Akamai Diversity

The Akamai Blog

Simplifying the Transcoding Conundrum

This is the third installment in a series of posts that discuss various challenges of online video and how Akamai's Sola Media Solutions can be used to address those challenges.

One of the most common challenges that we hear from customers - of any type - is the uphill battle they face when attempting to prepare content for delivery to multiple devices. Even dividing devices into categories seems daunting: mobile phones, smart phones, tablets, phablets, laptops, desktops, connected TVs, game consoles ... the list can go on and on. And it only gets worse when you introduce operating systems and brands into the mix.  

There are a lot of ways to go about getting content to multiple devices, but one way or another the content is more than likely going to require transcoding. And when you are talking about adaptive bitrate (ABR) streaming, transcoding is a necessity. To date, transcoding has been an arduous process. There are a myriad of input possibilities - video format, audio format, file container, source file quality - that make up yet another list that goes on and on (seeing a trend here?). Once you've got the file into the transcoding system, what outputs do you want? Or maybe more appropriately asked, what outputs do you need? Considerations like bitrate, number of renditions and frame rate are all in play at this point.  
Earlier I mentioned ABR, and how it requires transcoding. Let's dig into that. ABR, by design, requires multiple renditions - or copies - of the content; renditions that the media player can dynamically switch among during streaming playback based on factors such as network congestion, the device's available resources and video quality. These renditions are created by transcoding the content. The most common method of creating these rendition sets is to take a high-quality source file and run it through the transcoding system multiple times, each time changing the output settings to match the rendition required. This can mean five to 10 or more separate processes. This also means crucial elements, such as key frame alignment, aren't always achieved. Each rendition may locate the key frames at slightly different points, which means that when the stream switches, the video will initially play poorly or not at all - resulting in a bad experience for the viewer.

All of this begs the question, "If transcoding is so hard, how are folks doing it?" Unlike transcoding, the answer is simple: Not easily.

As you may well know, some of the world's largest and most recognized media companies have been trusting Akamai to deliver high-quality video for over a decade. During this time we've observed that nearly 70% of user playback quality issues were a result of improper content preparation. Based on all of this complexity, that shouldn't come as much of a surprise. When Akamai introduced our VOD Transcoding service earlier this year we knew that it had to produce high-quality content, optimized for streaming, and it had to do it simply. And simple doesn't just mean an easy-to-use, web-based UI (which it has). It means that users can get as granular as required and as they'd like on their and rest assured that a uniform rendition set will be created every time. It also means dropping a file into a watch folder and getting a streaming URL in return - not a series of files that you then have to manage. 

The bottom line here is that transcoding can be a difficult and confusing part of the media preparation workflow - but it doesn't have to. To learn more about Sola Vision VOD Transcoding and how it can help you simplify your overall workflow, visit Akamai.com/Sola or sign up for a free demonstration from one of our product experts.

Barrett Mononen is a product marketing manager at Akamai.