Part 1. Optimizing Input

Audio and Video Sources

It is important to start with the best-quality source. This section lists possible sources:

Serial digital interface (SDI) video. Used for digital video cameras and camcorders. Because the content stays in a digital format throughout the capturing and encoding processes, this results in the least amount of data translations, and results in the best-quality video.

Component video. Used when sourcing from DVDs. With this source, the video signals are separated, for example, into the RGB or Y/R-y/B-y format. Results in good-quality video.

S-Video. Used for S-VHS, DVD, or Hi-8 camcorders. The video signal is divided into luminance and chrominance. Results in good-quality video.

DV video. Used with DV devices, such as MiniDV digital camcorders connected through an IEEE 1394 video port. Results in good-quality video.

Composite video. Used for analog cameras, camcorders, cable TV, and VCRs. Composite video should only be used as a source as a last resort. With composite video, luminance and chrominance components are mixed, which makes it difficult to get good-quality video.

Audio. If possible, capture digital audio. If you must capture audio from an analog source, balanced audio connections are better than RCA.

Optimizing Your Computer

Before you start capturing, optimize your computer using the following steps:

Defragment your hard disk.

Turn off network and file sharing.

Close all other programs, especially if a program accesses the hard disk.

Monitor system resources, making sure that the computer is sufficiently powerful to keep pace with the data feed.

Watch for direct memory access (DMA) buffer conflicts between the capture card and the SCSI card, which can result in frame dropping. This is less likely to occur now than in the past. If conflicts occur, one solution is to use a dual PCI bus motherboard configuration, in which the capture card and the SCSI card are on different buses.

Part 2. - Compression 

In addition to bitrate and framerate there are several settings you can adjust for optimal quality. In many cases the preset values will be adequate, however it is important to know how and when to change these settings.

• Key frame. A key frame is a point in encoded video where the data for the entire frame is transmitted, rather than just the changes. Key frames are generally inserted when there is a scene change. The key frame setting is the minimum time between points where the encoder will insert key frames (they may be inserted more often automatically, if necessary). Decreasing the distance between key frames can improve the quality of the video. If you use a longer key frame distance, additional key frames are inserted when necessary; for example, when a scene changes. Keep in mind, though, that a long key frame distance will affect the amount of time a user may need to wait for video in a broadcast scenario.

• Buffer size. The bit rate and quality of content fluctuate within the confines of the buffer size. A larger buffer size enables more bits within the buffer range to be allocated to complex scenes. For example, if you set the buffer size to 10 seconds, the codec may choose to allocate x number of bytes to the first 8 seconds, and the rest during the last 2 seconds. This allows for the more complex parts of the video to have more bits allocated within the buffer. Typically, increasing the buffer will improve overall quality. However, it also increases the delay between the time when the user requests the content, and playing starts. For lower bit rates, it is recommended to increase the buffer size. For higher bit rates, increasing the buffer size has a smaller impact on quality.

• Video smoothness. Video smoothness determines the tradeoff between sharp images and smooth motion. Video appears smooth when objects move easily from one position to another on the screen, and the edges of objects are not jagged. Video appears clear when images and motion are well-defined and clearly delineated. The bit rate setting determines the number of bits that can be allocated over a period of video. Based on this setting, the codec can choose to include more frames, which results in the images appearing smoother. However, each frame uses fewer bits. With a higher video smoothness value, the codec may include fewer frames. This increases the number of bits allocated per frame, resulting in sharper images; however, the image may not appear as smooth. The video smoothness setting only comes into effect when there are not enough bits to encode at the specified frame rate, and a tradeoff must be made. At higher bit rates, this value can be increased. If you are dropping frames during encoding, consider decreasing video smoothness.

Part 3. Shooting video for streaming

The mechanics of streaming

When streaming video is encoded it will use the available amount of information (defined by the chosen bitrate) to create the highest quality video possible. Unlike film, each ‘frame’ in a streaming video is not a complete image, it is simply an update of the image to show any changes that were made between frames (generally movement). The only complete updates are what is known as ‘key frames’, which are inserted at regular intervals and also for scene changes.

In a video with very little movement, only a small portion of the available data for each frame needs to go to these ‘updates’, the rest is available to enhance the picture’s quality. In a video with frequent movement, the available data needs to be used to update large portions of the image, resulting in less data available for image quality.

If it is possible to minimize the amount of change between frames, you can deliver a high quality stream at a lower bitrate, or increase the quality of your stream without increasing the bitrate.

If you are shooting the video you intend to stream, you should attempt to minimize unnecessary movement in your video. Try to focus on your subject and keep the background either simple, or softly out of focus. If you are not producing your own content you will have to rely on compression techniques to optimize the video for streaming.