Multipoint Control Unit (MCU), or videoconferencing bridge

Multipoint Control Unit in brief MCU is a device or software that enables multilateral connections in classic solutions for video conferencing. It uses the transcoding process to match the received audio / video signals and create, for example, a picture divided into 4 parts in the case of a four-party videoconference.

The transcoding process requires very high computing power so that it can be executed in real time and therefore requires a dedicated DSP processor per participant (the so-called port) or in the case of programmable MCUs, at least 2 vCPU per participant. The higher the resolution of the multilateral connection and the more participants in the meeting, the higher the demand for computing power, the simultaneous increase in the MCU price and the requirements for the physical platform in the case of software MCU (where the hidden cost for a comprehensive solution arises).

Traditional MCU-based video conferencing architecture is derived from times when the medium was ISDN and there were no technical possibilities for large-scale videoconferencing. This technology has been adapted for video conferencing via the Internet, but it has a number of limitations:

  • transcoding introduces additional delays that affect the quality of connections, especially in networks without QoS
  • Multiple coding / decoding is not a lossless process, so it has a reduction in quality
  • HD or FullHD ports are expensive, so you often have to compromise in a lower quality, videoconferences in UltraHD quality are beyond the technical capabilities of the current MCU
  • poorly scalable solution, in the case of physical MCUs you need to make a “backup” purchase because the expansion often involves exchanges with other devices, program MCUs are heavily saturated with resources, which means that the expansion often involves the need to expand the server infrastructure

 

Let’s analyze an example of connecting four video terminals in one video conference using MCU.

1. The image is recorded by the cameras in each of the 4 locations.

2. The image is then coded (compressed to save transmission links) and sent to the MCU.

multipoint connection unit mcu

3. Here, decoding (decompressing) and merging the received signals.

4. In order for the merge to be carried out, transcoding occurs first (scaling of received signals to the desired resolution and frame rate)

5. Recoding the signals and sending back to the video terminals where the final decoding takes place.

mcu bridge

Let’s analyze an example of connecting four video terminals in one video conference with no MCU , using TrueConf Server.

1. The image is recorded by the cameras in each of the 4 locations.

2. The image is then coded with a unique, multi-layered codec with SVC extension and sent to TrueConf Server.

3. TrueConf Server sends back to the participants as many streams as there are participants in the meeting, separating from the received streams the appropriate video quality for each participant independently (on the principle of elimination of redundant layers). The resource occupation for this intelligent process of routing audio / video signals is negligible.

4. In the end stations, the received signals are decoded once.

svc

Our Clients