cyanspeech
cyanspeech 为 Minecraft 添加了本地文本转语音功能,拥有庞大的 LibriTTS/Piper 声音目录、游戏内语音控制,以及通过 Simple Voice Chat 进行可选播放的支持。它面向那些希望使用打字聊天、角色扮演台词、无障碍语音或快速客户端语音输出,而无需运行独立服务器的玩家。
警告
cyanspeech 目前在 Linux 系统上表现不佳!
功能
- 模组加载时自动准备的本地文本转语音渲染。
- 基于
en_US-libritts_r-medium Piper 语音模型的 904 人声库。
- 支持目录名称的语音 ID,例如
libritts-...、piper-000、voice-0 和数字说话人索引。
- 当安装并连接了 Simple Voice Chat 时,提供可选的 Simple Voice Chat 输出。
- 可选的本地桌面播放,以便在自己的客户端上听到生成的语音。
- 可选的非命令聊天消息自动语音发送。
- 持久化的客户端设置,包括语音、播放、聊天语音、键位绑定、收藏语音、星标颜色及语音效果。
- 在 Fabric、Quilt 和 NeoForge 版本中提供完整的游戏内设置界面。
- 可折叠的设置部分,涵盖语音、播放、预览、效果和许可证。
- 可排序的语音选择器弹窗,支持搜索、收藏语音、星标颜色、语音元数据和预览按钮。
- 平铺式语音效果控件,包括音调、速度、音量、音色、质感、情绪、广播风格处理、回声、混响、合唱、颤音、震幅、失真和比特压缩。
- 内置的许可证和来源归属信息,用于包含的语音模型元数据和数据集谱系。
游戏内控制
cyanspeech 提供了一个紧凑的客户端命令界面,用于语音、语音选择、播放切换和效果调整。
| 命令 |
描述 |
/cyanspeech |
显示 cyanspeech 状态和当前设置。 |
/cyanspeech <text> |
朗读提供的文本。 |
/cyanspeech say <text> |
朗读提供的文本。 |
/cyanspeech speak <text> |
朗读提供文本的别名。 |
/cyanspeech enable |
启用生成的语音输出。 |
/cyanspeech disable |
禁用生成的语音输出。 |
/cyanspeech toggle |
切换生成的语音输出。 |
/cyanspeech voice |
显示当前活动语音并提供可用目录的预览。 |
/cyanspeech voice <voice> |
从 cyanspeech 目录中选择一种语音。 |
/cyanspeech settings |
显示当前持久化设置。 |
/cyanspeech settings voice <voice> |
通过设置命令选择语音。 |
| `/cyanspeech settings svc <on |
off>` |
| `/cyanspeech settings local <on |
off>` |
| `/cyanspeech settings chat <on |
off>` |
/cyanspeech settings effect <name> <value> |
调整语音效果值。 |
/cyanspeech options |
/cyanspeech settings 的别名。 |
支持的效果名称包括 pitch(音调)、speed(速度)、volume(音量)、brightness(明亮度)、warmth(温暖度)、nasality(鼻音)、breathiness(气声)、roughness(粗糙度)、urgency(紧迫感)、fear(恐惧)、anger(愤怒)、amusement(趣味)、sadness(悲伤)、tired(疲惫)、outOfBreath(气喘吁吁)、injured(受伤)、cold(寒冷)、whisper(耳语)、robot(机器人)、chorus(合唱)、echo(回声)、reverb(混响)、distortion(失真)、tremolo(震幅)、vibrato(颤音)、radio(广播)和 bitcrush(比特压缩)。
设置界面
游戏内设置界面将相同的控制选项集中在一个位置:语音选择、播放路由、预览短语、效果滑块、收藏语音管理和许可证归属。语音选择器可以搜索和排序整个目录,预览选定的语音,筛选收藏语音,并在本地自定义星标颜色。
兼容性
cyanspeech 通过 cyancore 多加载器构建系统发布,目前包含以下客户端适配器:
- Fabric 1.21.11
- Fabric 26.1.2
- Quilt 1.21.11
- NeoForge 26.1.2
针对旧版 26.1 和 26.1.1 的目标请求会被路由到 cyancore 构建矩阵中可用的 26.1.2 适配器。
备注
Simple Voice Chat 是可选的。当它存在且已连接时,cyanspeech 可以通过语音聊天通道发送生成的语音。本地播放也可以独立使用,供客户端收听。
语音模型和数据集的归属信息显示在游戏内的许可证部分。当前的声音目录基于 Piper 语音元数据以及 LibriTTS-R/OpenSLR 141 的来源谱系。
cyanspeech
cyanspeech adds local text-to-speech to Minecraft with a large LibriTTS/Piper voice catalog, in-game voice controls, and optional playback through Simple Voice Chat. It is built for players who want typed chat, roleplay lines, accessibility speech, or quick client-side voice output without running a separate server.
