1.4. Lite versus Full IP Variants
For the full variants, the IPs transmit some control information together with video data over the streaming interfaces. For the lite variants, write all the control information to IPs using a processor interface.
You may still select run-time control via a processor interface when using the full variant of IPs (for example, to update clipper offsets or mixer settings). The control information about the size and type of incoming video is always handled automatically using streamed control packets.
You can turn on or off lite mode in the GUI for most of the Intel video and vision processing IPs. When off, the IP is a full variant. You can interface between full and lite variants of IPs using the protocol converter IP. However, system design is easier if a video pipeline standardizes on one version of the protocol.
When choosing between full and lite consider the following advantages:
For full IP variants:
- Ease of use. Full variants of the IPs handle more of the system control aspects automatically.
- Video and Image Processing Suite interoperability or upgradeability. Interfacing or modifying existing systems based around the Video and Image Processing Suite IPs to the full variants is easier, as these handle video in a similar way.
For lite IP variants:
- Building a system primarily using one of the Warp, 3D LUT or TMO IPs or a system requiring interoperability with third-party AXI-Stream video steaming IP. The Warp, 3D LUT, and TMO IPs use the lite variant of the protocol and do not pass control along the streaming interface.
- Device area. Expect to save 100-200 ALMs per IP for lite variants.
Regenerate IPs with either version of the protocol and compare overall system fMAX and area before deciding.