论文解读
记忆即变换：LETHE 用能量判别与随机扰动让混响参数自己走
LETHE 把 3×3 反馈延迟网络当作可被改写的房间，用五特征线性判别器对照初始声音 DNA、以单样本扰动更新十一个参数，45 组 180 秒对照显示只有生成器开启时 c22 才会演化，但稳定工作点偏离经典纳什均衡且扰动先验仍待剥离。
开源工具
论文解读
打开混响黑盒：从 Schroeder 结构到可核对的分析与移植路径
论文以 Schroeder 历史混响为案例，主张用开放架构重建 comb 与 all-pass 组合，通过 Csound 脉冲响应分析与 par/seq 重建验证理解过程，代价是遇到效率限制而把 C 编译 opcode 留作后续工作。
论文解读
当混响不再模仿房间：用声码器思路把噪声织成尾响
音频生成 | 7.2/10
论文解读
把拾音器从琴上焊死的位置里解放出来：64 路沿弦采样如何把音色选择推迟到混音台
音乐源分离 | 7.7/10
论文解读
把 100 小时 MEG 拆成两种稀缺资源，才看得见神经语音解码该怎样扩展
基准测试 | 8.9/10
论文解读
Arbitrary Polygon Oscillator: Generalizing Polygonal Synthesis to Arbitrary Shapes, Morphing, and Three-Dimensional Polyhedra
音乐生成 | 8.7/10
论文解读
One Timeline, Many Renderings: A Wolfram Language Paclet for heterogeneous musical output
音乐生成 | 7.0/10
论文解读
The ISCSLP 2026 Real-World Audio-Visual Speech Enhancement Challenge
音视频语音分离 | 8.8/10
论文解读
MusPyExpress: Extending MusPy with Enhanced Expression Text Support
音乐理解 | 8.7/10
论文解读
Computational Features for Symbolic Melody Analysis
音乐理解 | 7.6/10
论文解读
Automatic Transcription of Microtonal Free-Rhythm Vocal Music: A Case Study in Iranian Classical Music
音乐转录 | 7.5/10
论文解读
Sonifying I2S Transport Signals to Detect Transmission Faults
音频分类 | 5.9/10
论文解读
FastThaiG2P: Lightning-fast Thai Grapheme-to-phoneme Conversion for Voice Agent Pipelines
语音合成 | 7.0/10
论文解读
OmniScientist: An Omni-Modal Omni-Discipline AI Scientist
多模态模型 | 7.6/10
每日研究速递
语音/音乐/音频论文速递 2026-08-14
共分析 12 篇语音/AI 论文
论文解读
Easper: An Accessible ASR Pipeline for Language Documentation
语音识别 | 7.0/10
论文解读
DIY e-HandPan: A new DIY Low-Cost Handpan Interface based on Arduino and ESP32 Microcontrollers
音频交互 | 7.2/10
论文解读
omni-macos: On-Device Omni-Modal Search on Apple Silicon
音频检索 | 7.2/10
论文解读
GraphIDyOM: A graph-native Python reimplementation of IDyOM for musical expectation modelling
音乐理解 | 7.8/10
论文解读
JarvisHub: An Open Harness for Canvas-Native Multimodal Creative Agents
音视频生成 | 6.2/10