Automating QC is the key to quality in file-based workflows
File-based workflows are supplanting earlier, less integrated digital workflow architectures in cable plants everywhere. File-based environments streamline content production and delivery, a crucial benefit given proliferating end-user platforms and the diverse formats that support them. Cable operators need a workflow that can store, process and distribute what amounts to thousands of “channels” of content simultaneously and a file-based workflow brings together all these functions.
With a file-based workflow comes the requirement to ensure content is always ready for immediate distribution. This is “content readiness” and it means having all content available in a form and condition suitable for distribution. It means content that’s free from objectionable distortion and equipped with the right markers, metadata and resolution for the best possible service, regardless of receiving platform.
This level of content readiness requires a rigorous quality control (QC) program that inspects every individual file—from 15-second ads to an HD movie—as it enters the cable plant and when stored for playout.
The sheer volume of content dictates some kind of automated QC scheme.
Automated quality control
An automated QC strategy that operates 24/7 and pinpoints errors so technicians can spend their time fixing rather than finding errors is clearly a “must have” for operators as they deliver more VOD, insert more local ads and provide multi-screen TV Everywhere services. Automated QC is more thorough and consistent, whether it’s evaluating one short commercial or an entire archive of old TV westerns. Equally important, automated QC systems can detect errors that are invisible to the human eye: metadata that doesn’t match the play time, for instance, or syntax errors that may crash set-top boxes.
What could go wrong?
File-based video, like all other media, is susceptible to flawed source material.
Many errors originate in the baseband product including improper camera or microphone levels to amateurish, outof- gamut “homemade” graphics. These errors are part of the file and must be rectified eventually.
Incorrectly encoded files are another problem area. These can arise from a faulty encoder that produces syntax errors due to buffer overflows or similar technical issues.
Transfer errors are another hazard.
Problems can attach themselves during any transfer from system to system, including the ingest transfer itself. Even content that is sent correctly can be received with errors due to interruptions or faulty equipment.
Applying automated QC is particularly important to three file-based applications, namely VOD, local ad insertion and adaptive bitrate (ABR) streaming.
Video on Demand
VOD implementation begins at the headend with the acquisition of content via satellite receivers or by file transfers.
Incoming VOD assets can undergo automated QC checks when received or while stored on servers. These checks include: • Identifying file integrity problems: These can occur during the automated pitching-catching process at ingest with files becoming corrupted or truncated.
• Identifying format compliance issues: Files must be checked for compliance with CableLabs’ specification CEP 3.0 (content encoding profile) to ensure that bit rates are within the prescribed range, that PID numbers are correct (e.g. 481 for video or 482 for audio) and that GOP lengths are also correct.
• Ensuring compliance with government regulations: Are the mandated CEA 608/708 captions present? Is the audio loudness within the limits set by the CALM Act? In every case these characteristics must be not only checked but also corrected when errors are found.
• Verifying that machine-readable metadata agrees with measured values such as frame rate, displayed picture size and play time.
Local ad insertion is a lucrative moneymaker for cable operators, which earned them more than $4 billion in 2011. The challenge of the ad insertion workflow is managing myriad codec, container and media formats. Local commercials are delivered in the form of tapes, DVDs, memory cards or digital files, whereas national ads typically arrive as files from a CDN. The ingest process must ensure that ads are stored in a consistent file format. In every case, QC ensures continuity of visible quality as programming switches back and forth between ads.
At ingest the content gets transcoded to the format the splicer needs to insert it into the broadcast feed. Because the splicer can only switch ads in and out on GOP (group of pictures) boundaries, quality checks must ensure that the clip contains an integral number of GOPs so that there are no references to GOPs preceding or following the clip.
Ultimately the ad server should only contain correctly formatted ads ready for insertion. The active broadcast feed passes through the splicer, which inserts local ads at the correct time.
Ad Insertion QC
The majority of errors in the ad insertion workflow are quality-related. Local advertisers simply can’t afford the production values of a big national campaign, so compromises are made. For instance, a small video house might not have a loudness meter, resulting in audio levels that are too high. Or there may be over-compression, causing blockiness and artifacts.
It’s an operator’s prerogative to reject this content or to accept and repair it.
Regardless, it’s crucial to have a QC regime that can detect these flaws.
One of the most bothersome errors in ad content is format mismatch. For example, ads may be submitted in 4:3 standard definition, even though it will be broadcast in HD. This causes letterboxing as the subscriber’s equipment tries to make the best of a 4:3 aspect ratio on a 16:9 screen.
One way to avoid this is to set up separate ingest paths for SD and HD deliveries.
Additionally, the QC tool should verify that the video fills the active image and that letterboxing/pillarboxing effects are not a permanent part of the file. It’s also becoming common practice to check the Active Format Descriptor (AFD) to confirm that the playout code is correct.
Cable content distribution used to be straightforward, with scheduled programming going out to subscribers via operator-provided set-top boxes. But that market has changed forever. The set-tops are still there, but so are tablets, phablets, phones, laptops and PCs. To be competitive, operators must offer streaming services for all these devices.
The solution for this multi-faceted challenge is over-the-top (OTT) delivery using adaptive bitrate (ABR) streaming.
HTTP Live Streaming (HLS) and Smooth Streaming are two of the leading streaming architectures in use today. Both HLS and Smooth Streaming rely on transcoding each asset to multiple bit rates—high, medium and low. Thus there are several coexisting versions of each asset. Standard HTTP network transport protocols are used for client (subscriber) access.
An ABR implementation can continually change the bit rate, switching it higher or lower depending on network conditions and the requirements of the viewing device.
ABR demands are complex, and optimizing bit rates is only part of the story.
A file must be broken into short segments rather than being transmitted in one full-length delivery to the subscriber.
A dedicated “segmenter” tool creates segments that are just a few seconds long, usually six seconds or so. The reason for this layout is to provide boundary points at which the bit rate can be switched.
As bandwidth changes, the client player requests the optimum bit rate and the switch occurs at an appropriate Groupof- Pictures (GOP) boundary.
ABR quality control issues
A key part of the QC regime in an ABR workflow is to verify the content’s readiness for segmenting. It is important to establish early in the workflow that the file can be correctly divided into usable segments. In this context “content readiness” implies that IDR frames are embedded at regular timing intervals to form segment boundaries. The QC checks should confirm that stored content— whether movies or commercials—has the IDRs in place and timed correctly.
Another check confirms picture quality in an environment in which content is stored in numerous versions, each with a separate bit rate. Are the lowest bit rates still delivering acceptable picture quality? For that matter, are all of the rates providing the expected image quality? It’s not practical to perform this test on every piece of content in real-time; instead it is best used in designing an effective ABR workflow. Using a picture quality analysis system, encoder performance can be finetuned to get the most out of each bit rate.
In H.264, for example, there are numerous settings that can be adjusted to maximize picture quality on the receiving device.
File-based workflows are key
Expanding VOD offerings, maximizing revenues from local ad insertion and delivering multi-screen TV Everywhere services are at the top of every cable operator’s “to do” list and file-based workflows are key to the success of effectively managing all this new content. Ensuring the quality of these services is equally important as operators face mounting pressure from satellite, telco and OTT competitors.
Fortunately, automated QC has emerged as an approach that can scale to accommodate the exploding amount of content and which can operate with the speed required in an on-demand environment. ■