Visual-level filters provide the most powerful filter conditions in Power BI exclusive of custom filter conditions specified in DAX expressions. Unlike report and page-level filters, DAX measures can be used in visual-level filter conditions, such as Net Sales greater than $5,000. Additionally, top N filter conditions can be implemented referencing a column and measure that are included or excluded from the visual per the Top N visual-level filters section following this example.

In the following example, a table visual of customers has been filtered by the Internet Net Sales and Internet Sales Orders measures:

Specifically, the visual only displays items (customers) with more than $8,000 in net sales and more than three sales orders. Per the sales-ranking measure, certain customers that meet the net sales condition are excluded based on the sales order condition. Unlike the top N visual-level filter condition, filters based on measures, such as the following conditions, are only applied when items (for example, customers, products) are displayed on the visual:

By removing the two customer columns, the remaining measures (Internet Sales Orders, Internet Net Sales) are not filtered by the visual-level filter conditions. In other words, the visual-level filters based on measures are only applied against the dimension column or columns in the visual, such as Customer Name or Customer Postal Code. 

Although analytically powerful, report authors should exercise caution with visual-level filters. From a usability standpoint, reports can become confusing when visuals on the same report page reflect different filter conditions. Additionally, executing complex filter conditions against large or dense report visuals can result in performance degradation. If a complex filter condition is repeatedly needed at the visual level, it's likely the case that the dataset should be modified to include some or all of this logic.