search

saml / sphinx4-forced-alignment

sphinx4 forced alignment demo
5 stars 4 forks source link

Weird output - contains only the final two words #1

Open prasvin opened 13 years ago

prasvin commented 13 years ago

Hey there, I followed as specified in the readme and got it up and running fairly easily. However, I am getting somewhat weird output.

This is the command executed from command line : java -Xmx2048m -Dconfig=config/aligner.xml -Daudio=output.wav -Dtext=transcript.txt -jar ForcedAlignment.jar

I had to place '-Xmx2048m' to avoid 'java.lang.OutOfMemoryError: Java heap space' error.

I have placed output.wav at http://www.mediafire.com/?sclm6s9dvhrfwwk if anyone wanna check out. Its actually from youtube http://www.youtube.com/watch?v=xCw8oQyDf5g and converted to .wav

The following are the contents of transcript.txt (without quotes) : "Message pretend how are you doing. Note buy eggs. Play music."

And the output I obtain is : play(4.98,5.55) music(7.73,10.26) music(61.64,70.22) music(93.67,95.69) music(107.22,107.49)

The output contains only the last two words & the timestamp is way off (total duration of input is 1 minute 05.34 seconds. I did not find any place to post this, so I had to here. I would appreciate if you would shed some light on this. Thanks.

abhayathapa commented 13 years ago

Its a very weird error, I was also facing the same problem.