For the photoelectric detection technology of music signal, an improved harmonic reconstruction algorithm was proposed to reduce the system noise while retaining the timbre of music. Firstly, the priori signal-to-noise ratio of the signal was calculated, and Wiener filter for noise reduction processing was constructed. Then a comb filter design scheme was proposed to address the characteristic of multiple harmonic components in music signals. After performing the inverse Fourier transform on the power spectrum of the music signal, the number of sampling points within one fundamental period of the music signal was determined. The number of samples was used as the node parameter of the comb filter, and the harmonic enhancement of the music signal was thus accomplished. The noise reduction effect was theoretically analyzed through the calculation of the priori signal-to-noise ratio, and then the noise reduction effect of different noise reduction algorithms was obtained through experimental verification. The results show that the algorithm can perform denoising well. Compared with the traditional harmonic reconstruction algorithm, the performance of reducing the log-spectral distortion of music signal is improved by more than 50% when the signal-to-noise ratio is -5 dB, 0 dB and 5 dB, and the timbre of the music signal can be well preserved. This study provides a reference for timbre analysis and evaluation in performances of different types of musical instruments.