Most consumer devices, such as PCs or smart phones, can downscale video on-the-fly, as media content streams in from the Internet. But consuming high definition data on a low definition device adds expense at both ends, because it consumes far more bandwidth than the device can display. In fact, watching a 1080p video clip on a 360p device wastes as much a 90% of the data.

Transcoding is a job best performed at the service provider (or content source, if they are one and the same). An outgoing process can make intelligent, cost effective, on-the-fly, trade offs that go beyond individual device resolution.

Smart source transcoders are fully device aware. Not only can they adjust for bandwidth and device resolution, but for the recipient codec (AVI, MPEG, MKV), color palette or even display in black & white. All together, a smart transcoder can result in significant cost savings.

Consider, for example, a YouTube enhancement that you may have noticed this month. YouTube began adjusting the default video resolution to the default player window within your browser. For at least some devices and browsers, the YouTube server has sufficient information about your initial display resolution to optimize the transcoder and resulting media stream.

Let’s peek into the transcoder decision process offered to content admins. You might think that this is a concern only for big organizations that serve up lots of data. But more often, lately, it intersects with casual consumers who operate media-streaming clouds from a home-based NAS (such as a router-attached PogoPlug), a remote PC, or from an ISP or hosting service.

The Amazon Elastic Transcoder is a service component of their cloud storage and service suite (AWS, EC2, S3. Refer to this page for the full set of Amazon cloud services). The cost of transcoding depends on the resolution (e.g. from HD to SD, etc), the service region and the minutes of video content converted. For example, a minute of SD transcoding in the Northeast USA region is about 0.45¢ (just to be clear, that less than ½ penny).

This morning, Engadget highlighted improvements to the media transcoder, specifically, its ability to transcode audio as well as video. But, for me, the video tutorial seems to point to a deficiency rather than a bragging point…

A Wild Duck Opinion . . .

For organizations and individuals seeking to serve AV in user-demanded formats, the Amazon Elastic Transcoder barely scratches the surface. In my opinion, the video demonstrates a fine set of underlying processes, but from a user perspective (even from the perspective of a content admin), it is a crude and unfriendly beginning. The overall process fails the most important user tests: simple, automatic and transparent.

I envision this changing rapidly to something that requires no thinking by either the content admin or consumer/user. This is where market evolution must take us: On-the-fly transcoding at the optimum resolution for each user device—as it is demanded. Streams should be scaled even lower for individual users that have elected to conserve bandwidth at their end (using only their device interface). Decisions to prioritize pipelines or retain a cache of transcoded versions (instead of simply creating them anew for each user-demand) should never be a burden to the content admin. S/he shouldn’t be required to think about these things any more than an elevator needs an operator to decide at which floor it should hoover. I don’t mind the admin controls, but such forced decisions are evidence of either a start up alpha-process or a poorly implemented design. I expect that the feature rich controls of Amazon Elastic Transcoder falls into the latter category.

Instead, these should be driven by optimization algorithms that weigh current demand (at each resolution or quality) and the cost and availability of storage -vs- bandwidth. As with the most recent iteration of YouTube, users will ultimately trust that a video watched on a 360p device is not rushing toward their data cap by transmitting at 1080p. (And in fact, their own device will warn them that 90% of incoming data is being discarded, unless they are also saving to memory).

Cloud-based video transcoding, or cloud transcoding, is a service available from Qencode which gives the user access to all the latest formats, codecs and resolutions for improved video quality and reduced transcoding costs.