Skip to content

v1.2.0 — Cozy Redesign

Choose a tag to compare

View all tags
@nightt5879 nightt5879 released this 07 May 09:27
· 14 commits to main since this release
v1.2.0
ad4d9f0
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

完整重做了浏览器界面，并将 12 条试听样例换成现场算法生成。
A full studio refresh, plus rebuilt listening samples that now flow through the live audio pipeline.

🎨 UI 升级 / UI

  • Cozy 手作设计系统：抛弃原 Apple 风中性 SaaS，整体改为 2px 墨线 + 硬偏移阴影 + 暖纸底色 + 9 色 character cast 的"贴纸笔记本"风格。Self-hosted Nunito / JetBrains Mono / Noto Sans SC（约 344 KB）。
    Brand-new sticker-style visual system: 2px ink outlines, hard-offset shadows, warm paper canvas, and a nine-color character cast. Nunito / JetBrains Mono / Noto Sans SC ship self-hosted.
  • 9 个角色专属虚拟形象：每个 preset 配独特 SVG 表情画像 —— 圆眼粉腮 Cute NPC、像素 LED 眼 Robot Guard、闭眼小笑 Forest Spirit、皇冠紫袍 Deep Boss…
    Each of the nine presets now has its own expressive SVG face.
  • 整页一屏布局：收紧 padding / 字号 / 高度，studio + preset 行 + 高级抽屉在 768p+ 屏幕一屏装下。
    Tightened spacing so the whole studio fits on a single screen at 768p+.
  • 试听页角色 tag 配头像：listen.html 12 条卡片角色标签前置 SVG 头像，与主 app 一致。
    Showcase role tags now lead with the matching character avatar.

🎤 内核升级 / Core

  • 多人同时讲话面板：勾选任意 preset 组合 + 每行一句台词，一键全员同时开口；起声 0–0.18s deterministic 错开避免硬同步，混音自动 -10·log₁₀(N) dB 减幅防削峰。
    New Crowd Chatter panel — pick any subset of presets, multiple lines per row, fire on the same tick. 0–0.18s deterministic stagger and -10·log₁₀(N) dB headroom keep the mix natural.
  • 试听样例重新生成：12 条 WAV 现在全部由真正的 buildSchedule + renderEventsToWav 跑出 —— jieba 分词、pinyin 声调、Web Audio formant / ring mod / compressor 全在线，所听即所得。
    Showcase clips are now produced by the live audio pipeline — jieba segmentation, pinyin tones, formant filtering, ring modulation, and the dynamics compressor all participate.
  • npm run samples 改用 tsx + node-web-audio-api：Node 端直接跑真实音频管线，未来再生成不会和现场脱节。
    Sample regeneration now runs through tsx with node-web-audio-api as the OfflineAudioContext polyfill — future regens stay locked to the live algorithm.

🛠 其他 / Misc

  • UiCopy.heroCopy 移入 i18n（zh + en）
  • 主题切换器暂时隐藏（设计目前只 ship 亮色 token，state + localStorage 保留以便未来恢复）
  • public/fonts/ 自托管 8 个 woff2 字体
  • 路线图刷新，标注 V1.1.0 / V1.2.0 完成项

📦 升级 / Upgrade

无破坏性改动。Cache-buster 已 bump 到 ?v=1.2.0-cozy，浏览器会重抓 UI 与音频。
No breaking changes — cache-busters bumped to ?v=1.2.0-cozy so browsers refetch the new UI and audio.

Pull request: #1
Live demo: https://nightt5879.github.io/mumble-voice-lab/?v=1.2.0-cozy
Listening showcase: https://nightt5879.github.io/mumble-voice-lab/showcase.html?v=1.2.0-cozy

Assets 2
Loading