Aventus is a cloud-based end-to-end media processing solution for live events, 24x7 live linear channels, and VOD content. Complete with flexible acquisition options, scalable transcoding, adaptive bitrate packaging, robust security, and ad opportunities, Aventus helps you scale and monetize your content for millions. Provided as a managed service, Aventus simplifies the complexities of the changing media industry.
Table of Contents
Release date: 7/1/19
Build information: 8.0.1416.0 (General Public Release)
This feature allows a channel to be configured to ignore “segmentation_descriptor ID” when accompanied by a “splice_insert”. Instead, the “splice_insert ID” will be used.
Enabled support for “segmentation_upid_type” == 0x0D to enable multiple UPID types, allowing the signaling of local distributor ad availabilities (useful for dynamic ad insertion).
Enhanced how Aventus handles channel configuration changes for AC-3 and EAC-3 (e.g. switching from 5.1 to stereo) which had been creating audio and video to become out of sync on some occasions.
Table of Contents
Release date: 7/3/18
Build information: 7.9.30616.1 (General Public Release)
Functionality to properly handle multi-audio manifest generation during instances where more than one language track is present.
Some transient stream errors could cause the channel to incorrectly enter a "Ready" state if they happened during startup. We now detect such cases and restart the channel automatically when that happens, so it won't transition to Ready until it actually is.
Reduce small segments due to SCTE tolerance (i.e: Coalesce SCTE messages to a specified tolerance). The tolerance becomes the minimum duration of a segment that is generated as a result of SCTE.
Table of Contents
Release date: 1/30/18
Build information: 7.8.30196.1 (General Public Release)
Enhanced functionality to switch between two Kafka source feeds.
Corrects potential discrepancies with parsing and calculating UTC from the SCTE-35 signal.
In instances where a blackout occurs, allow SCTE-35 placement signalling opportunities to optionally only insert sample ads instead of provider ads.
Upon editing a channel within the Aventus UI, the AudioSourceID will be synced forward when updating the JSON.
Updated ability to turn off / on the sending of the HTTP header during publishing. By default, this feature will now be turned off.
Relaxed the Aventus hard-coded timeout to now be configurable via the JSON file based on customer request.
Table of Contents
Release date: 12/19/17
Build information: 7.7.29954.1 (General Public Release)
Aventus now provides support for publishing of mixed 30p and 60p streams in Amazon Web Services
Table of Contents
Release date: 10/24/17
Build information: 7.6.29558.1 (General Public Release)
Aventus has improved layer alignment in supporting mixed frame rate configuration. Customers will now be able to support scenarios with lower framerates at lower layers and higher framerates at the top most layer to ensure highest quality to the viewer. For example, the following is a typical scenario a customer may wish to use as their channel setup.
Aventus supports key rotation based on configured time interval for Fairplay. Specifically, Aventus integrates with the Atlas DRM service to offer seamless integration and content protection to our customers.
Aventus is capable of ingesting SCTE-35 markers even if the markers arrive late. Aventus respects the SCTE-35 spec today such that any SCTE-35 marker should arrive with a pre-roll of 4-seconds earlier than its intended schedule time or may face the risks of be discarded. The newly added windowing concept provides finer control based on current time.
Improved robustness on handling source loss when time domain change occurs. During source loss, there exist a condition that can cause the audio and video transcoders to get out of synchronization and cause stalls in publishing. The improved model now automatically handles time domain change more gracefully and ensures monotonic increase order.
In instances where multiple HLS packagers are being used, this guarantees that video will be copied prior to encryption to create seamless integration.
Table of Contents
Release date: 6/15/17
Build information: 7.5.28975.1 (General Public Release)
DEPRECATION NOTICE: This will be the last platform release that includes support for HDS streams
Based on a customer request this addresses downstream consumption when a short segment is produced due to SCTE-35 segmentation. This feature enables Aventus to offer a separate mode of segmentation, allowing better accuracy in video segmentation while reducing the number of short frame segments.
There are two new values that can be set for the HLS Stream Package Parameters:
This feature causes Aventus to maintain a hard cap on the number of retries during HLS segment publishing. Instead of leveraging automatic retry algorithm, we also offer a hard limit to further fine control this behavior to reduce flooding the CDN with requests.
It is now possible to set the total number of attempts the HLS publisher will try to send an item to a publishing point:
Content Identification messages can now be passed into the HLS manifest. The Content Identification message will stay at the top of the DVR Window until a Program End or a new Content Identification SCTE-35 has been received.
Added additional error handling and more robust implementation for sending SCTE-35 messages in addition to self-healing ability if an ingest error occurs.
