代码 GPLv3 开源 · “HonestMix / 诚听”名称为商标（申请中），衍生产品请勿使用本名称 → TRADEMARK.md
A free headphone-correction and monitor-translation plugin (VST3/AU), built in Chinese, for the hundreds of thousands of home-studio producers who mix on headphones alone.
Status: early design stage (pre-alpha). This repository currently hosts the project book and research plan. Code will follow.
Most home-studio producers in China mix entirely on headphones. Speakers and treated rooms are simply beyond their budgets. Headphone mixing suffers from three structural distortions:
- Frequency response - every headphone colors the sound (solvable with measurement + filtering)
- Missing crosstalk - speakers reach both ears, headphones do not (addressed with HRTF filtering)
- Missing room - the brain uses early reflections to judge depth and reverb (approximated with BRIR convolution)
The result: mixes that do not translate, and long sessions at ever-rising listening levels. We want to fix what software can honestly fix - no more, no less.
- VST3/AU plugin on the master bus, near-zero learning curve
- Headphone correction profiles (domestic Chinese models included and prioritized)
- HRTF crossfeed + measured-room simulation
- Translation checks (phone speaker / car / small room)
- SPL calibration and gentle listening-exposure awareness as first-class features
- Chinese UI, Chinese tutorials,人民币 pricing if ever commercial
- We use headphone measurement data only with the explicit permission of its authors, and we credit them in the plugin and documentation from day one.
- Sources known to forbid commercial use (e.g., RTINGS) will not be used at all.
- The project is completely free and non-commercial during its public testing phase. If it ever becomes commercial, we will return to every data author first and accept whatever licensing terms they consider fair.
- No medical claims. This tool cannot prevent ear disease; it can only help people listen at saner levels.
- Headphone model names remain trademarks of their owners.
We pre-commit to a blind evaluation protocol:
- 20-30 experienced home-studio producers each mix the same song twice, with and without the plugin, order randomized.
- Both versions are judged blind on five playback systems (studio monitors / car / phone / bluetooth speaker / another headphone).
- Win rate >= 60%: the tool works. 50-60%: iterate. < 50%: we stop and publish the data anyway, so the community learns from it either way.
| Phase | Content |
|---|---|
| M0 | Trademark & licensing decisions, DSP latency spike |
| M1 | MVP: correction + crossfeed + bypass A/B |
| M2 | + first measured room, SPL calibration, UI freeze |
| M3-4 | Blind internal test (20-30 producers) |
| M5-7 | Free public beta, monthly iteration |
| M8 | Data-driven decision: monetize / stay free / open-source |
JunXian Li - mixing engineer, SOLAR&TERRA DESIGN STUDIO, China. Issues and discussions are open. 中文也完全没问题。
诚听是一款免费的中文耳机监听矫正插件（VST3/AU），为只能用耳机混音的 home studio 制作人而生。
- 三层失真全部覆盖：频响矫正、HRTF 串扰、实测房间模拟
- 翻译检查：一键切换手机外放 / 车载 / 小房间
- 听力健康：SPL 校准与聆听暴露量提醒，做成功能而非口号
- 伦理先行：测量数据只在获得作者明确授权后使用并署名；禁商用来源 （如 RTINGS）一概不碰；公测期完全免费非商用；若将来商用，先回到每位 数据作者面前接受其公平的授权条款
- 用盲测说话：20-30 位制作人双盲混音对比，五种回放系统盲评， 事先承诺判定标准——胜率不到 50% 就停止并公开全部数据
当前阶段：设计期（pre-alpha），本仓库先存放项目书与研究计划，代码随后。
发起人：李俊贤，混音师，太阳与大地设计工作室（SOLAR&TERRA DESIGN STUDIO）。