I understood that payment is combined of the audio and video minutes. Can I index only audio?
Sure, you can index video or audio files and apply AIs only for the audio, only for the video (vision) or for both. Pricing is based on the duration of the input file. Indexing is charged for audio analysis, video analysis, or both. See more information here - https://azure.microsoft.com/en-us/pricing/details/media-services/#analytics