.2
+ word_duration = words / 2.7
+ other_text_duration = other_text_chars / 4.0
+ pause_duration = max(sentence_count - 1, 0) * 0.35
+ return max(3.0, cjk_duration + word_duration + other_text_duration + pause_duration)
+
+
+def generate_silent_audio(duration_seconds: float, output_file: str) -> bool:
+ """
+ 生成 MP3 静音音频,作为“无配音”模式的时间轴占位。
+
+ 使用 FFmpeg 的 anullsrc 直接生成静音,比先构造临时 WAV 再转码更少中间
+ 文件。失败时返回 False,让上层按普通 TTS 失败路径处理并记录日志。
+ """
+ ensure_file_path_exists(output_file)
+ duration_seconds = max(float(duration_seconds or 0), 0.1)
+ ffmpeg_binary = utils.get_ffmpeg_binary()
+ command = [
+ ffmpeg_binary,
+ "-y",
+ "-f",
+ "lavfi",
+ "-i",
+ "anullsrc=r=44100:cl=mono",
+ "-t",
+ f"{duration_seconds:.3f}",
+ "-codec:a",
+ "libmp3lame",
+ "-q:a",
+ "4",
+ output_file,
+ ]
+
+ logger.info(
+ f"generating silent audio for no-voice mode, duration: {duration_seconds:.2f}s"
+ )
+ result = subprocess.run(
+ command,
+ capture_output=True,
+ text=True,
+ check=False,
+ )
+ if result.returncode != 0:
+ logger.error(
+ "failed to generate silent audio: "
+ f"{(result.stderr or result.stdout or '').strip()}"
+ )
+ return False
+ if not os.path.exists(output_file) or os.path.getsize(output_file) <= 0:
+ logger.error(
+ "silent audio output file is missing or empty, "
+ f"file: {output_file}, duration: {duration_seconds:.2f}s"
+ )
+ return False
+ return True
+
+
def tts(
text: str,
voice_name: str,
@@ -209,6 +305,18 @@ def tts(
voice_file: str,
voice_volume: float = 1.0,
) -> Union[SubMaker, None]:
+ if is_no_voice(voice_name):
+ duration_seconds = estimate_no_voice_duration(text)
+ if not generate_silent_audio(duration_seconds, voice_file):
+ return None
+
+ sub_maker = ensure_legacy_submaker_fields(SubMaker())
+ return populate_legacy_submaker_with_full_text(
+ sub_maker=sub_maker,
+ text=text,
+ audio_duration_seconds=duration_seconds,
+ )
+
if is_azure_v2_voice(voice_name):
return azure_tts_v2(text, voice_name, voice_file)
elif is_siliconflow_voice(voice_name):
diff --git a/app/utils/utils.py b/app/utils/utils.py
index 69d46010..411a473d 100644
--- a/app/utils/utils.py
+++ b/app/utils/utils.py
@@ -2,6 +2,7 @@ import json
import locale
import os
import re
+import shutil
from functools import lru_cache
from pathlib import Path
import threading
@@ -121,6 +122,42 @@ def public_dir(sub_dir: str = ""):
return d
+def get_ffmpeg_binary() -> str:
+ """
+ 解析当前进程应该使用的 FFmpeg 可执行文件。
+
+ 增加原因:
+ 1. 视频编码、静音音频生成、pydub 音频转码都依赖 FFmpeg;
+ 2. Windows 便携包、Docker 和用户自定义安装目录经常出现 PATH 不一致;
+ 3. 集中解析可以让所有调用方使用同一套优先级,减少某条链路能跑、
+ 另一条链路找不到 FFmpeg 的现场问题。
+
+ 优先级:
+ 1. IMAGEIO_FFMPEG_EXE:MoviePy/imageio 约定的显式配置;
+ 2. 系统 PATH 中的 ffmpeg;
+ 3. imageio-ffmpeg 依赖提供的内置二进制;
+ 4. 字符串 "ffmpeg" 兜底,交给 subprocess 在运行时暴露更具体错误。
+ """
+ configured_ffmpeg = os.environ.get("IMAGEIO_FFMPEG_EXE")
+ if configured_ffmpeg:
