Agile Live 3.0.0

First release of Agile Live in 2023.

Agile Live 3.0.0 is the third major release of the Agile Live system. This release contains a major reorganization and improvement of the REST API as well as support for ingesting interlaced sources and JPEG encoding on Nvidia GPUs.

See the Release documentation for more details.

Reorganized REST API

The REST API has seen some major reorganizations in this release to make it more user friendly and simplify some operations. One such improvement is the management of Control Connections between Control Panels and the Control Receivers. These connections are now categorized under a new endpoint called /controlconnections. Creating, updating and closing these connections are now much easier, as each connection get its own unique UUID at creation and the same endpoints are used, no matter if the connection is originating from a Control Panel or a Control Receiver. Similar changes have been made to the Stream endpoints.

Support for interlaced sources

“Interlacing is dead” has been declared for years now, but despite this, many broadcasters are still locked into interlaced workflows due to legacy systems. To simplify the movement towards a progressive future we have added support for ingesting interlaced sources in Agile Live in this release. De-interlacing is implemented for both the Intel and Nvidia encoder and it is also possible to make JPEG thumbnails of interlaced sources. The system will automatically identify and de-interlace interlaced sources, so no additional procedures are required from the user.

In order to keep the maximum video quality, it is recommended to always use progressive scan when possible.

JPEG thumbnail encoding using Nvidia GPUs

This release adds support for JPEG thumbnail encoding using a Nvidia GPU alongside the old Intel GPU implementation. This is especially important in cloud environments where the Intel CPUs usually don’t feature an integrated Intel GPU to do the JPEG encoding. The user can explicitly choose which implementation should encode the thumbnail in the REST API call, or leave the decision to the Ingest to select whichever implementation is supported by the hardware.