In Exchange 2013/2016, Exchange Search includes built-in support for indexing many file formats. In order to enable or disable specific file formats for Exchange Search, we use the Set-SearchDocumentFormat cmdlet (which is only available in on-premises Exchange 2013/2016).
When we disable a file format for content indexing by Exchange Search, contents of the file become unsearchable by Exchange Search clients such as Outlook Web App, Outlook in online mode and In-Place eDiscovery.
Let us say that we do not want to include ZIP files. To do this, we simply run the following cmdlet:
Set-SearchDocumentFormat ZIP -Enabled $False
If we disable indexing for a supported file format, such as in the example above, items containing an attachment of that file type are not considered unsearchable. When we perform an In-Place eDiscovery search, and we select the option to include unsearchable items, only items that are actually unsearchable are returned. Items that were not searched because the associated file format is set as unsearchable are not returned.
Please note that JPG and GIF formats are not really used even if enabled. Exchange indexes their metadata but will not scan/OCR the image itself. Exchange will deliberately skip over images and mark the items as partially processed.