+ return configured_ffmpeg
+
+ system_ffmpeg = shutil.which("ffmpeg")
+ if system_ffmpeg:
+ return system_ffmpeg
+
+ try:
+ import imageio_ffmpeg
+
+ bundled_ffmpeg = imageio_ffmpeg.get_ffmpeg_exe()
+ if bundled_ffmpeg:
+ return bundled_ffmpeg
+ except Exception as exc:
+ logger.warning(f"failed to resolve bundled ffmpeg binary: {str(exc)}")
+
+ return "ffmpeg"
+
+
def run_in_background(func, *args, **kwargs):
def run():
try:
diff --git a/test/services/test_subtitle_background_settings.py b/test/services/test_subtitle_background_settings.py
index ff7c46aa..e7133ae7 100644
--- a/test/services/test_subtitle_background_settings.py
+++ b/test/services/test_subtitle_background_settings.py
@@ -15,6 +15,7 @@ class TestSubtitleBackgroundSettings(unittest.TestCase):
required_keys = {
"Enable Subtitle Background",
"Subtitle Background Color",
+ "No Voice",
}
for locale_file in i18n_dir.glob("*.json"):
diff --git a/test/services/test_video.py b/test/services/test_video.py
index 31660cdb..53440a45 100644
--- a/test/services/test_video.py
+++ b/test/services/test_video.py
@@ -185,7 +185,7 @@ class TestVideoService(unittest.TestCase):
def test_get_ffmpeg_binary_uses_configured_env_path(self):
"""配置中显式指定 ffmpeg 时,应优先使用该路径。"""
with patch.dict(os.environ, {"IMAGEIO_FFMPEG_EXE": "/tmp/custom-ffmpeg"}, clear=True):
- self.assertEqual(vd.get_ffmpeg_binary(), "/tmp/custom-ffmpeg")
+ self.assertEqual(utils.get_ffmpeg_binary(), "/tmp/custom-ffmpeg")
def test_get_ffmpeg_binary_falls_back_to_imageio_ffmpeg(self):
"""
@@ -197,9 +197,9 @@ class TestVideoService(unittest.TestCase):
)
with patch.dict(os.environ, {}, clear=True), patch.object(
- vd.shutil, "which", return_value=None
+ utils.shutil, "which", return_value=None
), patch.dict(sys.modules, {"imageio_ffmpeg": fake_imageio_ffmpeg}):
- self.assertEqual(vd.get_ffmpeg_binary(), "/tmp/bundled-ffmpeg")
+ self.assertEqual(utils.get_ffmpeg_binary(), "/tmp/bundled-ffmpeg")
def test_get_effective_video_codec_falls_back_when_encoder_missing(self):
"""
diff --git a/test/services/test_voice.py b/test/services/test_voice.py
index 0c6f6649..6e18904f 100644
--- a/test/services/test_voice.py
+++ b/test/services/test_voice.py
@@ -61,6 +61,98 @@ class TestVoiceService(unittest.TestCase):
all(v.startswith(("zh-CN", "en-US")) for v in filtered)
)
+ def test_no_voice_tts_generates_silent_audio_and_subtitle_timeline(self):
+ """
+ 无配音模式不调用任何外部 TTS provider,只生成静音音频作为时间轴占位。
+ 这里 mock FFmpeg,验证请求参数、输出文件和 legacy 字幕结构都符合后续
+ 视频合成链路的预期。
+ """
+
+ def fake_run(command, capture_output, text, check):
+ self.assertEqual(command[0], "/tmp/fake-ffmpeg")
+ self.assertIn("anullsrc=r=44100:cl=mono", command)
+ Path(command[-1]).write_bytes(b"fake-silent-mp3")
+ return SimpleNamespace(returncode=0, stdout="", stderr="")
+
+ with tempfile.TemporaryDirectory() as tmp_dir, patch.object(
