A modified additive synthesis method using a combination of additive synthesis and source filter modeling has been proposed for musical instruments. When applied to a trumpet, the results are perceived as being less synthetic than additive synthesis by itself. Since harmonics as well as inharmonics are important for resynthesized sound quality, the authors segment the frequency range into bands each of which contains a single harmonic with inharmonics. A sinusoidal model is not sufficient to represent both harmonics and inharmonics. A more complex model of a quasipolynomial in each of the frequency bands is proposed. The quasipolynomial can be obtained as the response of a special linear filter to the unit impulse. Thus each harmonic is obtained using a source-filter model. The inputs of this filter are periodic pulses with time-varying amplitudes and slightly varying periods. Two listening tests were performed to evaluate the perceived quality.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.