Zixi Product Updates
Release Notes: Zixi Broadcaster Version 18
15 min
v18 6 46249 (28 july 2015) version 18 (v18 6 46249) of the zixi broadcaster and zixi edge compute is now generally available and includes the following new capabilities and bug fixes note not all features are available in zixi edge compute improvements improved overlay blend performance on html5 overlay improved source switching on no merge failover group to now detect an exit point (i or p frame on the source stream) before switching to the new stream improved memory allocation and performance in the tr 101 analyzer v18 5 46206 (8 july 2025) improvements added additional stream stats, such as cc errors within a segment, on hls inputs on input stream stats json api endpoint v 18 3 46065 (24 april 2025) improvements added feature for including web audio in the html overlay during transcoding updated no merge failover mode to use 'search window' parameter for inactivity detection, instead of 200m fixed time updated cpu detection for graviton to be “graviton2” or “graviton3” instead of “neoverse” added support for billing codes on transcoded sources, bytes going to the transcoder are counted added support for keeping rtp headers in srt output when srt input contains rtp headers added support for chroma sub sampling up conversion for software transcoding, including with html overlay v 18 2 45997 (13 march 2025) improvements added feature for including web audio in the html overlay during transcoding added feature to convert hls inputs into a valid mpeg ts which meets cbr and pat/pmt repetition requirements added support pid mapped and/or error concealment on transcoded sources added support for seamless switching on audio only streams the switch will occur on the first audio frame detected and is supported on aac, aac latm, ac3, eac3, opus, mpga1, mpga2 and aes3 added support for seamless stream switch on rtp/udp push outputs updated ui message about license expiration to appear two weeks before the expiration date added display of sdt and nit table in tr 101 analyzer added support for bind to local port for srt push output and srt pull input updated jitter alert on rtp/udp sources to turn off after 5 seconds added support for srt inputs on the same port from different hosts with unique remote stream names added support for time and date table (tdt) / time offset table (tot) in dvb transport stream on pid 0x14 in tr 101 analyzer improved performance of html overlay when using software transcoding added ability to disable metadata injection from ui for non transport streams, such as atsc 3 0 stltp (studio to transmitter link) streams added option to delay stream by one segment for hls inputs in order to inject scte 35 packet in advance enabled via checkbox in the ui updated scte 35 report to calculate wall clock time = scte 35 pts – scte 35 packet pcr updated failover groups to reattempt a source previously considered faulty if there are no alternatives improved upload performance to s3 using aws recommended settings added support billing codes for transcoded sources updated srt library to v1 5 4 added support for capturing scte 35 mpu format identifier on hls inputs added ability to dejitter rist sources either by arrival times (default) or by rtp timestamps, as is done on rtp inputs added checkbox to keep rtp headers in input streams over a unix socket on output streams over a unix socket, added checkbox to keep rtp headers and a checkbox to add new rtp headers to the stream v18 0 (30 dec 2024) upstream protocol metadata the zixi broadcaster now saves stream health statistics on zixi input streams as metadata stored in the transport stream when a stream is sent through multiple zixi broadcasters, health statistics from each stage in the transport chain will be saved and if packet loss has occurred or cc errors have been introduced, it can be tracked down to a specific stage of the transport the saved metadata can be viewed by hovering on the status light on inputs in the input tab in the ui bitrate reduction on congested network links the zixi broadcaster and zixi edge compute can now reduce stream bitrate on zixi protocol push and pull outputs when the network link is congested without transcoding it reduces the bitrate by selectively removing p frames and b frames while keeping the audio and metadata in sync with the video settings to enable this feature, called “stream thinner”, are on push and pull output dialogs in the ui when advanced mode is enabled for automatic pull outputs there is a global config as well the stream thinner latency setting should be set to the gop of the stream irdeto digital rights management (drm) integration the zixi broadcaster and zixi edge compute can now retrieve an encryption key from irdeto drm servers and encrypt adaptive hls streams using that encryption key this capability allows customers to use business rules enabled by irdeto to allow or block playback of encrypted streams settings for this feature including credentials for irdeto servers can be found on the adaptive live stream dialog in the ui when advanced mode is enabled both fairplay and widevine are supported video preview using mpeg ts over http the zixi broadcaster has a new option for video preview which plays mpeg ts streams directly in the browser via an http connection from the broadcaster with this feature, vlc is no longer required for video preview however, not all codecs are supported by browsers so the vlc option is still available for maximum flexibility note, automatic http pull outputs need to be enabled in the broadcaster configuration to use this feature netint transcoding accelerator support the zixi broadcaster now supports accelerated transcoding using netint quadra vpu accelerator cards the price vs performance vs quality of the netint quadra vpu offers advantages over gpu and cpu based transcoding this feature is currently in preview while we www zixi comwork closely with netint to optimize performance netint quadra cards can be purchased for bare metal servers and will also be available on akamai connected cloud in early 2025 scte 35 replacement via pois to support placement opportunity information system (pois) scte 35 replacement methodologies, the zixi broadcaster can now remove all existing scte 35 messages on an input stream while maintaining the existing scte 35 pid and then inject new scte 35 messages on the existing pid or a new custom pid the feature works in conjunction with an esam agent provided by zixi and 3rd party pois servers protocol updates listed below are improvements and bug fixes to the various protocols supported by the zixi broadcaster zixi implemented performance optimizations for zixi protocol hls updated behavior of injecting scte 35 on cue tags in hls inputs now injects scte 35 message in advance, pointing to the pts of the first video frame following the tag, instead of immediate injection at the time of transition updated behavior on hls inputs to increment splice event id when converting cue points to scte 35 messages srt updated srt implementation to allow multiple srt outputs to share the sameport by using the stream name to uniquely identify them transcoding updates improvements to transcoding in the zixi broadcaster are listed below added ability to recognize and show nvidia ada cards on the status page in the gui added support for netint as discussed above added ability for user to select the transport stream id (tsid) on an input being transcoded added display of transcoded sources in red if the original source is encrypted and not able to be decoded tr 101 analysis must be enabled added support for transcode of audio only sources implemented performance optimization in transcoder additional improvements added failover group events to logs exposed dar and display resolution in the tr 101 analyzer exposed inter packet gaps maximum, average and standard deviation on udp outputs, instead of timing histogram added ability to synchronize the timeline of file/http sources by specifying iso time to use for pts at beginning of the file added count of 608 and 708 closed captions to input stream statistics data can be used to determine whether or not input streams contain 608 captions and/or contain 708 captions, as may be required by law currently the data is only available via the api added ability to delay inputs in memory by up to three seconds added 'steal time' into reported cpu usage added support for delivering payloads of up to 1486 bytes, which may be necessary for payloads such as atsc 3 0 implemented performance and memory optimizations in tr 101 analyzer