I have downloaded the database from timit they are .wav files.
when I opened it shows following header NIST_1A
1024
database_id -s5 TIMIT
database_version -s3 1.0
utterance_id -s11 kah0_si2158
channel_count -i 1
sample_count -i 41575
sample_rate -i 16000
sample_min -i -4941
sample_max -i 4924
sample_n_bytes -i 2
sample_byte_format -s2 01
sample_sig_bits -i 16
end_head
How can i convert to .sph file. Is there any code available to convert this.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If your file starts with "NIST_1A" then it's already in NIST Sphere (i.e.
.sph) format, not MS WAV. No need to convert to MS WAV.
Older databases containing audio files in sph format uses extensions like
wav, wv1, etc. Keep in mind these were recorded before Microsoft's ".wav"
files became popular. Only after MS WAV became popular is it that the .sph
extension was adopted.
MS Wav files start with "RIFF", as you already noticed, and the format is
supported by sphinx_fe, as Arseniy already pointed out.
I have downloaded the database from timit they are .wav files.
when I opened it shows following header
NIST_1A 1024 database_id -s5 TIMIT database_version -s3 1.0 utterance_id
-s11 kah0_si2158 channel_count -i 1 sample_count -i 41575 sample_rate -i
16000 sample_min -i -4941 sample_max -i 4924 sample_n_bytes -i 2
sample_byte_format -s2 01 sample_sig_bits -i 16 end_head
How can i convert to .sph file. Is there any code available to convert
this.
Thank you sir for quick reply.
I have an another wav file which is of RIFF not NIST header. It is not supported by sphinx.
I have attached that wav file below. For that also i need to use the same.
sox -t wav file.wav -t sph ofile.sph
You do not need to convert it. Decoder should work with this file directly. For training, you need to specify in the config file the following options:
I have downloaded the database from timit they are .wav files.
when I opened it shows following header
NIST_1A 1024 database_id -s5 TIMIT database_version -s3 1.0 utterance_id -s11 kah0_si2158 channel_count -i 1 sample_count -i 41575 sample_rate -i 16000 sample_min -i -4941 sample_max -i 4924 sample_n_bytes -i 2 sample_byte_format -s2 01 sample_sig_bits -i 16 end_headHow can i convert to .sph file. Is there any code available to convert this.
You can convert wav to sph with sox:
sox -t wav file.wav -t sph ofile.sphSphinx supports both .sph and .wav files, so I think you do not need to convert actually
Pradeep,
If your file starts with "NIST_1A" then it's already in NIST Sphere (i.e.
.sph) format, not MS WAV. No need to convert to MS WAV.
Older databases containing audio files in sph format uses extensions like
wav, wv1, etc. Keep in mind these were recorded before Microsoft's ".wav"
files became popular. Only after MS WAV became popular is it that the .sph
extension was adopted.
MS Wav files start with "RIFF", as you already noticed, and the format is
supported by sphinx_fe, as Arseniy already pointed out.
--Evandro
On 6 November 2016 at 06:46, Pradeep S V pradeepsv@users.sf.net wrote:
Thank you sir for quick reply.
I have an another wav file which is of RIFF not NIST header. It is not supported by sphinx.
I have attached that wav file below. For that also i need to use the same.
sox -t wav file.wav -t sph ofile.sph
Your file looks OK at first glance:
file 0001.wavYou do not need to convert it. Decoder should work with this file directly. For training, you need to specify in the config file the following options:
$CFG_WAVFILE_EXTENSION = 'wav';
$CFG_WAVFILE_TYPE = 'mswav';