1 2CMU ARCTIC KSP 3 4This directory contains a recording of the phonetically balanced US 5English CMU ARCTIC database by KSP, a male Indian English speaker. 6 7See http://www.festvox.org/cmu_arctic/ for details on the 8database coverage and other recordings of this dataset 9 10The format follows the Festvox (http://www.festvox.org) directory 11structure. 12 13The directory structure is 14 bin/ 15 basic scripts for building prompts, labelling feature files etc.s 16 cep/ 17 Ceptrum files dynamically created in phone autolabellingl 18 dic/ 19 Final diphone dictionary final (used at run-time) 20 etc/ 21 prompt file, and some labelling templates 22 festival/ 23 Not used in diphone bases 24 festvox/ 25 scheme voice definition files (used at run-time) 26 group/ 27 extracted diphones into signle group file for distribution 28 lab/ 29 autolabelled phone labels (generated by EHMM) 30 lar/ 31 recorded EGG signal files (not used in this example) 32 lpc/ 33 LPC parameters plus residuals, (used at run-time for nongrouped version) 34 mcep/ 35 MFCC (Mel Frequency Cepstrum Coefficients) not used in diphone databases 36 pm/ 37 Pitchmark files as extract from waveforms (or EGG signal) 38 pm_lab/ 39 derived pitchmark labeled files from pm/ enabling emulabel (and others 40 display programs) to show the pitchmarks and waveform files. 41 prompt-cep/ 42 cepstrum files for 43 prompt-lab/ 44 label files for synthesized prompts 45 prompt-wav/ 46 waveforms of synthesized prompts 47 prompt-utts/ 48 utterances of synthesized prompts 49 wav/ 50 recorded spoken nonsense words (in Microsoft riff (wav) format). 51 If you are using Xwaves you should convert these to NIST format 52 53 54 55 56 57 58