一种用于估计板混响物理参数的多分辨率频谱图方法
A multi-resolution spectrogram approach for estimating the physical parameters of a plate reverb
浏览论文内容
中文总结 AI 辅助
本文提出用ResNet-18结合多分辨率频谱图与相位信息，从脉冲响应估计板混响物理参数，在5000个IR验证集上达到0.02920的平均NMSE。
中文摘要 AI 辅助
采用ResNet-18图像分类模型，从脉冲响应的录音中确定板混响的物理参数。该模型通过使用从提供的脉冲响应（IRs）计算出的归一化和降采样的多分辨率频谱图来适应参数推导。为了细化输出位置的预测，频谱相位响应也作为额外的输入通道加入网络，因为多个输出位置可能对高阶谐振模式产生相同的幅度响应。在5000个IR验证集上，我们的模型在所有参数上实现了平均归一化均方误差（NMSE）为0.02920，其中最低的平均NMSE出现在参数yo（0.00228）、Ly（0.00347）和xo（0.00574）上。
英文摘要
The ResNet-18 image classification model is employed to determine the physical parameters of a plate reverb from a recording of the impulse response. The model is adapted to derive parameters using normalized and downsampled multi-resolution spectrograms computed from the provided impulse responses (IRs). To refine the prediction of the output location, the spectral phase response is also included as an additional input channel to the network since multiple output locations can give the same magnitude response for high-order resonant modes. On a 5000 IR validation set, our model achieves an average normalized mean squared error (NMSE) of 0.02920 across all parameters, with the lowest average NMSE occurring for parameters yo (0.00228), Ly (0.00347), and xo (0.00574).
发表机构
- University of Minnesota Duluth（明尼苏达大学德卢斯分校）
- University of Rochester（罗切斯特大学）
机构由 AI 辅助整理，请以论文原文为准。
