We've released a new version of the /media/analyze v1.1 API.

Changelog

• With this release you can now get more insight into the scenes present in your media. The Analyze API response includes a breakdown reporting Speech, Music, Silence, and Other sections so that you can get an understanding of the percentage and num_sections identifying the count of discrete sections attributable to each type. A section is defined to be a consecutive block of a section type that is at least 0.5 seconds long.
• The presence of clipping has been identified as the number of discrete sections within the content. You can find this as the num_sections attribute and the num_events is no longer provided.
• Some general performance improvements have been made for media files with a duration greater than 30 minutes. If your media is not long enough to be analyzed you will get an invalid-input-file error.
• The start and end timestamps for detected silence segments have been removed. If you relied on this functionality, please contact Support and we can discuss a beta preview of how this functionality will be changing in a future release.
• The loudness validation results will only be returned going forward if you specify a specific loudness profile in your request.