Unverified60% confidenceFactExact time
Mel features are obtained by first performing an FFT on short-time audio frames (e.g., 25ms window, 10ms step) to get a spectrum, then compressing it through a triangular Mel filter bank to 20-40 frequency bands, and finally taking the logarithm to obtain the feature map.
2
Sources
60%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMel频谱图是模拟人耳听觉感知非线性特性的时频表示,通过STFT后经Mel刻度上均匀分布的三角滤波器组(通常80或128个)并取对数得到,相比原始波形将数据维度压缩约一到两个数量级71% similarUnverifiedWhisper将输入音频重采样为16kHz,通过25ms窗口、10ms步进的STFT生成频谱,再经80个梅尔滤波器组转换为80维特征向量序列67% similarUnverifiedAt an 11 kHz sample rate and a 1024-sample frame, one FFT bin is about 11 Hz wide, wider than the spacing between adjacent notes below A3, so it cannot be used for harmonic analysis.64% similarVerifiedAudio visualization is typically based on Fast Fourier Transform (FFT) for frequency-domain analysis of audio signals.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1010399API
curl https://kongchang.com/api/v1/knowledge/claims/1010399MCP
get_claim(id=1010399)