The paper introduces the reassignment method, which improves the readability of time-frequency and time-scale representations by repositioning signal components to better reflect their true locations. This method, originally applied to the spectrogram 15 years ago, is generalized to all bilinear time-frequency and time-scale distributions. The reassignment method moves the values of a representation away from their computed positions to enhance the localization of signal components. The method is theoretically analyzed and demonstrated for various time-frequency and time-scale distributions, including the Wigner–Ville distribution, smoothed pseudo Wigner–Ville distribution, and the scalogram. The reassignment method is shown to preserve key properties such as time and frequency shift invariance, energy conservation, and perfect localization of chirps and impulses. It is also shown to be computationally efficient and applicable to a wide range of signal representations. The method is validated through numerical examples, demonstrating its effectiveness in improving the readability and accuracy of time-frequency and time-scale representations.The paper introduces the reassignment method, which improves the readability of time-frequency and time-scale representations by repositioning signal components to better reflect their true locations. This method, originally applied to the spectrogram 15 years ago, is generalized to all bilinear time-frequency and time-scale distributions. The reassignment method moves the values of a representation away from their computed positions to enhance the localization of signal components. The method is theoretically analyzed and demonstrated for various time-frequency and time-scale distributions, including the Wigner–Ville distribution, smoothed pseudo Wigner–Ville distribution, and the scalogram. The reassignment method is shown to preserve key properties such as time and frequency shift invariance, energy conservation, and perfect localization of chirps and impulses. It is also shown to be computationally efficient and applicable to a wide range of signal representations. The method is validated through numerical examples, demonstrating its effectiveness in improving the readability and accuracy of time-frequency and time-scale representations.