Shooting and Digitizing Techniques
for Great-looking Indeo® Video Interactive

Introduction Intel's Indeo® video is software that allows you to capture, compress, decompress, and play digital video files on a desktop PC. It is available free of charge to PC users, video producers, and multimedia software developers, who can distribute it royalty-free with their applications. This document describes how to most effectively shoot, digitize, and compress video using the Intel Smart Video Recorder III and Indeo video interactive. This document assumes that you have installed on your PC: the Intel Smart Video Recorder III board and (Indeo video) drivers software such as Adobe Premiere* or Asymetrix Digital Video Producer* for video capture, edit, and playback The information is divided into four sections: Planning Describes considerations to bear in mind before you begin. Shooting the Video Provides hints for video quality, lighting, color, and camera techniques that can help your compressed video look better. Capture Techniques Discusses the best way to capture audio and video. Compression Techniques Describes how to use Indeo video interactive's software compression capabilities to create videos with the highest quality and lowest data rates for such applications as CD-ROM publishing. For technical assistance with Indeo video, contact Indeo video technical support. For technical assistance with the Intel Smart Video Recorder III, contact ISVR III technical support. Planning To save time and use Indeo video as effectively as possible, consider how the resulting video is to be used in the final product before you begin. The final image window size and the maximum playback rate affect how you compress a video with Indeo technology. In addition, proper management of your computer's hard disk and video capture files will help avoid dropping frames during capture. Image window size Larger windows provide greater visual impact but typically result in larger Indeo video files. They also usually require faster PCs to play back video at higher frame rates. The following image window sizes, in pixels, are commonly used: Window size Description 320 x 240 Considered ideal for many multimedia applications. 240 x 180 A good intermediate window size. 160 x 120 The smallest commonly used video window size. Maximum playback rate Below are typical playback rates. Higher rates (more frames per second (fps)) allow for smoother video playback. Playback rate Description 5-10 fps Used in multimedia, such as video conferencing or the Internet, that requires low data rates. 15 fps Considered satisfactory for most multimedia applications. 24 fps The frame rate at which motion pictures are projected in theaters. 25 fps PAL (Europe) broadcast television frame rate. 30 fps NTSC (USA, Japan) broadcast television frame rate. Shooting the video Use the following video quality tips and techniques for lighting, use of color, and camera movement to ensure that your footage is suitable for creating high-quality compressed digital video, avoiding situations that make compression more difficult. Video source quality The most important factor affecting the quality of your final Indeo video file is the quality of the original source video. Compression cannot recreate detail that was never present in the original image, but it can take artifacts present in source video and potentially make them worse. Therefore, it is important that the best video source possible be used to create compressed video files. Compression algorithms analyze the digitized video input stream searching for redundant or predictable data patterns, which can be compressed and reconstructed later. The compressor interprets noise or artifacts digitized from the analog source as non-redundant, unpredictable data and therefore wastes valuable CPU time and file space attempting to accurately compress and store these deficiencies. That's why the quality of the original source video tape is so critical to the quality of the resulting compressed digital video file. Video artifacts can in fact look worse after compression than they did originally. Video compression has created a new variation on an old computer saying: "garbage in, garbage squared out." Capture your source video using a high-quality video tape format. The best source video formats use S-Video, rather than composite. S-Video signals carry separate signals for luminance (brightness) and chrominance (color), resulting in higher bandwidth and an improved signal-to-noise ratio. Composite video sources modulate the luminance and chrominance together on one signal, which lowers both the bandwidth and the signal-to-noise ratio. Composite video signals are also subject to video artifacts such as color bleeding. From highest to lowest quality, component video formats include: D1, Digital Betacam Betacam SP Hi8, Super-VHS D1 and Digital Betacam actually store video information digitally, rather like an audio CD. Both formats are broadcast quality but expensive, even for a professional budget. Betacam SP is the most widely used analog recording format for video creation and post-production; its high quality and ubiquity make it an ideal choice for professional and semiprofessional multimedia authoring. Hi8 and S-VHS are consumer formats, and therefore inexpensive and easily available. Both produce good results for desktop and consumer use. From highest to lowest quality, composite video formats include: D2 1" Type C, 3/4" U-Matic videodisc VHS D2 is the digital composite cousin of D1 and is similar to D1 in both quality and expense. 