Closed Tortoise17 closed 3 years ago
You will likely need to install the sox plugin for mp3. Its usually in a separate package on most linux distros.
Tortoise17 @.***> schrieb am Mi., 31. März 2021, 18:04:
I am facing error.
steps/make_mfcc.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.run.pl: 28 / 28 failed, log is in exp/make_mfcc/commonvoice_train/make_mfcc_commonvoice_train.*.log
and inside the log is
compute-mfcc-feats --write-utt2dur=ark,t:exp/make_mfcc/commonvoice_train/utt2dur.1 --verbose=2 --config=conf/mfcc.conf scp,p:exp/make_mfcc/commonvoice_train/wav_commonvoice_train.1.scp ark:- | copy-feats --write-num-frames=ark,t:exp/make_mfcc/commonvoice_train/utt2num_frames.1 --compress=true ark:- ark,scp:/home/desktop/Desktop/research_and_development/lab_work/workshop/speech_lab/kaldi/egs/csj/s5/mfcc/raw_mfcc_commonvoice_train.1.ark,/home/desktop/Desktop/research_and_development/lab_work/workshop/speech_lab/kaldi/egs/csj/s5/mfcc/raw_mfcc_commonvoice_train.1.scp
Started at Wed Mar 31 18:00:26 CEST 2021
# copy-feats --write-num-frames=ark,t:exp/make_mfcc/commonvoice_train/utt2num_frames.1 --compress=true ark:- ark,scp:/home/desktop/Desktop/research_and_development/lab_work/workshop/speech_lab/kaldi/egs/csj/s5/mfcc/raw_mfcc_commonvoice_train.1.ark,/home/desktop/Desktop/research_and_development/lab_work/workshop/speech_lab/kaldi/egs/csj/s5/mfcc/raw_mfcc_commonvoice_train.1.scp compute-mfcc-feats --write-utt2dur=ark,t:exp/make_mfcc/commonvoice_train/utt2dur.1 --verbose=2 --config=conf/mfcc.conf scp,p:exp/make_mfcc/commonvoice_train/wav_commonvoice_train.1.scp ark:- sox FAIL formats: can't open input file `data/wav/cv/clips/common_voice_de_17298952.mp3': No such file or directory ERROR (compute-mfcc-feats[5.5.899~1-3d0e4313]:Read4ByteTag():wave-reader.cc:56) WaveData: expected 4-byte chunk-name, got read error
while I tried to change at prepare_commonvoice.py in local and replaced sox with lame but didn't help.
If I can use ffmpeg? and where is required/ if you can guide me.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/uhh-lt/kaldi-tuda-de/issues/52, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACKGA6XFH4IMJB3OEURNATTTGNBYHANCNFSM42E7NWMA .
Can you tell me package name? so that I can install for CentOS? Because I searched alot and failed already before my first message.
yum install sox soxr sox-plugins-freeworld -y
you mean this? or if I misunderstood?
For ubuntu its:
sudo apt-get install libsox-fmt-mp3
Sorry I dont have yum/centos
On Wed, Mar 31, 2021, 9:38 PM Tortoise17 @.***> wrote:
yum install sox soxr sox-plugins-freeworld -y you mean this? or if I misunderstood?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/uhh-lt/kaldi-tuda-de/issues/52#issuecomment-811386245, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACKGA6W66KFZXC6CIANINPTTGN24RANCNFSM42E7NWMA .
Thank you ..
I am facing error.
and inside the log is
while I tried to change at prepare_commonvoice.py in local and replaced sox with lame but didn't help.
If I can use ffmpeg? and where is required/ if you can guide me.