Home > Storage > PowerScale (Isilon) > Industry Solutions and Verticals > Media and Entertainment > PowerScale OneFS: NFS over RDMA for Media > Video files and image sequences
Professional content creation workflows deal with both video files and image sequences. Video files and image sequences can be compressed or uncompressed. This compression is referred to as the video codec (compressor/decompressor). Video is compressed to save space or bandwidth and then decompressed while being read. Video codecs can either be lossy (where some of the image data is discarded), or lossless (where no data is lost). Examples of codecs are H.264 (a common consumer delivery codec), ProRes (Apple’s codec designed for video editing), and Sony XAVC (used by some Sony cameras to capture video). PIZ compression is a lossless codec used with EXR image sequences, as described below. Technical and artistic requirements drive choice of codec, compression, and file type.
Video files can contain multiple tracks of audio along with video and are stored in a “container” format such as .mov or .mxf. The file format is referred to as a “container” because the file extension does not define which codec the video (and audio) stored within uses. For instance, a .mov file could contain video that is encoded with ProRes, H.264, or some other codec.
Image sequences store each frame of video in a discrete file. Some common image sequence formats are DPX and OpenEXR (EXR). DPX is an uncompressed format. Reading back an uncompressed DPX image sequence puts less load on the workstation because the machine does not have to expend CPU cycles decompressing the video frames. However, this benefit comes at the expense of more storage space and bandwidth consumed while playing back these large files. EXR image sequences can be compressed or uncompressed. As mentioned above, PIZ compression is a lossless compression type that can optionally be used in EXR image sequences. PIZ compression saves storage space, because each frame is compressed, and all image data is retained on playback because the compression is lossless. Playing back PIZ compressed EXR image sequences puts a heavy load on the computer reading those image sequences because the system has to decompress each frame in real time.
Image sequence playback also tends to be challenging for storage performance. To address this challenge, PowerScale OneFS has specific file-format aware techniques for dealing with image sequences, as outlined in this white paper: Isilon Filename Based Prefetch.
So why use a video file or an image sequence? As with codec choice, there are technical and artistic reasons to use each type. For example, some visual effects and animation software function better with image sequences because the intraframe compression of a video file does not play nicely with those applications. However, image sequences, with their file-per-frame format, can be unwieldy to store and transfer, making video files a better choice in many circumstances.