Photo courtesy of unsplash.com
In SimilarWeb’s July 2019 rankings of the most used websites, three of the top 10 were dedicated video streaming websites, while another three offered video sharing without being the website’s primary purpose.
The first video sharing website launched in 1997. ShareYourWorld.com was different from the modern video sharing giant YouTube since internet speeds of its time were often not fast enough to stream videos. After ShareYourWorld shut down in 2001, the next popular video sharing website was the South Korean Pandora.tv. This one found more success, primarily localized in South Korea. Four months after Pandora’s creation, the first video streaming website was launched: YouTube. YouTube was different from predecessors because it used video transcoding, which allowed users to upload their own videos and made it easier for users to watch videos. YouTube exploded in popularity and is now the second most used website in the world. This is made possible by advancements in video encoding and video transcoding technologies.
When a camera takes a picture, the image is broken into individual squares, called pixels. Using the commonly used RGB color space, each pixel can be described by three numbers between 0 and 255, each number representing the amount of red, green and blue in that pixel. Since it takes one byte to store the number range, each pixel must consist of at least three bytes. The standard image resolution, described by two numbers representing the count of horizontal and vertical pixels, is 1920 x 1080, meaning that each image contains slightly more than 2 million pixels, or 6 million bytes. To make a video, most cameras record 30 images per second. At this rate, a DVD would be filled in less than a minute. This is why video encoding is so important.
Video encoding uses algorithms that find similarities within sections of images or between images in a video and exploit those similarities to describe the images using less storage. This process of decreasing the size is called compression, and it is used for many types of files, even some of the simplest text-only files. A file compressed using lossless compression can be perfectly reconstructed by the user. To further decrease file sizes, files can use lossy encoding. This means that the encoded file likely can’t be perfectly reconstructed, though it usually remains close enough that most people can’t notice a difference. Several prominent video creators have created videos illustrating the data loss by uploading a video to their preferred video sharing website, downloading it, then uploading the downloaded version, repeating the process hundreds and even thousands of times.
Transcoding is a similar process, but instead of starting with a raw video, it starts with an already encoded video and converts it to other formats. This process is what enables a website to offer the same video at different qualities. When a website detects that a user’s internet connection is not fast enough to play the video they are trying to watch, the website begins sending the same video transcoded to a lower quality. This lower quality video has less data in it, meaning less information needs to be sent, allowing for the user to continue watching despite a slower connection.
Over 300 hours of video are being uploaded to YouTube every minute. Video sharing websites have become so popular that it is difficult to imagine the internet before transcoding made these websites possible.