Aventus now has the option to send AES-128 keys to multiple publishing points. The AES-128 key can be configured to push to a primary and secondary location, helping with redundancy and eliminating a single point of failure. If the origin region goes down entirely there will be a backup and the client will still receive both video and encryption keys. The key is considered valid when successfully pushed to either location in a 1+1 redundant failover scenario. There is a single playback URL that points to the publishing point that had a successful key published. For each key that is generated it is pushed simultaneously to all of the listed publishing points. Whichever one returns a success first will be the primary publishing point. Please note, if multiple publishing points exist, you can only have 1 playback URL.
Friendlier representation of the SCTE-35 history API that shows the in-band SCTE-35 messages in a more readable format.
A set of new buttons in the Aventus UI to allow graphical injection of chapter start and chapter end SCTE-35 messages.
Added logic to the HLS Packaging to keep the Elementary Streams moving forward at the same pace. Before this change, Packaging read the streams in round robin. If a stream, like audio, fell behind but produced all of the missing data, Packaging would not read the stream to catch back up. This would cause additional permanent latency in the pipeline for GOP Synchronization to occur and if the time delta was too great the Elementary Streams would never synchronize because data would be thrown away to keep the queues at a manageable level. This is an “always-on” feature.
RTMP now synchronizes the Elementary Streams before publishing. This fixes Akamai RTMP Transmux where layers are not aligned. Also, the RTMP timestamp is now assigned at one place before publishing so all RTMP streams have the same time.
Updated the Aventus internal decoder / encoder package for improved functionality in the cloud environment, which results in increased picture quality in lower bitrate streams.
New functionality which will allow for the updating of an Aventus slate while the channel is still on-air.
Updated functionality which will now re-download an updated Aventus slate when a channel is restarted.
Updated HLS publishing to handle additional DNS failure scenarios to allow better self-healing to avoid CDN disconnection.
Ability to host Aventus in AWS
Improvements in “Signal to Noise” ratios have led to a reduction in “false positives” and improved reporting accuracy for overall channel health
MD5 content header validation which supports segment validation prior to publishing
Aventus now supports multiple differing slate images per segmentation type
Introduced a new whitelist capability to improve flexibility in controlling whether the flag should be honored for different segmentation type coming from the source
Aventus now supports EXT-X-PROGRAM-DATE-TIME tags in HLS manifest with an optional millisecond accuracy
Aventus will not start the channel until at least one key is published successfully. Once a key is published, if the new key fails to publish for any reason, we will continue to use the known good key until a successful publish occurs. In addition, we will allow the keys to be pre-published and cached for use to ensure consistent playback.
Changes have been made to the channel timeline pipeline to give better diagnostic coverage. This will allow active monitoring work to move forward without being bounded by Aventus release cycles, as well as allow users better and more immediate response information across the board.
Provided Aventus the ability to listen to Content Identification (0x01) SCTE-35 messages
Aventus will now try up to 20 times to publish a file. The timeout for time to start a connection and the time to send the key is capped at 1 second
Added functionality to not allow Aventus to send Audio/Video Samples to RTMP that are not in order.
In cases where HLS channels implement more than a single DRM content protection scheme, Aventus now publishes multiple, distinct manifests in addition to "compound" ones where all DRM tags coexist in a common manifest. This helps ensure player app compatibility for cases where client players reject HLS playlists containing tags they do not recognize. Aventus will not publish duplicate TS segments, meaning each specific manifest refers to a common set of TS segments. Work with iStreamPlanet Operations team members for additional details.
Aventus now supports HTTPS publishing of "AES-128 CBC" DRM content decryption keys. This complements existing Aventus support for publishing to FTP endpoints.
RTMP timestamp can now be reset whenever the connection between Aventus and Twitch is dropped. To enable this feature, edit the channel JSON for a Twitch packager or publisher by adding the following StreamPackage field:
"Package": {
"StreamPackages": [
{
"Format": "RTMP",
"Parameters": {
"ResetTimestampOnReconnect": true
},
"StreamCollection": [
{
"VideoStreamIndex": 0
},
{
"VideoStreamIndex": 1
},
{
"AudioStreamIndex": 0
}
]
}
]
}
This feature is undergoing continued development. In cases where segment alignment is critical to your live streaming scenario, Aventus will now perform additional synchronization during channel start-up to ensure all HLS renditions are initiated across a common IFrame. When enabled, Aventus may require up to six additional seconds to deliver HLS segments to your CDN. Thus, this behavior is currently disabled by default until we achieve a solution which does not incurr additional channel startup latency. The following Channel JSON enables this behavior and is declared at the root.
"ServicesConfiguration": {
"VideoTranscoder": [
{
"GopsToSkip": 2
}
]
}
The order in which partial presentations are declared in channel JSON is preserved in the channels Aventus publishes. This makes it easier for operators to reconcile JSON Channel Config with the Aventus UI and channel API results.