The sets are "Bell" (self explanatory), "Tri" (a triangle), "Egypt", and "Man" (music pieces).
In each set there are four files. The first (e.g. bell.wav) is the original uncompressed sound. The _q file (e.g. bell_q.wav) is the result of compressing the first file with Dolby Digital, the decoding and creating the new wave file from the decoded output. If the compression system is working well there should be little or no audible difference between the first two files of each set. The third file _sb (e.g. Bell_sb.wav) is the result of subtracting the second file from the first. It is all the differences between the original sound and the output of the encode/decode system. This file contains the increased quantization noise, and all the elements that were "masked out" in the encoding process (see Chapter 17). One problem in appreciating this is that this third file is low-level, because the compression system is very good, so the differences are small! So, it may be difficult to hear all the components of the difference without a high-gain amplifier. To simplify the process I have taken each of the _sb files and made a fourth file _sbNorm (e.g. Bell_sbNorm) normalized to raise the peak level in the difference signal to the nominal 0 dB level. This make the differences much easier to hear but remember, in fairness to Dolby, that this is very artificial. The sound levels you will hear from these normalized files do NOT represent the differences due to the compression system!
Back to Top Level Contents
Back to Introduction