eyeson_logo_white

MCU+ Data Sheet

The eyeson MCU+ service is an enhanced software cloud MCU (Multi Control Unit, based on eyeson Single Stream Technology) that provides an API for group video conferences adding custom data & media content into the video stream.

MCU Versus MCU+

Table of functions

The superiority of eyeson's MCU+ functions

 

eyeson’s video conferencing technology is based on the MCU mode which supports server-side data processing and collects all streams from all participants in a central entity which then delivers a single stream to each participant.

This is aimed to guarantee high-quality video combined with a stable meeting experience.

 FUnction

other

MCus

eyeson

mcu+

Add/Remove participant 
Hide/Show participants
Use screen casting/presentations
Enable broadcasting/streaming
Synchronized view ("all see the same")
Set custom video layout
Set custom positioning of participants 
Cloud-based snapshots & recording
Set back-/foreground layer per video frame
Add live or playback video streams
Inject any data on custom positions

MCU+ transcodes video, audio & data streams into one single stream

 

transcoding-mcu

1 | DECODING/CONVERTING

Incoming video streams from each client are decoded into a raw video format. Incoming data stream are converted into pictures or single frames.

2 | MIXING/LAYOUT MANAGER

Video frames from incoming video stream and media conversions are mixed into one single frame. The layout manager can position each source separately.

3 | ENCODING/STREAMING

The mixed stream is encoded in one video stream and send to all connected client, as well used for recording and live streaming services.

Let our server do the computinG

Comparison on the client-side with 9 participants on the server

 

Save resources at the client-side

 

MCU makes it possible for a multipoint audio/video conference to be controlled and moderated from a single location.

With eyeson MCU+ the video streams are mixed on the server. Therefore the CPU-/GPU-Usage and the bandwidth usage are as low as possible – with just 1 outgoing and 1 incoming stream and with the video streamed from the client to the stream "as it is" without the need to do a lot of computing on the clients.

eyeson MCU+ can handle up to 100 participants, while other MCU are mostly limited to 16 participants.

 

OTHER
MCUs

eyeson

MCU+

Outgoing streams 9 1
Incoming streams 8 1
CPU usage (%) 35-50 5-24
GPU usage (%) 10-35 5-7
Upload MBit/s 0,5-2,5 0,3-1,0
Download MBit/s 1,5-7,0 0,3-1,5
Screencast MBit/s + 0,4-2,0 + 0,0
Recording resource Local Client Cloud Server

(Test performed with Windows 10/Chrome on Intel Core i5-4590, 16 GB Ram, NVIDIA GeForce GT 730, 4 GB Ram

Technology overview

  • Basic technology: MCU/SFU hybrid, WebRTC compliant
  • Video codecs: vp8/h264 video codec
  • Audio codecs: opus/aac audio codec
  • API: REST (restful service)/websocket control interface
  • Recording: direct webm-recording
  • Streaming: DASH/RTMP(S) live-streaming client
  • RTMP: RTMP(S) client (live-stream forwarding) & RTMP server (video injection)

High-performing core software

eyeson single stream technology core MCU service written in C/C++ including RTMP live streaming and recording functions.

MCU controller written in one go

It is the main component to manage the cloud instances. Implements cloud API’s for different cloud providers and manage the registration of the on-demand resource allocated meetings started by users

Bandwidth requirements

The connection for up- and downstream of each client automatically adjusts to the real-time streaming capabilities of the network the client is located in.

Upstream

  •  40 kbps audio-only
  •  900 kbps with video
  • 1500 kbps with presentation/screen-casting

Upstream maximum (adaptive)

Downstream

  •  1280x960 @ 1500 kbps @ 25fps
  • 640x480 @ 800 kbps @ 25fps
  • 320x240 @ 400 kbps @ 25fps

Regardless of the number of participants
(up to 100)

MCU+ is based on eyeson's patented Single Stream Technology

  • Low constant bandwidth needed from clients.
  • Bitrate does not depend on the number of participants, nor on the amount of data sent and received.
  • The client is directly connected to the conference server, not to the other participants. Server-side recording is available.
  • Devices using H323/SIP protocols can participate in conferences.
  • Maximum compatibility ensured by using standards supported by a maximum number of devices.
  • No expensive server is required to mix multiple media streams into one because eyeson takes care of it
  • Up to 100 participants in one meeting plus recording and live streaming.

No expensive devices needed

The server sends a single stream, which allows access to video conferences even through low-performance devices (smartphones, mobiles or tablets).

Data integration with patented technology

eyeson MCU+ can integrate data (triggered by the API) into the video stream

In-depth tech insights

Each cloud instance is controlling a wide range of functions for group video conferencing and data processing such as video and audio transcoding, mixing, streaming, recording, data & media injections and some more. 

Clients are streaming audio and video from a web browser to the cloud instance, where all incoming streams are decoded. The decoded audio and video streams are mixed into one single video stream, which is used to send back to the clients, to store as a recording or to stream to broadcasting services (e.g. YouTube). (read more)

cloudmcu-sst-v3

Transparent All-in-One Pricing

Independent of the number of participants and features, we charge only based on the duration of the meeting.