We introduce the analysis of variables measured outside a direct current electric arc furnace that are related to the arc coverage and its electric energy efficiency. We derive an on-line reference to the arc coverage suitable for control decisions. In order to do so, we study the furnace shell vibration frequency spectrum as structure-borne sound sensed by a microphone. The data is analysed using real time software tools to obtain the sound signal spectrum associated with arc coverage likewise related to the optimal performance of the furnace. A statistical test is performed to prove the relationship between sound and electric energy efficiency.