Content of the database concerning to each sentence

The database contains spoken sentences (about 2000/person) of 5 male and 5 female speakers . The RF data are given in Hz for each speech sound as RF1, RF2, RF3 in the function of time (in 5 measuring points of the sound inside: 10%, 25%, 50%, 75%, 90%.

Data representation

The Rf data (Hz) are given in txt files as the example shows below (sounds: ju:li.....). The speech sounds are given in SAMPA codes, the position of the measurements in milliseconds (time).

  file label num time pos rf1 rf2 rf3
1 f_bm_1701 j 2 0.1067 10 263 2014 2566
2 f_bm_1701 j 2 0.1142 25 289 1998 2529
3 f_bm_1701 j 2 0.1266 50 332 1917 2492
4 f_bm_1701 j 2 0.1391 75 357 1831 2399
5 f_bm_1701 j 2 0.1466 90 366 1536 2323
6 f_bm_1701 u: 3 0.1642 10 360 1230 2241
7 f_bm_1701 u: 3 0.1831 25 328 938 2303
8 f_bm_1701 u: 3 0.2146 50 304 899 2362
9 f_bm_1701 u: 3 0.2461 75 327 1018 2404
10 f_bm_1701 u: 3 0.2650 90 361 1234 2371
11 f_bm_1701 l 4 0.2821 10 343 1425 2377
12 f_bm_1701 l 4 0.2888 25 300 1418 2368
13 f_bm_1701 l 4 0.3000 50 311 1475 2439
14 f_bm_1701 l 4 0.3112 75 301 1485 2498
15 f_bm_1701 l 4 0.3179 90 302 1479 2550
16 f_bm_1701 i 5 0.3292 10 396 1655 2548
17 f_bm_1701 i 5 0.3393 25 399 1774 2550
18 f_bm_1701 i 5 0.3562 50 388 1965 2541
19 f_bm_1701 i 5 0.3730 75 382 2107 2594
20 f_bm_1701 i 5 0.3832 90 383 2121 2617

Every speech sound is represented by 5 lines as to the measuring points.

RF data

The database contains 11 693 100 db RF data in two groups.
In group 1. (for vowels and for  sounds represented by v, j, l, m, n, ny letters in Hungarian) the data represent the measured resonance frequencies in Hz.
In group 2. (consonants represented by  b, p, d, t, g, k, gy, ty, h, f, z, sz, c, dz, zs, s, cs, dzs, r letters)  the data do not represent measured data, but only derived ones by linear calculations from the neghburing sounds. So these data are virtual (VRF) but they help to see the continuous articulation process during speaking. On the homepage only examples are shown.