What’s new in Watcher ver. 21.12

  1. Improved facial recognition. In some cases, the person in the photo was not identified and the video analysis module recognized that individual as someone else. This issue has been resolved.

  2. We continue to develop the IRIS firmware (secure firmware for IP cameras that works in any browser, without installing additional plug-ins or special settings, and allows to implement a maximum of connections at the same time). The behavior of interrupted broadcasts was fixed (due to available RAM running out, something that shouldn’t happen).

What’s new in Flussonic ver. 21.12

  1. Opus codec is now supported for MPEG-TS containers

    Our client, the UGC platform, implements one-to-many communication by receiving and distributing video streams via the low-latency WebRTC protocol.

    Using a third-party transcoder, they transcode these streams into multiple qualities to cover all types of devices. Audio via WebRTC is always streamed in the Opus codec (this is a free and effective audio codec that compresses better and faster than others).

    opus

    To distribute the streams in different qualities to the audience at the output, the service had to transcode the audio twice. First: from Opus to AAC (to be able to transfer multiple video tracks via TSHTTP to distribution servers). Then back from AAC to Opus (to distribute multi-bitrate streams to ultra-low latency end devices via WebRTC).

    The key task was to find a way to receive audio from third-party transcoders using the Opus codec, but packaged in a container that would support multiple video tracks.

    There were 4 ways to post videos on Flussonic. RTMP was removed because it didn’t support more than one video track. Multiple video tracks can be streamed over RTSP, but it didn’t fit either (it turned out that a third-party transcoder didn’t support multiple video tracks over RTSP). Streaming via WebRTC was also not suitable as it is a transfer protocol between browsers. After investigating all the options, it was found that it was possible to transfer Opus audio + multi-bitrate video packaged in an MPEG-TS container. Therefore, in version 21.12, we taught Flussonic to accept multi-bitrate video and audio streams in the Opus codec from third-party transcoders.

  2. Added adaptive bitrate support for WebRTC (Ultra-Low Latency Data Transfer Protocol). The WebRTC ABR option is useful for one-to-one and one-to-many communication in online education, game streaming, and video conferencing (VC).

    Responsiveness works for both publishing and playback. The video quality will adjust to the Internet speed of each user.

    PD WebRTC Flussonic has one important difference from other solutions. When receiving a post via WebRTC (thus with an audio track in the Opus codec), Flussonic can give different protocols to different viewers for playback. For example, some viewers will watch video via WebRTC, others via HLS (where the sound is in another codec, AAC). This is possible without unnecessary configuration, within a single stream name. Flussonic’s built-in packer can add another audio track with the AAC codec to the existing Opus audio track.

  3. The ARM64 docker image is now supported (develop on Apple M1 and deploy on Amazon, Google Cloud, and Microsoft Azure).