For audio filter and equalizer design it is desirable to take into account the frequency resolution of hearing. Therefore, various specialized filter design methodologies have been developed, from which warped and parallel filters are particularly appealing options due to their simple design and good approximation properties. This paper compares the quantization noise of two different warped IIR implementations with that of fixed-pole parallel filters in single-precision floating point arithmetic. It is shown by simulations that the parallel filter provides the best compromise between quantization noise and computational complexity, since it significantly outperforms the series second-order warped IIR implementation in terms of noise performance, while requires less computational resources compared to the original warped IIR structure.
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=18712
(140KB)
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 Join the AES. If you need to check your member status, login to the Member Portal.