+ vs.utils,
+ "get_ffmpeg_binary",
+ return_value="/tmp/fake-ffmpeg",
+ ), patch.object(vs.subprocess, "run", side_effect=fake_run):
+ voice_file = str(Path(tmp_dir) / "silent.mp3")
+ sub_maker = vs.tts(
+ text="第一句话。Second sentence.",
+ voice_name=vs.NO_VOICE_NAME,
+ voice_rate=1.0,
+ voice_file=voice_file,
+ )
+
+ self.assertEqual(Path(voice_file).read_bytes(), b"fake-silent-mp3")
+
+ self.assertIsNotNone(sub_maker)
+ self.assertEqual(getattr(sub_maker, "subs", []), ["第一句话", "Second sentence"])
+ self.assertEqual(len(getattr(sub_maker, "offset", [])), 2)
+ self.assertGreater(vs.get_audio_duration(sub_maker), 0)
+
+ def test_no_voice_alias_none_is_supported_temporarily(self):
+ """
+ 兼容 PR #981 曾使用过的 none sentinel,避免少量直接调用 API 的用户
+ 升级后立即失效。新 UI 和新代码仍统一使用 no-voice。
+ """
+ self.assertTrue(vs.is_no_voice("none"))
+ self.assertTrue(vs.is_no_voice(vs.NO_VOICE_NAME))
+ self.assertFalse(vs.is_no_voice(""))
+
+ def test_no_voice_duration_estimates_non_ascii_languages(self):
+ """
+ 无配音没有真实 TTS 音频,只能根据脚本文字估算阅读时间。俄语、阿拉伯语、
+ 日文假名、韩文等非 ASCII 文本也必须参与估算,不能都落到最短 3 秒。
+ """
+ russian_text = (
+ "Это длинный тестовый сценарий без озвучки. "
+ "Он должен получить достаточно времени для чтения субтитров."
+ )
+ arabic_text = "هذا اختبار طويل بدون تعليق صوتي، ويجب أن يحصل على وقت كاف لقراءة الترجمة."
+
+ self.assertGreater(vs.estimate_no_voice_duration(russian_text), 8.0)
+ self.assertGreater(vs.estimate_no_voice_duration(arabic_text), 8.0)
+
+ def test_generate_silent_audio_rejects_missing_output_file(self):
+ """
+ 即使 FFmpeg 进程返回成功,也要确认输出文件真实存在且非空。这样可以把
+ 异常收敛在 TTS 阶段,而不是拖到后续视频合成阶段才暴露。
+ """
+ with tempfile.TemporaryDirectory() as tmp_dir, patch.object(
+ vs.utils,
+ "get_ffmpeg_binary",
+ return_value="/tmp/fake-ffmpeg",
+ ), patch.object(
+ vs.subprocess,
+ "run",
+ return_value=SimpleNamespace(returncode=0, stdout="", stderr=""),
+ ):
+ voice_file = str(Path(tmp_dir) / "missing-silent.mp3")
+
+ self.assertFalse(vs.generate_silent_audio(3.0, voice_file))
+
+ def test_empty_voice_name_does_not_enable_no_voice_mode(self):
+ """
+ 空 voice 通常意味着配置缺失或接口参数错误,不能自动切到无配音模式。
+ 否则用户填错 TTS 配置时也会得到一个“成功”的静音视频,定位成本更高。
+ """
+ sentinel = object()
+
+ with patch.object(vs, "azure_tts_v1", return_value=sentinel) as azure_tts_v1:
+ result = vs.tts(
+ text="empty voice should still use the default TTS path",
+ voice_name="",
+ voice_rate=1.0,
+ voice_file="/tmp/empty-voice.mp3",
+ )
+
+ self.assertIs(result, sentinel)
+ azure_tts_v1.assert_called_once()
+
def test_siliconflow(self):
# SiliconFlow 的 API Key 存在 [siliconflow].api_key 中,运行时代码也是从
# config.siliconflow 读取;这里必须使用同一配置源,避免正确配置凭据时
diff --git a/webui/Main.py b/webui/Main.py
index a31b42b8..d47a1d36 100644
--- a/webui/Main.py
+++ b/webui/Main.py
@@ -262,6 +262,7 @@ if not config.app.get("hide_config", False):
("OneAPI", "oneapi"),
