train timit data use refered to tedlinu v1 and get mtrc0_si479.wav: Permission denied error

[zhangjiulong]

Hi

I want to test eesen on timit data, but I got some error when extracting fbank features. the error like this

copy-feats --compress=true ark:- ark,scp:/home/zhangjl/git/asr/eesen/asr_egs/timit/v1/fbank/raw_fbank_train.20.ark,/home/zhangjl/git/asr/eesen/asr_egs/timit/v1/fbank/raw_fbank_train.20.scp
extract-segments scp,p:data/train/wav.scp exp/make_fbank/train/segments.20 ark:-
compute-fbank-feats --verbose=2 --config=conf/fbank.conf ark:- ark:-
sh: /home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_si479.wav: Permission denied
ERROR (extract-segments:Read4ByteTag():wave-reader.cc:74) WaveData: expected 4-byte chunk-name, got read errror
WARNING (extract-segments:Read():feat/wave-reader.h:149) Exception caught in WaveHolder object (reading).
WARNING (extract-segments:HasKeyInternal():util/kaldi-table-inl.h:1370) RandomAccessTableReader: error reading object from stream '/home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_si479.wav |'
WARNING (extract-segments:main():extract-segments.cc:125) Could not find recording mtrc0_si479, skipping segment mtrc0_si479-0000000-0000436
WARNING (extract-segments:Close():kaldi-io.cc:446) Pipe /home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_si479.wav | had nonzero return status 32256
sh: /home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_sx119.wav: Permission denied
ERROR (extract-segments:Read4ByteTag():wave-reader.cc:74) WaveData: expected 4-byte chunk-name, got read errror

I process the data as follows:

I convert the sounds of timit to wav format as follows

run cmd as follows:

file mtrc0_si479.wav

and got the result

si2005.wav: NIST SPHERE file

so I use sox convert sphere format to wav format, the result like this

RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16 bit, mono 16000 Hz

I transform timit data like tedlium data format, the only difference is the wav.scp file.

tedlimu' wav.scp is like

ichaelSpecter_2010 /home/zhangjl/git/asr/eesen/asr_egs/tedlium/v1/../../../tools/sph2pipe_v2.5/sph2pipe -f wav -p /home/zhangjl/git/asr/eesen/asr_egs/tedlium/v1/db/TEDLIUM_release1/test/sph/MichaelSpecter_2010.sph |

my wav.scp is like

faem0_si1823 /home/zhangjl/dataCenter/asr/timitTedFormat/test/wav/faem0_si1823.wav |

Look forward your suggestion. Thx

[naxingyu]

your wav.scp is not correct. you are using the wave file itself as a feeding into the pipeline. remove the '|' at the end of each line.

在 2016/6/2 17:57, zhangjiulong 写道:

Hi

I want to test eesen on timit data, but I got some error when extracting fbank features. the error like this

|copy-feats --compress=true ark:- ark,scp:/home/zhangjl/git/asr/eesen/asr_egs/timit/v1/fbank/raw_fbank_train.20.ark,/home/zhangjl/git/asr/eesen/asr_egs/timit/v1/fbank/raw_fbank_train.20.scp extract-segments scp,p:data/train/wav.scp exp/make_fbank/train/segments.20 ark:- compute-fbank-feats --verbose=2 --config=conf/fbank.conf ark:- ark:- sh: /home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_si479.wav: Permission denied ERROR (extract-segments:Read4ByteTag():wave-reader.cc:74) WaveData: expected 4-byte chunk-name, got read errror WARNING (extract-segments:Read():feat/wave-reader.h:149) Exception caught in WaveHolder object (reading). WARNING (extract-segments:HasKeyInternal():util/kaldi-table-inl.h:1370) RandomAccessTableReader: error reading object from stream '/home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_si479.wav |' WARNING (extract-segments:main():extract-segments.cc:125) Could not find recording mtrc0_si479, skipping segment mtrc0_si479-0000000-0000436 WARNING (extract-segments:Close():kaldi-io.cc:446) Pipe /home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_si479.wav | had nonzero return status 32256 sh: /home/zhangjl/dataCenter/asr/timitTedFormat/train/wav/mtrc0_sx119.wav: Permission denied ERROR (extract-segments:Read4ByteTag():wave-reader.cc:74) WaveData: expected 4-byte chunk-name, got read errror |

I process the data as follows:

1. I convert the sounds of timit to wav format as follows run cmd as follows:

|file mtrc0_si479.wav |

and got the result

|si2005.wav: NIST SPHERE file |

so I use sox convert sphere format to wav format, the result like this

|RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16 bit, mono 16000 Hz |

1. I transform timit data like tedlium data format, the only difference is the wav.scp file. tedlimu' wav.scp is like

ichaelSpecter_2010 /home/zhangjl/git/asr/eesen/asr_egs/tedlium/v1/../../../tools/sph2pipe_v2.5/sph2pipe -f wav -p /home/zhangjl/git/asr/eesen/asr_egs/tedlium/v1/db/TEDLIUM_release1/test/sph/MichaelSpecter_2010.sph

my wav.scp is like

|faem0_si1823 /home/zhangjl/dataCenter/asr/timitTedFormat/test/wav/faem0_si1823.wav | |

Look forward your suggestion. Thx

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/srvk/eesen/issues/57, or mute the thread https://github.com/notifications/unsubscribe/ADKpxL7E1PRCkqrpoIrjkpxIwGkXZWaZks5qHqkUgaJpZM4IsY7x.

[zhangjiulong]

@naxingyu Thanks very much, it works.