Getting Started

[shawn.p...@gmail.com]

Hello,

I'm trying to get started using AuToBI and I am having some trouble. I'm trying to run everything on a 64-bit Windows 7 machine. Here is the process I have followed so far:

1) Installed Java Version 1.7.0.15
2) Downloaded AuToBI version 1.3 compiled jar file from http://eniac.cs.qc.cuny.edu/andrew/autobi/version_1.3/AuToBI.jar
3) Downloaded the trained model version 1.3 from http://eniac.cs.qc.cuny.edu/andrew/autobi/models/version_1.3/burnc_bdc_games.AuToBI_1.3.10022012.tgz
4) Downloaded some sample Praat files and textgrids from http://aune.lpl.univ-aix.fr/~welby/PRAAT/Praat-Welby-9Feb2006.zip
5) Put the following files in the same directory (C:\AuToBI):

AuToBI.jar
longfile.wav (obtained from zip file in step 4, located in ~\Praat-Welby-9Feb2006\SOUNDFILES\SET3)
longfile.TextGrid (obtained from zip file in step 4, located in ~\Praat-Welby-9Feb2006\TEXTGRIDS\SET3, TextGrid has 0 point tiers and 1 interval tier named words with word annotations)
burnc_bdc_games.AuToBI_1.3.10022012

6) Executed the following command:

java -jar AuToBI.jar -input_file="longfile.TextGrid" -wav_file="longfile.wav" -pitch_accent_detector="burnc_bdc_games.AuToBI_1.3.10022012" -pitch_accent_classifier="burnc_bdc_games.AuToBI_1.3.10022012" -intonational_phrase_boundary_detector="burnc_bdc_games.AuToBI_1.3.10022012" -intermediate_phrase_boundary_detector="burnc_bdc_games.AuToBI_1.3.10022012" -phrase_accent_classifier="burnc_bdc_games.AuToBI_1.3.10022012" -boundary_tone_classifier="burnc_bdc_games.AuToBI_1.3.10022012" -out_file="longfile.TextGrid2"

When I open the longfile.TextGrid2 output file, there are 3 tiers:
words (modified from original TextGrid, silence intervals have been removed)
pitch_accent_hypothesis (same number of tiers + aligned with words, but blank)
phrase_hypothesis (same number of tiers + aligned with words, but blank)

Am I doing something wrong here since the 2 new tiers are blank?

[shawn.p...@gmail.com]

One other note - I ran the longfile.wav file through SoX (command: sox.exe "~\AuToBI\longfile.wav" -t .wav -r 16000 "~\AuToBI\longfile_16.wav") to make sure it was 16 kHz, and then ran AuToBI again, and I had the same results. I'm thinking the blank tiers are not due to a .wav file format issue.

[Andrew Rosenberg]

Hi Shawn,

To get this up and running, you'll want to uncompress the trained models in the tgz file.  Then specify the appropriate model with each flag.  Something like the following:

-pitch_accent_detector=burnc_bdc_games.pitch_accent_detector.model

-pitch_accent_classifier=burnc_bdc_games.pitch_accent_classifier.model

-intonational_phrase_boundary_detector=burnc_bdc_games.intonational_phrase_detector.model

-intermediate_phrase_boundary_detector=burnc_bdc_games.intermediate_phrase_detector.model

-phrase_accent_classifier=burnc_bdc_games.phrase_accent_classifier.model

-boundary_tone_classifier=burnc_bdc_games.phrase_accent_boundary_tone_classifier.model

This bug mentions an potentially useful (and simplifying) to include shortly.  There's no reason that AuToBI shouldn't be able to decompress and load the models from a single compressed archive file.  

The current approach asks a user to specify each model separately, so he or she can choose which of them to use, while ignoring others. But again, this can be done more simply.

Thank you for your interest in AuToBI, and don't hesitate to get in touch if you have more questions or problems.

-Andrew

--
You received this message because you are subscribed to the Google Groups "AuToBI Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to autobi-users...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[shawn.p...@gmail.com]

Yup, that got it working. Not sure what happened on the original unzip that only produced 1 file, but once I had the 6 extracted files everything was good to go.