- Classification of room volume from reverberant speech signals can be useful in acoustic scene analysis applications such as forensic audio, rescue, and security. Previous work required the room impulse response (RIR) to explicitly either estimate or classify the room volume. In this work different approaches are investigated to apply to reverberant speech room volume features that are usually extracted from the RIR. The room volume is then classified with a pattern recognition based-system that uses the room volume features. Room volume classes are trained using Gaussian mixture models and room volume is classified using log-likelihood ratio with a background model. Feature selection is used to achieve minimum equal error rate (EER) of room volume verification. An EER of 22.38% is achieved when RIR features are extracted from abrupt stops in speech.