Abstract: The purpose of this study is to investigate the perceptual confusions of Putonghua consonant initials among native Cantonese and Putonghua speakers under different signal-to-noise ratio conditions. The study focuses on the relationship between the acoustic characteristics of consonants and perceptual confusions, while examining the influence of native language background and signal-to-noise ratio on the perception of Putonghua consonant initials. Using a confusion matrix method, 13 Cantonese participants and 22 Putonghua participants are required to identify consonants in speech stimuli with different signal-to-noise ratios. The results of repeated measures analysis of variance indicate a significant difference in the ability to perceive Putonghua consonants between the two groups of subjects with different native languages. Additionally, white noise is found to reduce the perception ability in both groups. The common perception results are as follows: (1) Sonorants exhibit better resistance to noise. (2) There are less perceptual confusions between fricatives with prominent spectral peaks and those with flatter spectra. (3) Due to the masking effect on transitional cues and obstruction pulses of stops, white noise can lead to more prominent perceptual confusions among unaspirated stops, as well as among aspirated stops, aspirated affricates, and fricatives. (4) The tongue position of following vowels has a certain degree of influence on the results of consonant perceptual confusions. The above results show that in terms of the physiological (or linguistic theoretical) aspects of pronunciation, categorizations based on place of articulation, manner of articulation, and voicing fall short in explaining the perceptual confusions of Putonghua consonant initials in noisy environments. Rather, spectral features and duration offer more insightful explanations for these confusions. The role of acoustic features in consonant perceptual confusions sometimes surpasses that of phonetics.