Speech Emotion Recognition (SER) is a critical subfield within affective computing aimed at automatically detecting and categorizing emotional states conveyed through speech signals. SER’s main objective is to identify speakers’ underlying emotional nuances, including sadness, fear, happiness, rage and neutrality. This discipline has become pivotal in psychology, HCI and artificial intelligence research, facilitating advancements within emotion-aware computing, opinion mining, and psychological health assessment. Convolutional Neural Networks are used in SER to analyze sequential data, including audio spectrograms. CNNs were first developed for image recognition. Among these architectures, VGG16 stands out for its capacity to automatically extract discriminative features from raw audio data. Integration of CNNs with complementary modalities like text or facial expressions in multimodal systems enables the capture of richer emotional representations, thereby enhancing the performance of emotion recognition tasks. This abstract provides an overview of recent advancements and methodology in SER utilizing VGG16, highlighting its application in extracting and interpreting emotional cues embedded in speech signals.