Digital Video Applications for Broadcast, Cable, Content Distribution, Education and Related Markets

Video Transcoding & Transrating - Video in 31 Flavors!

Tweet

 

In this post, we will discuss the basics of video transcoding. My apologies in advance, for what has turned out into a super-long blog post! I hope to make up for the excessive length by providing a somewhat comprehensive treatment of the topic, and showing some examples from a broadcaster's or cable MSO's  perspective.

Video transcoding is the process of digitally transforming a piece of multimedia (video and audio) content from one format to another. Such transformation can include any one or more of several attributes of digitally encoded content... encoding protocol, bit rate, resolution, frame rate, picture size.

Transcoding is used widely to take a piece of encoded content, and expand its use beyond the use/s initially anticipated at the time the video was originally encoded. For instance, a commercial that is created and encoded for insertion into a broadcast network's programming in the US or Canada, might need to be delivered over an IPTV channel, or over cable TV in an Asian market, such as India, Japan or Korea. In such cases, the required transformations might include changing  bitrates, resolutions, standards (e.g., NTSC to PAL), frame rate, format (e.g., MPEG2 to MPEG 4 Part 10 or H.264). The transcoded content can then be utilized for further monetizing, syndication or other business purposes. If a file that is originally encoded for the specific intent of use in a broadcast studio environment, with the view that it would be played out in a play-to-air server, is then sought to be delivered to much wider (read, worldwide) audiences, it might be syndicated to a whole lot of national or regional broadcast or cable channels, or it might be made available over an IPTV or streaming video service.

Though entertainment video, advertisements and historical footage have very definitive and high value business applications for, and demands of, video transcoding, I think Electronic News Gathering (ENG) and dissemination is one of the most quintessential applications for transcoding. The same piece of video footage (most of the time, short-form footage, lasting at the most a few minutes or half an hour), is very rapidly disseminated all across the world, over a multitude of networks and display devices. Increasingly, the use of transcoding has become a basic, essential aspect of maximizing the utility of high value video content.

The power structure in the entertainment, news, education, history, sports and related content areas is increasing reverting back to the owners of unique, high value content. Such owners of content, in an attempt to make their content entirely agnostic to delivery means and consumption means, are resorting to transcoding, and theby ensuring that their valuable content is ready to be effectively, maximally and efficiently consumed by all viewers, located anywhere, over any type of distribution (broadcast TV, cable TV, IPTV, internet, intranets), and possessing any type of viewing device (television sets, computer screens, jumbotrons, PDA/handheld devices, tablets). Content is King, and the owners are trying to keep it so.

That is not at all to say that video distribution and delivery companies do not add extremely important value in getting the content from the content owner, to the content consumer (or viewer) in a manner that is effective, instantaneous,  efficient, and highly cost effective. These businesses, broadcasters, cable operators, overbuilders, IPTV providers, Internet services, wireless companies (offering video delivery over 3G or 4G handheld and cellular devices) are where an overwhelming majority of users of transcoding equipment are located. These businesses, of course, bring huge scale to the distribution though their networks.

Transcoding of video content is also used in several other less obvious cases. One major set of such cases, are those where the content owner, syndicator or distributor is seeking to actually enhance the original content (beyond rendering it capable of being distributed in different formats, over different network or physical layer alternatives, or to be displayed on different devices). Such enhancements include the addition of metadata or annotations that enable the content to be more easily controlled by VCR trick modes or machine control equipment, for non-linear editing and manipulation devices, and for adding different types of interactivity to content.

Another set of applications of transcoding a previously encoded video file, relate to the need to add certain additional information to the video, such as logos, watermarks or data to enable digital rights management. Transcoding is very much conducive to such operations, and it enables the protection, identification and enhanced monetization of previously encoded content.

Technically, the age-old view of transcoding is one wherein the encoded file is decoded, and then the decoded content is re-encoded using the desired transcoding parameters. This approach has significant problems inherent to it. It takes considerable computational resources, introduced punitive time delays, and very importantly, causes generational deterioration losses. That is, each time a piece of content is encoded, then taken back to its original form by decoding, then re-encoded, each lossy coding step introduces losses, which are cumulative.

I will leave out discussions of the technical aspect of different architectures and techniques of transcoding - those used to reduce bit rates, to change formats and standards, to vary resolutions, to optimize content for different applications such as streaming in this blog post. Should there be an interest in such a discussion, one of Vela's R&D engineers or I will cover those techniques in a separate post in the future.

From a layman's perspective, one basic technical characterization is that transcoding is somewhat akin to encoding in the first place.  A major difference is that when content is being first encoded, there is little data available to the encoder about the coding and other parameters specific to the video content being encoded. A comprehensive set of parameters, data and estimates are readily available to the transcoding process, however - and that enables the transcode process to operate at a significantly stepped-up level of competence and efficiency. For those of you who know something about two-pass or multi-pass encoding, transcoding is a bit like that. Two (or multi) pass encoding is a process wherein a piece of content is sequentially passed through two (or more) encoding "stages", with the first stage "sniffing" the data to estimate, measure, determine key parameters, then passing those stats on to the second stage which,  in a quasi-heuristic manner, presets its encoding parameters to the optimal ranges derived from the first pass. Such multi-pass encoding process can generate considerably higher quality encode results than single-pass encodes. As such, an effective transcoder can generate files that may be superior to that of the originally encoded video file.

The objective - and challenge - of each new transcoding techique, is to be able to effectively draw upon the statistics from the prior encode, to generate transcoded streams that provide superior video quality, that are ideally matched to the specific new use that is intended, and to allow the transcoding process to take place with minimal computational complexity or resource-demands.

Shown below, is a small clip of video that has been transcoded from one bit rate to another (4 Mbps to 1 Mbps), one format to another (MPEG-2 to MPEG-4 or H.264), one resolution to another (1080i to 720p). The former version is suited for playout from a cable headend. The transcoded version is suited for a wide variety of purposes, from broadcasting to IPTV to delivery over the public internet.

These clips (incidentally, all of them generated using Vela's Argo Navis encoding system) demonstrate how dramatically a user can transform the same video asset into a broad selection of programming content, suited to a wide variety of delivery and viewing methods and options.

Vela offers multiple advanced and highly cost-effective  encoding solutions that can be used conveniently to transcode and transform video content.

Here is the Original Clip which is 1920x1080i MPEG-2 at 50 Mbps

Here is a clip transcoded to H.264 and transrated to 1 Mbps

Here is a clip transcoded to H.264, transrated to 1 Mbps, and re-scaled to 1280 x 720P

Here is a clip transcoded to H.264, transrated to 256 kbps, and re-scaled to 320 x 240

That is it for this blog post. As always, please feel free to post your comments or tell us all about unique experiences you might have had with transcoding in your environment, or to ask questions that you might have.