1" and 3/4" literally refer to the width of the recording tape used in each format: 1" is an open reel-to-reel format and 3/4" is a cartridge format. Both are older professional studio formats; while both produce better results than consumer formats, they have been almost entirely supplanted by the newer and better Betacam SP format. Despite being a composite format, videodisc (also known as laserdisc) nevertheless provides good quality. Some videodisc players provide S-Video outputs, but the original signal is still composite, and low-cost electronics can sometimes make these S-Video outputs worse than the composite. Under the circumstances, it's probably better to use the composite output, but try both to determine the specific characteristics of your videodisc equipment. VHS is the familiar home consumer format. Its composite signal and low-cost recording and playback mechanisms make it the lowest in quality of any available format. Lighting Use enough light. Adequate lighting is vital for creating high-quality video. Most consumer video cameras generate noise in low light. Random noise in an image decreases the redundant information, leading to poor compressed video quality and higher data rates. Use natural light when possible. The best light occurs outside on a cloudy or overcast day, because the light is evenly diffused. Direct sunlight can create areas of deep shadow where video noise can be prominent. Use a reflector to bounce light onto your subject. Reflectors produce more even lighting and can also reduce areas of deep shadow. Avoid fluorescent light. The color temperature of fluorescent light makes videos look blue-green. If you're in an office, shoot near windows to use the natural light. Place the camera between the windows and the subject; don't use the windows as a background, which would place your subject in relative shadow. Color Avoid saturated colors in the scene. Video with highly saturated colors (especially red) can bleed or appear blocky, particularly when viewed on 8-bit (256-color) displays. Avoid adjacent areas of high contrast. For example, a white shirt with black stripes is a poor choice. Sudden changes in brightness levels can emphasize color bleeding and create edges that are more difficult to compress. Avoid extremely thin horizontal or vertical lines. Areas of fine detail are not always particularly visible after compression, especially when displayed in a multimedia application window. Moreover, extremely fine lines or patterns (such as those that are one pixel wide) can distort the source video image. For example, a certain late-night talk-show host was notorious for wearing seersucker suits and garish ties that caused these problems on television displays. Such distortion looks bad and is extremely difficult to compress successfully, resulting in poor video quality. Camera Techniques Make the camera invisible. Your viewer should notice the subject, not the camera. Limit camera movement. Use a tripod. Even the best hand-held video camera moves, and this movement reminds the viewer that someone is there, holding a camera. Also, because a camera on a tripod is steadier, more information is redundant between frames, yielding better compression. Don't overuse pans and zooms. These camera techniques put the focus on the camera instead of the subject. Pans and zooms also limit how much can be compressed because nearly every pixel can change between frames. Rather than zooming, break the action up into two static shots: one close-up and one long shot. Use other techniques to create visual excitement. For example, instead of centering your subject in the frame, create a more dynamic shot by placing the subject to one side of center and directing her attention to the other side of the frame. Use close-ups. Remember that your video will probably be played within a relatively small window on a computer monitor. Close-ups work well to emphasize your visual message in this medium. Use auto-focus judiciously. Auto-focus works well for most medium and wide shots. Don't use it, however, in close-ups where the subject is moving, or when zooming in on a subject. The entire image gets first blurry and then sharp as the camera attempts to stay focused on the object in the center of the frame. This is not only distracting, it also limits the effectiveness of compression, because every pixel changes as the camera refocuses. Instead, before you shoot, lock your camera on non-auto focus, zoom in tight on the subject, then manually focus and zoom back out to frame the shot. Capture Techniques Follow these tips to optimize your video quality and performance when using the Intel Smart Video Recorder III. Use high quality video playback equipment as your capture video source. Consumer camcorders provide a good signal when shooting live, but some models do not play back as well from recorded video tape. When digitizing from video tape, therefore, it is better use a high-quality editing deck rather than the camcorder itself. Choose S-Video. If your video source offers a choice between composite and S-Video formats, choose S-Video. Connect the S-Video output of your video source to the S-Video input on the Intel Smart Video Recorder III to ensure the highest possible video capture quality. A digital time base corrector is a useful piece of studio hardware. A time base corrector between the source video deck and the video digitizing board can correct certain signal deficiencies. These deficiencies are sometimes created by lower-end video equipment, such as weak or inaccurate synchronization signals from the source video deck, particularly common in VHS and other consumer-grade equipment. Inadequate synchronization signals appear as tearing along the top or bottom edge of the video capture window. Many higher quality video decks, including even some of the better S-VHS and Hi8 equipment, feature built-in digital time base correction. Setting up the capture In your video capture program, experiment with the input Video Source settings. If you're unhappy with the appearance of your captured video, you can adjust the incoming signal using the Brightness, Contrast, Saturation, and Tint controls in the Video Source menu. For most video captures, the default slider settings are correct, and any noticeable problems with the appearance of the captured signal are probably due to the analog video signal itself and not the capture board. But you can use the Video Source controls to enhance the incoming signal. If the analog source video is monochrome (black and white), or if you'd like to make it so, turn the Tint and Saturation sliders in the Video Source menu down to zero. Turning these controls off ensures no color data is recorded into the video file. This results in better-looking monochrome video and better compression; it also prevents any imperfections in the source signal from being recorded into the video as erroneous color data. If the preview of the video appears too slowly, set your Windows* display to 8-bit mode. Some programs allow you to see the incoming video signal on your computer monitor. When capturing, the preview may appear slowly or even freeze; this is normal and does not mean that frames are being dropped. To maximize the speed with which the preview appears, set your Windows display mode to 8-bit (256 colors). Your Windows display mode has no effect on the capture quality of Indeo video, which is always captured