Speech recognition systems are the most natural way of communication between the machines and humans. Human beings have a potential to speak and interpret the spoken speech utterances. Machines have no intelligence so Natural Language Processing assumes a significant part, progressively uses machine human interaction in real time applications. Sentiments are the expression of emotions like anger, happiness, sadness, fear, suspense, neutral etc. where they become tied to a social object. It has been observed that spoken speech is more informative than the text based speech. Sentiment analysis from speech recognition system is an active area of research in digital signal processing used to identify the sentiments of speakers using speech rather than text. Deep learning is a robust machine learning method with lots of applications in various fields such as hand-written character recognition, image processing, computer vision and natural language processing. In this paper, deep learning methods employed in the sentiment analysis from speech recognition system have also been discussed and analyses. The proposed work can be utilized to conduct a marketing research for a product by businessman, social media marketer, healthcare department, etc. This work will be also helpful to understand the mental health of the employees of any organization. Lastly, a comparative analysis of various deep learning classifiers was discussed in this paper.