Skip to main content

Speaker verification protocol on the NIST SRE 2012

Project description

The 2012 NIST Speaker Recognition Evaluation (SRE12) is part of an ongoing series that starts in 1996.

In this package, we implement speaker recognition protocols (both Male and Female) for the NIST SRE 2012. The file lists of the development set were designed by the I4U consortium during its participation to the competition. Special thanks to Rahim Saeidi for the good work (original link of the lists: http://cls.ru.nl/~saeidi/file_library/I4U.tgz). The file names were then normalized following the PRISM definition.

This package is automatically downloaded/used by xbob.spkrec.nist_sre12 to reproduce the results of Idiap Research Institute at SRE12. xbob.spkrec.nist_sre12 itself relies on xbob.spkrec, an open-source speaker recognition toolbox developed at Idiap. The list files can also be used independently as explained below.

If you use this package and/or its results, please cite the following publications:

  1. The original paper presented at the NIST SRE 2012 workshop:

     @inproceedings{Khoury_NISTSRE_2012,
       author = {Khoury, Elie and El Shafey, Laurent and Marcel, S{\'{e}}bastien},
       month = {dec},
       title = {The Idiap Speaker Recognition Evaluation System at NIST SRE 2012},
       booktitle = {NIST Speaker Recognition Conference},
       year = {2012},
       location = {Orlando, USA},
       organization = {NIST},
       pdf = {http://publications.idiap.ch/downloads/papers/2012/Khoury_NISTSRE_2012.pdf}
    }
  2. The paper that described the development set used by the I4U consortium:

    @inproceedings{Saedi_INTERSPEECH_2013,
       author = {Saeidi, Rahim and others},
       month = {aug},
       title = {I4U Submission to NIST SRE 2012: a large-scale collaborative effort for noise-robust speaker verification},
       booktitle = {INTERSPEECH},
       year = {2013},
       location = {Lyon, France},
       pdf = {to appear}
    }
  3. Bob as the core framework used to run the experiments:

    @inproceedings{Anjos_ACMMM_2012,
      author = {A. Anjos and L. El Shafey and R. Wallace and M. G\"unther and C. McCool and S. Marcel},
      title = {Bob: a free signal processing and machine learning toolbox for researchers},
      year = {2012},
      month = {oct},
      booktitle = {20th ACM Conference on Multimedia Systems (ACMMM), Nara, Japan},
      publisher = {ACM Press},
      url = {http://publications.idiap.ch/downloads/papers/2012/Anjos_Bob_ACMMM12.pdf},
    }

Installation

Just download this package and decompress it locally:

$ wget http://pypi.python.org/packages/source/x/xbob.db.nist_sre12/xbob.db.nist_sre12-1.1.1.zip
$ unzip xbob.db.nist_sre12-1.1.1.zip
$ cd xbob.db.nist_sre12-1.1.1

Use buildout to bootstrap and have a working environment ready for experiments:

$ python bootstrap
$ ./bin/buildout

This also requires that bob (>= 1.2.0) is installed.

Getting the data

You need to order the NIST SRE databases (Fisher, Switchboard, and Mixer):

http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2013S03

Please follow the instructions and the evaluation plan given by NIST:

http://www.nist.gov/itl/iad/mig/sre12.cfm

Depending on the release year, the data may need to be flatten and reorganized. Please, follow the file structure as appearing when running:

$ bin/bob_dbmanage.py nist_sre12 dumplist

For this purpose, you will need the utilities provided by NIST with the database, as well as sox.

Decompressing the data and splitting the audio channels

The data provided by NIST are compressed in a non-standard format. NIST supplies a binary called w_decode to perform the decompression. Therefore, you should decompress all the files using the following command (where NIST_FOLDER/bin is the one containing the w_decode utility:

$ NIST_FOLDER/bin/w_decode -o pcm $compressedfile $decompressedfile

Several files are in stereo and hence contain two audio channels. These files needs to be split using a script similar to the following one:

$ decompressedfileStereo=`basename $decompressedfile .sph`
$ num=`soxi $decompressedfile | grep Channels | cut -c 18`
$ echo $num
$ if [ $num -eq 2 ]
$ then # File is stereo
$   echo sox $decompressedfile -c 1 $outputDir/${decompressedfileStereo}-a.sph mixer -l
$   sox $decompressedfile -c 1 $outputDir/${decompressedfileStereo}-a.sph mixer -l
$   sox $decompressedfile -c 1 $outputDir/${decompressedfileStereo}-b.sph mixer -r
$ else # File is mono
$   echo cp $decompressedfile $outputDir/
$   cp $decompressedfile $outputDir/
$ fi

Adding noise

In order to better represent the SRE12 evaluation set, 2 noisy versions (SNR=6dB and SNR=15dB) of the same segments were included to the development set. This can be done using FaNT:

http://dnt.kr.hsnr.de/download.html

Speech enhancement

The denoising of the audio signal can be done using QIO:

http://www1.icsi.berkeley.edu/Speech/papers/qio/

Using independently the file lists

The file lists of the development and evaluation sets are shipped with this package. They can be used independently, and can be found here:

$ cd xbob/db/nist_sre12/lists/

The file lists of the development set were prepared by the I4U consortium.

In case you need any help, please contact us.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

xbob.db.nist_sre12-1.1.2.zip (11.8 MB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page