In the cost-conscious broadcast engineering industry, the luxury of independent production resources for each delivery platform — TV, Web, handheld and mobile — must eventually evolve to an integrated, adaptive production environment. This will enable the increase of ROI in infrastructure, reduce production time and maximize content repurposing.
Prior to the advent of digital, file-based production, content was acquired, edited and disseminated in linear real time. Compression and file formats were not an issue. Now, audio and video enter, exist and exit the system in an increasing variety of formats.
Production today is migrating to tapeless nonlinear workflows. Random access to stored content enables any portion of audio or video to be used by more than one person at any given time. There are two dominant methods of content management through the production workflow. Naturally, each method has its pros and cons.
One method uses an elaborate scheme where editing, generally cuts-only, is performed on low-resolution video and produces an edit decision list (EDL). The high-resolution video, kept on the main media servers, is assembled into a finished piece by applying the instructions in the EDL. A benefit of this approach is that high-resolution content is stored in only one location and file transfers and network traffic are kept to a minimum. A drawback is not working with full-resolution content and the absence of inter-segment transition effects.
The other approach downloads full-resolution video and audio to a local workstation. Once work is completed, the finished piece is published — stored in the main media server and logged by the asset management system. This technique allows editing to be performed on full-resolution content; frame accurate edits with elaborate transition effects are possible. The price, however, is that multiple copies of content must be stored and managed. Sufficient network bandwidth, at peak production times, can also become an issue.
A thorough analysis of production workflows and infrastructure resources is necessary to find the most efficient mix of both methods. Even the smallest local broadcaster can save time and money with an optimized production workflow. If the inherent workflow and equipment challenges are overcome, the benefits realized will grow exponentially when an integrated production workflow that supports multiplatform delivery is optimally engineered and implemented.
A starting point
In order to examine the design challenges of integrated production for multichannel distribution, consider a hypothetical broadcaster who wants to deliver content over the air, across the Internet and to handheld and mobile devices. At times, content delivery will be simultaneous; at other times, programming will be available later over non-broadcast delivery channels.
To facilitate an analysis of the core infrastructure and workflow, both audio and video production will be modeled as consisting of three fundamental steps: acquisition/ingest, storage (as a file) and the subsequent transcoding for dissemination. At which stage in the workflow the actual transcoding occurs will depend on detailed production process analysis. The optimal location in the integrated production workflow to transcode must be identified — this will be a key design precept.
The design goal is to minimize artifacts produced by format conversions; therefore, a fundamental design philosophy is to keep format conversions to a minimum.
There are a number of places the format conversion can occur including: when content enters or leaves the MAM, by discrete intelligently located conversion equipment and/or in an integrated production and/or master control room. The ultimate goal is for the content to be transferred to the transmission equipment in the appropriate format for the target delivery channel; source content may be transcoded concurrently and multiple formats will need to be managed.
The integrated infrastructure
Regardless of the workflow and supporting system architecture, the heart of all file-based production is asset storage and management. The fact that content can be acquired in many different formats complicates both the technical process of format conversion and the conceptual workflow design — should content be stored in its original format or should it be converted to a single “house” format?
This makes storage architecture an important design consideration. Keeping copies of content in multiple formats occupies large amounts of storage, and broadcast-reliable storage can be expensive.
Tracking multiple formats is required if more than one copy is stored, and it is important to consider how efficiently the asset management application can catalog the same content in a number of formats.
Proxy generation, used in editing or to aid content searches, may be done by the MAM system or ingest server. Again, this is dependent on the overall system architecture. Proxies, even low-resolution, low-bit rate images, will need a large amount of permanent storage. Once content is moved offline or archived, proxies and metadata will be the only way to find content. Tapes are physical and can be found, but files are bits and are lost without a tag. Proxies must reside in the MAM system forever.
Many choices have to be made. Dynamic demands on the infrastructure must be taken into account (the best architecture for Saturday afternoon sports production may be completely inappropriate and inefficient for a nightly news program). Another consideration is to take advantage of production downtime. For example, while a two-hour theatrical release is on the air, plan for scheduled maintenance, repairs and upgrades.
A detailed analysis of how content moves through the production workflow will reveal ingest, storage and conversion requirements. This should generate a design strategy that works 24/7. Try to stick to the fundamental premise that no format conversion is a good conversion, so the number of conversions should be kept to a minimum or zero. Also, storage costs money, and large files in HD formats eat up bandwidth.
Another concern to keep in mind is the SDI routing infrastructure. Plans must be made now for adequate scalability in the future. Can the emerging 3Gb/s format, SMPTE 424/425, be supported with existing equipment? Will implementing faster than real time SDI transfers be required to meet content transfers at peak production times?
An initial working hypothesis of a workflow could be: convert as content is ingested; store in highest quality format, the “house” format; convert as little as possible; and downconvert if necessary only for channel delivery. To maintain the highest video quality whenever possible, however, content may also be stored in its original format. Obviously, there is a complex interdependence between system resources and workflows.
As an experienced broadcast engineer understands, there is more than one way to implement an integrated production infrastructure that supports multiple delivery channels. But no matter how the design challenge is approached, small workflow improvements can have a large impact.
Design options that can be leveraged to support the myriad of audio and video source formats that feed the production process and core infrastructure will be the topic of the next Transition to Digital.