7. Encoding for Delivery
Digital video comes in many resolutions and frame rates; Digital Betacam, MP4 video phones, mini-DV, HDV, DVD and so on. One universal measure of these different formats is Bit Rate (also known as ‘bitrate’). Bitrate is the number of bits (units of data) processed per unit of time. When dealing with Video or Audio signals, bitrate refers to the number of bits used per second (after coding or compression) to represent the images and/or sound over a continuous period and corresponds to the signal’s ‘bandwidth’.
A simple analogy is to think of a garden hose. Roughly speaking, a hose’s bandwidth (the width of the hose) is restricted because perhaps one can only get half a pint per second through it. Use a drainpipe with its greater bandwidth, and many more pints can pass per second. The greater the bitrate, the more information or detail can be carried in the signal. Therefore, a video stream of 10Mbps (Megabits per second), will have the potential of containing ten-times greater resolution or information than a video stream of 1Mbps.
Bytes are the computer industry’s measuring units for data, and each byte contains eight bits. All computers deal with bytes, but pipes running between computers deal in bits (as, strictly speaking, not all bytes are equal). Any bitrate can be divided by 8 to find the equivalent in bytes. Understanding bitrates and bytes becomes important when converting video into a format acceptable for DVDs or the web.
As general information, the following list shows the bitrate of common forms of digital video.
- Mini-DV/HDV 1080i = 25Mbps
- DVD = 7.5Mbps (ranges between 2 and 9 are possible, with 3-7 being normal)
- HD-DVD and BluRay = 9-18 Mbps
- Freeview = 2MbpsThe following notes contain basic information for creating video for DVD and website delivery. As each encoding system or solution works differently, you will need to experiment in order to create a compatible and playable video programme.
7.1 Video for DVDs
Most single-layer recordable DVDs are limited to about 4.7 Gigabytes of data, but an hour’s worth of HDV in its native format occupies 14 Gigabytes.
To convert the material for use on the DVD, we must convert and compress it into a format called ‘MPEG2’. However, although a short movie could use a high MPEG2 bitrate to get the best quality within the space available, the DVD player may not be able to read the data fast enough. The maximum safe bitrate for MPEG2 on a DVD is around 7.5 megabits per second. Set at this bitrate you will be able to fit around 80-minutes on to a single-layer DVD. This is based on the following calculation:
- 4.5 Gigabytes x 8 = 36 Gigabits (or 36,000 Megabits)
- 36,000 Megabits @ 7.5 megabits per second = 4,800 seconds (or 80-minutes)You will also need to account for menus and additional data stored on the DVD. It is also not good practice to use the entire capacity as data stored on the outer edge of the DVD is more susceptible to corruption and playback errors. By reducing the bitrate of the MPEG2 stream you will be able to fit a longer duration programme on the DVD. However, reducing the bitrate will lower the resolution of the video. DVDs of feature films are generally longer than 80-minutes and they also normally include additional information such as special features, trailers and sub- title information. These films will often be compressed using bitrates of less than 5Mbps – without an appreciable loss in viewing quality. There are two methods of setting the bitrate on MPEG2 – ‘Constant Bit Rate’ (CBR) and ‘Variable Bit Rate’ (VBR). With CBR all video and audio will be encoded at the same bitrate. With VBR the video will be encoded at a bitrate determined by the amount of information or detail in the video. Segments with a lot of detail or movement will be encoded at a higher bitrate than those with less detail or movement. The upper bitrate limit can be set to ensure that it does not exceed playback restrictions or the capacity of the DVD.The benefits of CBR encoding are that it is generally quicker to encode, and is generally more compatible with DVD players. The benefit of VBR is that it is the most efficient in terms of data capacity.
7.2 Video for the Web
There are two main ways that video can be distributed over the web. These are ‘Streaming’ and ‘Download’. With streaming video you can select any point within a video and it will start to play from that point onward. Streaming technology is also used for taking a live video feed and uploading it to a file server so that it is available for an audience to watch there and then. Download refers to video content which is uploaded to a server and incrementally downloaded by the end user. Download web video can be as simple as a file which is copied from the host website by the end user and played once fully downloaded, or delivered by means of progressive download so that the video starts to play once a given proportion of the video has been temporarily transferred to the end user and played once fully downloaded, or delivered by means of progressive download so that the video starts to play once a given proportion of the video has been temporarily transferred to the end user’s computer.
Bitrates play an important role in preparing video for the web. Many compression programs will suggest helpful settings for certain applications, but the general measurement for all formats is bitrate. There are various bitrate bottlenecks that will impact on the delivery of video over the web. These include the internet connection speed of the end user, the speed at which the file server where the video file is stored can deliver the content, and the number of end users wanting to view the same video at any one time. These will all be affected by the quality or data rate at which the video has been encoded. The higher the quality, the greater the bandwidth required at each bottleneck.
The average 2Mbps ADSL internet connection is shared between 25-100 users, so it’s unlikely to provide more than 1Mbps at any given time. Whilst it can deliver near 2 Mbps in the middle of the night, it can dip far lower at peak times. So, video on the web tends to be measured in Kilobits rather than Megabits.
In most cases you will be unable to determine what internet connection speed your audience has available to them so you will need to encode your video according to the lowest possible connection speed. For the average broadband end user this could be as low as 300 to 500kbps. However, there are encoding solutions and server facilities that enable the video to be encoded at differing bitrates, and then served to the viewer according to their detected connection speed.
The main consideration is that that squeezing, say, DV material at 25,000kbps down to 500kbps requires a lot of compression. The resulting material will not be of the same resolution, may not include the same number of frames per second, and may contain audio of a much lower quality than the original material.
Whilst internet speeds are increasing, and encoding and server technologies improving, delivering video over the internet might always result in a trade-off in quality.
7.3 Better Web Video
There are four simple measures that the videographer can employ to lessen bandwidth requirements and deliver more effective web video.
- Smaller frame sizes require less bandwidth. When designing video content for the web, shoot material that can be seen at a smaller frame size. Avoid situations that will cause the end user to have to play the video at full-screen in order to take in the detail.
- Less movement and detail in the frame. The way compression and playback codecs work is that they only need to process changes from one frame to the next. Avoid using unnecessary pans, zooms and other camera movements. Also, avoid scenes with lots of moving content. If shooting a talking head shot, ensure that the background is static as this will require less encoding and decoding.
- Less noise in the picture. All detail in a picture takes up valuable bandwidth and this includes inherent noise in the picture. Avoid using artificial gain by lighting your shots properly. And do not treat the picture in post production in any way which causes additional detail, noise or movement.
- Quality In = Quality Out. Use the highest original production format as possible and endeavour to maintain that quality during post production.
- Finally, always use good quality encoding tools.