WARNING
cyanspeech does not presently work well with Linux!
Features
- Local text-to-speech rendering prepared automatically when the mod loads.
- 904-speaker voice catalog based on the
en_US-libritts_r-medium Piper voice model.
- Voice IDs that support catalog names such as
libritts-..., piper-000, voice-0, and numeric speaker indexes.
- Optional Simple Voice Chat output when Simple Voice Chat is installed and connected.
- Optional local desktop playback for hearing generated speech on your own client.
- Optional outgoing chat auto-speech for non-command chat messages.
- Persisted client settings for voice, playback, chat speech, keybinds, starred voices, star colors, and voice effects.
- Full in-game settings screen on Fabric, Quilt, and NeoForge builds.
- Collapsible settings sections for voice, playback, preview, effects, and licenses.
- Sortable voice selector popup with search, starred voices, star colors, voice metadata, and preview buttons.
- Tiled voice effect controls for pitch, speed, volume, tone, texture, emotion, radio-style processing, echo, reverb, chorus, vibrato, tremolo, distortion, and bit crush.
- Built-in license and source attribution for the included voice model metadata and dataset lineage.
In-Game Controls
cyanspeech exposes a compact client command surface for speech, voice selection, playback toggles, and effects.
| Command |
Description |
/cyanspeech |
Shows cyanspeech status and current settings. |
/cyanspeech <text> |
Speaks the provided text. |
/cyanspeech say <text> |
Speaks the provided text. |
/cyanspeech speak <text> |
Alias for speaking the provided text. |
/cyanspeech enable |
Enables generated speech output. |
/cyanspeech disable |
Disables generated speech output. |
/cyanspeech toggle |
Toggles generated speech output. |
/cyanspeech voice |
Shows the active voice and a preview of the available catalog. |
/cyanspeech voice <voice> |
Selects a voice from the cyanspeech catalog. |
/cyanspeech settings |
Shows current persisted settings. |
/cyanspeech settings voice <voice> |
Selects a voice through the settings command. |
| `/cyanspeech settings svc <on |
off>` |
| `/cyanspeech settings local <on |
off>` |
| `/cyanspeech settings chat <on |
off>` |
/cyanspeech settings effect <name> <value> |
Adjusts a voice effect value. |
/cyanspeech options |
Alias for /cyanspeech settings. |
Supported effect names include pitch, speed, volume, brightness, warmth, nasality, breathiness, roughness, urgency, fear, anger, amusement, sadness, tired, outOfBreath, injured, cold, whisper, robot, chorus, echo, reverb, distortion, tremolo, vibrato, radio, and bitcrush.
Settings Screen
The in-game settings screen brings the same controls into one place: voice selection, playback routing, preview phrases, effect sliders, starred voice management, and license attribution. The voice selector can search and sort the full catalog, preview a selected voice, filter starred voices, and customize star colors locally.
Compatibility
cyanspeech ships through the cyancore multi-loader build system and currently includes client adapters for:
- Fabric 1.21.11
- Fabric 26.1.2
- Quilt 1.21.11
- NeoForge 26.1.2
Older 26.1 and 26.1.1 target requests are routed to the available 26.1.2 adapter in the cyancore build matrix.
Notes
Simple Voice Chat is optional. When it is present and connected, cyanspeech can send generated speech through the voice chat route. Local playback can also be used independently for client-side listening.
Voice model and dataset attribution is shown in the in-game license section. The current catalog is based on Piper voice metadata and the LibriTTS-R/OpenSLR 141 source lineage.
评论 0
登录后参与评分
还没有评论,来抢沙发吧