Metadata is simply data about data or information about information. It can describe a single content or in more complex situations, it describes a database schema. Metadata is there to provide context for the specific data it is describing. Metadata is also used to facilitate and assist understanding and management of data.
Documents and spreadsheets have metadata. They usually describe the file by naming the author of the file, the file size and the date it was created. Of course, more metadata details can be added to such metadata.
Audio files, too, have metadata. The singer of the song would be named as well as the composer. The date of release and the album from which it was taken are also enumerated. While some of these are automatically created with the file. Certain details can also be added to the metadata. In some instances, audio metadata could also describe how a specific audio are to be recorded. These are manually created metadata files written for audio programming.
Image metadata typically describe the settings used in capturing a specific data. It would also describe the kind of camera used in taking the photo. When the original file is processed using a photo processing software, more details are added to the image. All of this additional information is saved with the original image when it is saved in TIFF or EXIF files. But if the said image was taken in RAW format and processed in Adobe camera RAW, a separate metadata file is usually recorded in the XMP format.
It is therefore important to take note that while metadata can actually be embedded with the original data separate metadata files are also created in other instances. With either cases, metadata can also be created automatically or it can be generated manually.