REAL-TIME PROCESSING

DC Live/Forensics has been optimized to "run" in real-time. When the operator/examiner changes a filter setting, the results of this change are heard immediately. Real-time processing enables the operator/examiner to select the optimum filter settings by adjusting the controls while listening.

Each of the filter modules has a "default" set-up and several set-ups that are optimized to handle common problems. Using the default or one of the custom settings as a starting point, the operator listens to the processed audio output and decides if the processing is satisfactory. If additional enhancement is needed, the operator/examiner adjusts each filter control in the direction that improves the intelligibility of the spoken words and reduces interfering noise until the intelligibility of the spoken words begins to get worst and/or the interfering noise(s) gets louder. The operator/examiner then moves the control back to the best position.

Real-time processing requires that the filtering programs be written by the programmer to run efficiently and that the computer performing the processing have a high clock speed, a fast front side bus, a large RAM memory, and a fast, large hard drive.

The user cannot change the efficiency of the software programs but can use a faster computer.

DC Live/Forensics is optimized to run in real time. If the user's computer is not fast enough to process in Real-time, two options are possible:

1. Processing to a destination file on the hard disk in non-real-time using the custom multi-filter you have designed, previewed, and optimized to reduce the interfering noise. Draw Back: the operator cannot monitor (listen to) the audio as it is processed to verify satisfactory performance of the filtering operation.
2. Process Sequentially by setting up two or more multi-filter chains such that each chain can run in Real-time to attain the desired enhancement. Draw Back: requires two or more times the playing time of the recorded information. Produces rounding errors when the processed sound files are converted from processing precision (32-bit floating point) back to WAVE file precision (8 to 24 bit fixed).