Transcript for Video Codecs 101
| Time | Content |
|---|---|
| 00:00 → 00:03 |
♪ |
| 00:04 → 00:06 |
If you watch videos on the web, you use video codecs. |
| 00:07 → 00:08 |
[galloping hooves] |
| 00:08 → 00:10 |
But what exactly is a video codec? |
| 00:11 → 00:14 |
Codecs reduce large videos to more manageable sizes. |
| 00:14 → 00:16 |
This is called compression. |
| 00:17 → 00:20 |
Without codecs, serving video on the web would be almost impossible. |
| 00:21 → 00:25 |
Uncompressed, or raw video, simply requires too much data to transmit, |
| 00:25 → 00:27 |
over even the fastest networks. |
| 00:28 → 00:31 |
For example a 5 minute mobile video without compression, |
| 00:31 → 00:33 |
uses over 3 gigabytes of data. |
| 00:34 → 00:38 |
A 5 minute HD video uses over 25 gigabytes. |
| 00:38 → 00:42 |
Files this large would take many hours to download to your computer for playback. |
| 00:43 → 00:45 |
On a mobile connection, they'd take weeks. |
| 00:46 → 00:48 |
So, how does a codec shrink all that data, |
| 00:48 → 00:50 |
to something that you can play on your device in seconds? |
| 00:51 → 00:54 |
Codecs use sophisticated algorithms to work their magic. |
| 00:55 → 00:58 |
But you don't need to know what a discreet cosign transform, |
| 00:58 → 01:01 |
or entropy encoder is, to understand the basics. |
| 01:01 → 01:03 |
The first thing to keep in mind is |
| 01:03 → 01:05 |
that digital videos are sequences of pictures, called "frames". |
| 01:06 → 01:08 |
Much like the frames in conventional movie film. |
| 01:09 → 01:13 |
In most web video, there are about 30 frames per second of video. |
| 01:14 → 01:16 |
Because there are so many pictures per second of video, |
| 01:17 → 01:20 |
the picture in each frame varies only slightly from the frames before it, |
| 01:20 → 01:22 |
and those after it. |
| 01:22 → 01:25 |
Codecs work by removing similarities in the uncompressed frames, |
| 01:25 → 01:28 |
keeping only the differences, and encoding the differences as symbols. |
| 01:29 → 01:33 |
After all the redundancies are removed, the resulting data, called the "encoded bitstream", |
| 01:34 → 01:37 |
uses far less data than the original uncompressed source. |
| 01:39 → 01:42 |
There are 2 parts to a codec - an encoder and a decoder. |
| 01:44 → 01:48 |
The encoder compresses the video stream into a binary encoded bitstream. |
| 01:49 → 01:51 |
The decoder restores the encoded bitstream |
| 01:51 → 01:54 |
back into sequential pictures, or frames, for playback. |
| 01:55 → 01:59 |
The encoding process results in some loss of the original video's quality. |
| 02:00 → 02:03 |
One test of a codec's quality, is measuring the size of that loss. |
| 02:05 → 02:09 |
On the decoding side, simplicity and speed are perhaps the most important qualities of a codec. |
| 02:10 → 02:14 |
After all, a decoder isn't much use if no one's computer is powerful enough to run it. |
| 02:15 → 02:18 |
So, that's a brief introduction to the world of video codecs. |
| 02:19 → 02:21 |
Next time you watch a video on the web, |
| 02:21 → 02:24 |
you'll know something about the amazing technology that makes it possible. |
| 02:24 → 02:27 |
♪ |
| 02:27 → 02:29 |
[Captions by Captionwire.com] |