This chapter develops low-delay filter banks (LDFBs) that decouple filter length from end-to-end latency while preserving perfect reconstruction (PR). Starting from the MDCT factorization, we extend the polyphase folding matrix into a cascade of simple structured stages: zero-delay matrices, which raise the polynomial order (and thus filter length) by one block without adding delay, and maximum-delay matrices, which increase both filter length and system delay efficiently. Practical Python realizations (offline and online) implement delays as stateful memory elements, and polynomial matrix products via efficient blockwise operations. The framework enables LDFBs with high stopband attenuation and controllable delay—useful for low-latency audio coding e.g., ELD-AAC [ISO/IEC JTC1/SC29/WG11 (2006) Enhanced low delay AAC (ELD-AAC)—MPEG-4 audio coding. ISO/IEC International Standard 14496-3, vol 3, Amendment 3; Schnell et al (2007) Low delay filterbanks for enhanced low delay audio coding. In: IEEE workshop on applications of signal processing to audio and acoustics, New Paltz], communication, and interactive DSP systems.