US-12627846-B2 - Just in time transcoder system, method and architecture for audio and video applications

Abstract

Methods, systems and architectures for video and/or audio processing and distribution are described. More particularly, a just in time transcoder/transcoding (JITT) is employed. If a matching optimal asset segment with calculated optimal asset attributes requested by a client device is not stored in a server, the server is configured to invoke the just in time transcoder (JITT) to produce the matching optimal asset segment by transcoding from one or more assets stored in an asset storage device to fulfill the request of client device.

Inventors

• Kumar Ramaswamy
• Jeffrey Allen Cooper
• John William Richardson
• Joaquín Sergio LOPEZ CERRO

Assignees

• igolgi Inc.

Dates

Publication Date

20260512

Application Date

20240917

Claims (12)

1. 1 . A system for delivering at least one media asset, comprising: at least one server; at least one client device; at least one just in time transcoder; the at least one server is configured to deliver a manifest comprising a single default profile to the at least one client device, wherein the single default profile is a top most profile with highest resolution and bitrate available for the at least one media asset in the system; the at least one client device is configured to calculate optimal asset attributes based at least on available bandwidth and screen resolution of the at least one client device, and is further configured to request, from the at least one server, an optimal asset profile with the calculated optimal asset attributes; and wherein: if the matching optimal asset profile requested is the single default profile, the at least one server is configured to deliver the matching optimal asset profile to the at least one client device to fulfill the request of the at least one client device; and if the matching optimal asset profile requested is not the single default profile, the at least one just in time transcoder is invoked to produce the matching optimal asset profile requested by transcoding from the top most profile to fulfill the request of the at least one client device, and wherein a processing speed of the transcoding from one format to another is at least multiple times faster than real-time arrival rate of the one format of the at least one media asset at the at least one just in time transcoder.
2. 2 . The system of claim 1 , wherein the at least multiple times are at least 10 times.
3. 3 . The system of claim 1 , wherein the transcoding from one format to another is by changing one or more of following attributes: 1) bit rate, 2) frame rate, 3) video compression codec, 4) audio compression code, 5) resolution, 6) overlay graphics, and 7) output pixel bit depth.
4. 4 . The system of claim 3 , further comprising at least one asset storage device, wherein the at least one asset storage device is configured to store only the top most profile.
5. 5 . The system of claim 4 , wherein the at least one server, the at least one just in time transcoder, and the at least one asset storage device are coupled together, and the at least one client device is coupled to the at least one server through a content delivery network.
6. 6 . The system of claim 1 wherein the at least one client device is coupled to the at least one just in time transcoder through an edge cache.
7. 7 . A method for delivering at least one media asset in a media delivering system, comprising: delivering, by at least one server, a manifest comprising a single default profile to at least one client device, wherein the single default profile is a top most profile with highest resolution and bitrate available for the at least one media asset in the system; calculating, by the at least one client device, optimal asset attributes based at least on available bandwidth and screen resolution of the at least one client device, and requesting from the at least one server an optimal asset profile with the calculated optimal asset attributes; and wherein: if the matching optimal asset profile requested is the single default profile, delivering, by the at least one server, the matching optimal asset profile to the at least one client device to fulfill the request of the at least one client device; and if the matching optimal asset profile requested is not the single default profile, invoking the at least one just in time transcoder to produce the matching optimal asset profile requested by transcoding from the top most profile to fulfill the request of the at least one client device, and wherein a processing speed of the transcoding from one format to another is at least multiple times faster than real-time arrival rate of the one format of the at least one media asset at the at least one just in time transcoder.
8. 8 . The method of claim 7 , wherein the at least multiple times are at least 10 times.
9. 9 . The method of claim 7 , wherein the transcoding from one format to another is by changing one or more of following attributes: 1) bit rate, 2) frame rate, 3) video compression codec, 4) audio compression code, 5) resolution, 6) overlay graphics, and 7) output pixel bit depth.
10. 10 . The method of claim 7 , further comprising, storing by at least one asset storage device, only the top most profile, and the transcoding from one format to another is transcoding using the stored the top most profile.
11. 11 . The system of claim 1 , wherein the optimal asset profile represents a media asset segment.
12. 12 . The method of claim 7 , wherein the optimal asset profile represents a media asset segment.