("Cloudflare", "cloudflare"),
("ERNIE", "ernie"),
+ ("MiniMax", "minimax"),
("MiMo", "mimo"),
("Pollinations", "pollinations"),
("LiteLLM", "litellm"),
@@ -801,6 +802,7 @@ with middle_panel:
# 添加TTS服务器选择下拉框
tts_servers = [
+ (voice.NO_VOICE_NAME, tr("No Voice")),
("azure-tts-v1", "Azure TTS V1"),
("azure-tts-v2", "Azure TTS V2"),
("siliconflow", "SiliconFlow TTS"),
@@ -829,7 +831,11 @@ with middle_panel:
# 根据选择的TTS服务器获取声音列表
filtered_voices = []
- if selected_tts_server == "siliconflow":
+ if selected_tts_server == voice.NO_VOICE_NAME:
+ # 无配音是显式模式,只提供一个稳定 sentinel。这样普通 TTS 的空配置
+ # 不会被误判为静音,后端也能继续通过同一条音频/字幕流程生成视频。
+ filtered_voices = [voice.NO_VOICE_NAME]
+ elif selected_tts_server == "siliconflow":
# 获取硅基流动的声音列表
filtered_voices = voice.get_siliconflow_voices()
elif selected_tts_server == "gemini-tts":
@@ -853,12 +859,15 @@ with middle_panel:
if "V2" not in v:
filtered_voices.append(v)
- friendly_names = {
- v: v.replace("Female", tr("Female"))
- .replace("Male", tr("Male"))
- .replace("Neural", "")
- for v in filtered_voices
- }
+ if selected_tts_server == voice.NO_VOICE_NAME:
+ friendly_names = {voice.NO_VOICE_NAME: tr("No Voice")}
+ else:
+ friendly_names = {
+ v: v.replace("Female", tr("Female"))
+ .replace("Male", tr("Male"))
+ .replace("Neural", "")
+ for v in filtered_voices
+ }
saved_voice_name = config.ui.get("voice_name", "")
saved_voice_name_index = 0
@@ -902,8 +911,12 @@ with middle_panel:
params.voice_name = ""
config.ui["voice_name"] = ""
- # 只有在有声音可选时才显示试听按钮
- if friendly_names and st.button(tr("Play Voice")):
+ # 无配音模式会生成静音占位音频,不展示试听按钮,避免用户误以为需要测试声音。
+ if (
+ friendly_names
+ and selected_tts_server != voice.NO_VOICE_NAME
+ and st.button(tr("Play Voice"))
+ ):
play_content = params.video_subject
if not play_content:
play_content = params.video_script
diff --git a/webui/i18n/de.json b/webui/i18n/de.json
index 3ea67aea..82b97e65 100644
--- a/webui/i18n/de.json
+++ b/webui/i18n/de.json
@@ -54,6 +54,7 @@
"Speech Rate": "Lesegeschwindigkeit (1,0 bedeutet 1x)",
"Custom Audio File": "Benutzerdefinierte Audiodatei",
"Custom audio will be used directly. TTS synthesis will be skipped for this task.": "Die benutzerdefinierte Audiodatei wird direkt verwendet. Die TTS-Synthese wird für diese Aufgabe übersprungen.",
+ "No Voice": "Keine Stimme (keine Sprachsynthese)",
"Male": "Männlich",
"Female": "Weiblich",
"Background Music": "Hintergrundmusik",
diff --git a/webui/i18n/en.json b/webui/i18n/en.json
index 48b83bbf..1b7657f6 100644
--- a/webui/i18n/en.json
+++ b/webui/i18n/en.json
@@ -51,6 +51,7 @@
"Speech Rate": "Speech Rate (1.0 means 1x speed)",
"Custom Audio File": "Custom Audio File",
"Custom audio will be used directly. TTS synthesis will be skipped for this task.": "Custom audio will be used directly. TTS synthesis will be skipped for this task.",