and stored in full-color format. Compression Techniques It's important to be aware of the effect capturing and editing video files has on your hard disk. Capturing video files can quickly consume hard disk space, and repeatedly creating and deleting files can cause disk fragmentation. Use the following techniques to maximize the performance of your hard disk for capture. Periodically defragment your hard disk, especially before capturing video. Contiguous free disk space helps avoid dropped frames. Avoid allocating disk space in your capture files that doesn't get used. There are two ways you can create files into which you capture video: Create a new capture file each time you want to capture a video clip. Create a single, large capture file, always capture your video clips into it, and then save each clip using a different name. Create a capture file in your capture program using its Set Capture File option. Your capture program may automatically pre-allocate 20% of the available hard disk space to the capture file. Suppose, for example, that 20% of your available hard disk space was 60MB. The capture program would allocate a file of this size. Then suppose that you capture only 5MB of video into the file. The resulting file will still be 60MB, wasting 55MB of your hard disk. To avoid this problem, override the default capture file size and set it to 1MB. If you capture more than 1MB, the capture program should continue to write your video to the hard disk until you stop capturing. At the end of the capture, your file will be exactly as large as it needed to be. The other way to capture video, however, is to always capture your video clips into the same capture file, but then use the capture program to save each clip you capture under a different name. Each file you save from the capture file will be exactly the size of the video you last captured. This technique is convenient because you almost always want to perform at least simple editing on each clip you capture, such as trimming a few frames or interleaving the audio and video. To try this technique, use your capture program's Set Capture File option to create and pre-allocate space for a large capture file, large enough to hold the longest clip you think you'll capture. (You may want to defragment your computer's hard disk before you create the capture file.) Then for each video clip you capture: Capture a video clip into the capture file. Open the capture file in your capture program. Edit the clip as necessary (for example, trim the first few and last few frames). Assuming that you do not wish to compress or recompress the video at this point, but simply to save it to your hard disk, make the output and input compressor settings the same (for example, Indeo video interactive). If the option is available in your capture program, set the audio/video interleave to 1. Save the edited clip using a different name. Repeat this procedure for each video clip you capture. Repeatedly capturing into the same capture file, trimming excess frames, and resaving each new clip to disk reduces both disk space consumption and fragmentation. Tips to avoid dropping frames If possible, you don't want to drop any frames during capture. Video files with dropped frames appear to stutter during playback because some of the original video frames are missing from the file. If frames are dropped during capture, try some of the following techniques: Add more memory to your computer. When using the Intel Smart Video Recorder III, your computer should have a minimum of 16MB (32MB is recommended), especially when capturing uncompressed video. Use a faster Pentium processor. The Intel Smart Video Recorder III real-time compression software runs on your computer's Pentium® processor, not on any special video hardware. A faster Pentium processor can compress video at higher frame rates, bigger window sizes, and with better video quality and compression. And PCs based on newer Pentium processors are more capable of keeping up with the amount of video data being captured over the PCI bus. This is especially important if you're capturing RAW video. Use a bigger, faster hard disk. Use a fast, 1GB or larger hard disk, one with less than 10 ms access time and 1MB per second sustained data rate. Newer SCSI drives (especially SCSI-2/Fast/Wide) are ideal for video because of their speed. You can also link SCSI drives together if you need more disk space. Defragment your hard disk regularly. Reduce the amount of data you're trying to capture. If your video capture tool supports it, capture using a lower frame rate, smaller window size, or lower Quality setting. Also, use lower audio quality settings to reduce the size of the audio being captured. Don't use disk compression programs. Compressed hard drives are slower and affect capture performance. Audio capture tips Don't capture audio if it isn't there. If you capture audio that isn't there, an audio stream is created in the .AVI file, wasting space. Properly calibrate your audio levels. Use the software that came with your sound card to properly calibrate the input audio levels on all of the audio sources to be used in your project. If your audio source has a volume control, you should always try to adjust it as loud as possible without causing distortion. If the volume coming from your sound source is too low, you'll have to turn the "Line Input" control on your sound card's mixer up too high, which adds noise to your audio. See your sound card documentation for more information. Use the lowest audio quality settings suitable for your project. Audio is typically not compressed and can therefore occupy a lot of disk space. Increasing any of the three audio quality parameters (sample size, sample rate, or number of channels) doubles the audio data rate. The following table shows examples (using the Intel Smart Video Recorder III) of how increasing the audio quality can affect the size of the file. Audio Quality KB per Second 8-bit, 11 KHz, mono 11 8-bit, 22 KHz, mono 22 8-bit, 22 KHz, stereo 44 16-bit, 22 KHz, mono 44 16-bit, 22 KHz, stereo 88 16-bit, 44 KHz, stereo 176 Interleave the audio and video to ensure synchronized playback. Audio and video are not interleaved during capture. Even if your captured video doesn't require video editing, you must edit each captured file and properly interleave the audio and video streams. If your capture program provides the feature, set the interleave audio parameter to 1.

* Legal Stuff © 1997 Intel Corporation