Description

RELATED APPLICATION This patent application claims the priority and benefits of U.S. Provisional Application No. 63/629,540, filed on Nov. 6, 2023, the disclosure of which is incorporated by reference herein in its entirety. FIELD OF INVENTION The present principles relate to video and/or audio processing, and more specifically, to Just-In-Time-Transcoding (JITT) video and/or audio systems and methods. BACKGROUND INFORMATION The surge in video streaming services and the increasing reliance on video in all web-related content necessitates the need for affordable, high-quality streaming technology. The varying usage scenarios and deployment situations greatly influence the streaming system costs. For instance, cloud DVR systems can place a substantial financial strain on storage, but the expenses associated with the Content Distribution Network (CDN)/edge distribution architecture should not be overlooked either. Factors such as the application (e.g., DVR, VOD, Live), the quantity of channels or number of media assets stored, the user count, and the CDN or distribution network's architecture/bandwidth all contribute to shaping the system's design and cost centers. While advancements in various technologies have boosted streaming performance and cost-efficiency, the deployment of streaming systems remains a considerable investment. Content-aware transcoding and just-in-time packaging (JITP) are two technologies that have positively impacted streaming deployments. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates an example of an Adaptive Bitrate (ABR) streaming system; FIG. 2 illustrates an example of a static ABR ladder and a content aware ladder; FIG. 3 illustrates an example of an ABR system with a Just-In-Time-Transcoder (JITT) to provide lower profiles according to an embodiment; FIG. 4 illustrates an example of a Just-In-Time-Transcoder (JITT) ladder compared to standard static or content aware ladders; FIG. 5 illustrates an example of a Just-In-Time-Transcoder (JITT) virtual ladder vs. traditional static and content aware ladders; FIG. 6 illustrates an example of an arbitrary profile request and generation using a Just-In-Time-Transcoder (JITT) system and architecture; FIG. 7 illustrates an example of using a Just-In-Time-Transcoder (JITT) system and architecture for legacy clients; FIG. 8 illustrates an example of server-client interactions using a Just-In-Time-Transcoder (JITT) system and architecture to serve arbitrary profile requests; FIG. 9 illustrates an example of server-side optimal profile delivery using a Just-In-Time-Transcoder (JITT) system and architecture; FIG. 10 illustrates an example of a Just-In-Time-Transcoder (JITT) system and architecture for edge caches; and FIG. 11 illustrates an example of a software stack to optimally use GPU resources for a Just-In-Time-Transcoder (JITT) system and architecture. FIG. 12 illustrates an exemplary workflow process which may be utilized for an exemplary Just-in-time Transcoder. FIG. 13 illustrates a block diagram of an exemplary apparatus 1300 in which various aspects of the exemplary embodiments may be implemented. DETAILED DESCRIPTION In this disclosure, we will introduce and discuss “Just-in-Time Transcoding/Transcoder” (JITT). We will explain how a JITT product sets a new standard for affordable video streaming deployment, eases generational codec transitions and enhances video quality and performance for the end-users in the process. An example of such a JITT product is a Blazar Just-in-Time Transcoder being designed by igolgi Inc. of Princeton, NJ. In addition, another benefit of the Blazar JITT is that it reduces energy consumption per stream, helping to reduce carbon emissions. ABR Streaming Background A video streaming system employs a technique known as adaptive bitrate (ABR) encoding. The audio and video streams are first chopped up into small time chunks known as segments. These segments may last anywhere from 2 to 10 seconds (this is only a guideline, not a requirement). The file segments are then encoded into a variety of resolutions and bitrates, referred to as a “bitrate ladder”. The audio signal is similarly divided into several audio channels and bitrates. The goal of ABR is to deliver the appropriate version of video and audio segments based on the capabilities of the receiving device and the available bandwidth. The entire streaming operation is thus converted to a series of file transfers (that can be supported by any Internet server) from the headend to the client devices. Rate selection is done by each individual client based on a manifest of rates provided by the ABR encoding head end. FIG. 1 shows a high-level view of an ABR streaming system 100 with a cloud DVR. The video/audio content is transformed into various profiles and kept in storage. When a device asks for this content, the Origin Server initially sends a manifest to the requesting device detailing the available profiles. The device, in turn, selects the profi