+ "No Voice": "No Voice (skip speech synthesis)",
"Male": "Male",
"Female": "Female",
"Background Music": "Background Music",
diff --git a/webui/i18n/pt.json b/webui/i18n/pt.json
index a9757209..26335f8e 100644
--- a/webui/i18n/pt.json
+++ b/webui/i18n/pt.json
@@ -42,6 +42,7 @@
"Speech Key": "Chave da API(:red[Obrigatório,[Obter Chave da API](https://portal.azure.com/#view/Microsoft_Azure_ProjectOxford/CognitiveServicesHub/~/SpeechServices)])",
"Speech Volume": "Volume da Fala (1.0 representa 100%)",
"Speech Rate": "Velocidade da Fala (1.0 significa velocidade 1x)",
+ "No Voice": "Sem voz (sem sintese de fala)",
"Male": "Masculino",
"Female": "Feminino",
"Background Music": "Música de Fundo",
diff --git a/webui/i18n/ru.json b/webui/i18n/ru.json
index c9175777..bff12772 100644
--- a/webui/i18n/ru.json
+++ b/webui/i18n/ru.json
@@ -42,6 +42,7 @@
"Speech Key": "API-ключ (:red[Обязательно,[Получить API-ключ](https://portal.azure.com/#view/Microsoft_Azure_ProjectOxford/CognitiveServicesHub/~/SpeechServices)])",
"Speech Volume": "Громкость речи (1.0 соответствует 100%)",
"Speech Rate": "Скорость речи (1.0 означает обычную скорость)",
+ "No Voice": "Без озвучки (без синтеза речи)",
"Male": "Мужской",
"Female": "Женский",
"Background Music": "Фоновая музыка",
diff --git a/webui/i18n/tr.json b/webui/i18n/tr.json
index 32e9eea8..e01d0960 100644
--- a/webui/i18n/tr.json
+++ b/webui/i18n/tr.json
@@ -49,6 +49,7 @@
"Speech Key": "API Anahtarı (:red[Gerekli, [API Anahtarı Al](https://portal.azure.com/#view/Microsoft_Azure_ProjectOxford/CognitiveServicesHub/~/SpeechServices)])",
"Speech Volume": "Konuşma Ses Seviyesi (1.0, %100 anlamına gelir)",
"Speech Rate": "Konuşma Hızı (1.0, 1x hız anlamına gelir)",
+ "No Voice": "Ses yok (konuşma sentezi yok)",
"Male": "Erkek",
"Female": "Kadın",
"Background Music": "Arka Plan Müziği",
diff --git a/webui/i18n/vi.json b/webui/i18n/vi.json
index 363971fa..cbe2a641 100644
--- a/webui/i18n/vi.json
+++ b/webui/i18n/vi.json
@@ -51,6 +51,7 @@
"Speech Rate": "Tốc độ đọc (1.0 biểu thị tốc độ gốc)",
"Custom Audio File": "Tệp Âm Thanh Tùy Chỉnh",
"Custom audio will be used directly. TTS synthesis will be skipped for this task.": "Âm thanh tùy chỉnh sẽ được dùng trực tiếp. Bước tổng hợp giọng nói (TTS) sẽ được bỏ qua cho tác vụ này.",
+ "No Voice": "Không có giọng đọc (không tổng hợp giọng nói)",
"Male": "Nam",
"Female": "Nữ",
"Background Music": "Âm Nhạc Nền",
diff --git a/webui/i18n/zh.json b/webui/i18n/zh.json
index 323934cc..6a8b666f 100644
--- a/webui/i18n/zh.json
+++ b/webui/i18n/zh.json
@@ -51,6 +51,7 @@
"Speech Rate": "朗读速度(1.0表示1倍速)",
"Custom Audio File": "自定义音频文件",
"Custom audio will be used directly. TTS synthesis will be skipped for this task.": "将直接使用自定义音频,本次任务会跳过 TTS 语音合成。",
+ "No Voice": "无配音(不合成语音)",
"Male": "男性",
"Female": "女性",
"Background Music": "背景音乐",