Real-time defect detection is essential for advanced in-process quality inspection systems, particularly for identifying complex crack formations in small-scale products through automated methods. Acoustic emission (AE) enables the rapid detection of crack initiation, making it a highly effective technique for modern manufacturing processes. This study presents a comparative analysis of AE signatures generated during the microblanking process of magnesium and SK-5 steel. Experiments were conducted using a proprietary 5 kN micro-forming machine on 0.1-mm-thick SK-5 steel and 0.5-mm-thick magnesium plates, representing materials with contrasting mechanical properties. This study evaluates and compares various signal processing methods, including the fast Fourier transform (FFT) and the wavelet scattering transform (WST), in terms of their effectiveness in identifying crack initiation events. The wavelet scattering transform exhibited enhanced sensitivity for detecting crack initiation in both materials, with distinct signature patterns identified between the magnesium and SK-5 steel specimens. However, there were limitations for softer materials or when the original signal amplitude was lower than the system noise level. Overall, this comparative analysis provides valuable insights for developing material-specific acoustic emission monitoring systems for microforming applications. © 2026 Faculty of Engineering, Universitas Indonesia. All rights reserved.