Since dailies have to be created at a reliable and predictable pace, Datalab has to operate at a reliable speed. The editing process has to be supported by consistent availability of information and material.
There are considerable performance requirements in terms of CPU, disk space and bandwidth. Datalab meets these needs by using commodity hardware and by managing workloads intelligently. In the digital environment, quality need not be affected by the price of the components used. This is a significant difference to the analog world where you usually get what you pay for. Datalab uses this concept extensively and shares the generated value in ways of speed and flexibility with the film maker.
The images created for Zodiac left the Thomson Viper camera by means of a realtime dual HD SDI connection. The is the only time in datalab that a realtime technology was used. (except for the various playback scenarios of course.) Datalab tries to avoid realtime operations. They generally involve expensive equipment whose performance is locked. Realtime operations are not as easily scriptable and are rarely open source. Integrating them into a bigger automated workflow can be counterintuitive. Many operations in the processing domain have already crossed into faster than realtime, upending the price-performance ratio of dedicated real-time equipment. ALso, processing operations scale much more efficiently than real time operations.
