Skip to main content
. 2012 Mar 29;7(3):e33993. doi: 10.1371/journal.pone.0033993

Table 2. Power-law fitting results for Bark-band code-words per database and window size.

DB/Window N words Inline graphic Inline graphic Inline graphic
Speech
46 ms 494,926 2,000 2.20Inline graphic.05 0.84Inline graphic.04
186 ms 219,595 501 2.22Inline graphic.05 0.82Inline graphic.03
1,000 ms 100,273 79 2.33Inline graphic.05 0.75Inline graphic.03
Music-W
46 ms 1,724,245 2,000 2.26Inline graphic.04 0.79Inline graphic.03
186 ms 798,871 794 2.33Inline graphic.06 0.75Inline graphic.03
1,000 ms 240,236 79 2.29Inline graphic.03 0.78Inline graphic.02
Music-nW
46 ms 1,905,444 126 2.17Inline graphic.01 0.85Inline graphic.01
186 ms 947,327 50 2.17Inline graphic.01 0.85Inline graphic.01
1,000 ms 306,682 5 2.17Inline graphic.01 0.86Inline graphic.01
Elements
46 ms 125,248 794 1.95Inline graphic.04 1.05Inline graphic.05
186 ms 34,171 20 1.79Inline graphic.02 1.27Inline graphic.03
1,000 ms 10,231 8 1.79Inline graphic.02 1.27Inline graphic.03

DB/Window means database name and window size, N words is the number of used code-words, Inline graphic is the minimum frequency for which the Zipf's law is valid, Inline graphic is the frequency-distribution exponent (Eq. 2), and Inline graphic corresponds to the Zipf's exponent (Eq. 1).