The following is an tentative schedule for VoicePrivacy 2024 and subject to change. Any specific times are for Anywhere on Earth (AoE).
| Release of evaluation data, software and baselines and evaluation plan | 8th March 2024 |
| Deadline for participants to submit a list for training data and models | 20th March 2024 |
| Publication of the full final list of training data and models | |
| Evaluation plan v2.0 with new baselines | 1st April 2024 |
| Submission of challenge papers to the joint SPSC Symposium and VoicePrivacy Challenge workshop | |
| Deadline for participants to submit system descriptions (containing the title, authors, team name, abstract and an expected number of systems to be submitted)) | 15th June 2024 |
| Deadline for participants to update descriptions | 18th June 2024 |
| Extended deadline for participants to submit evaluation results and anonymized data | 18th June 2024 |
| Author notification for challenge papers | |
| Final paper upload | |
| Joint SPSC Symposium and VoicePrivacy Challenge workshop | 6th September 2024 |
Registration for the VoicePrivacy 2024 Challenge
Participants are requested to register for the evaluation. Registration should be performed once only for each participating entity using the following form: Registration.
You will receive a confirmation email within ~24 hours after successful registration, otherwise or in case of any questions please contact the organizers:
organisers@lists.voiceprivacychallenge.org
Registration for the Workshop
Registration for the VoicePrivacy 2024 Challenge & SPSC workshop is available through the registration portal for Interspeech 2024 (https://interspeech2024.org/registration-payment/#) with prices listed below (https://interspeech2024.org/satellite/):
You will be able to add the workshop registration as an option during the main conference registration process.
Not attending Interspeech?
For those not attending Interspeech, a standalone registration option is available. Please fill the Registration Form below to register using this option (for both onsite & online participants):
Deadline extended: June 15th 22nd 2024 (AoE).
Submission link: ISCA-Speech SIG-SPSC 2024 Symposium paper submission.
In the submission interface window, for the VoicePrivacy 2024 papers, please add in the field:
TL;DR "Too Long; Didn't Read": a short sentence describing your paper
the following information:
VoicePrivacy-2024: <Your team name used in the VoicePrivacy 2024 challenge registration form>
The reviews for papers (unlike system descriptions) will be double-blind, please remove the author names from the submitted files for the papers.
Format is the same as for the Interspeech 2024 paper: template, 4 to 6 pages long excluding references. Submitted papers will undergo peer review via the regular ISCA SPSC Symposium review process, though the review criteria applied to regular papers will be adapted for VoicePrivacy Challenge papers to be more in keeping with systems descriptions and results.
Submission of the system descriptions is required for all VoicePrivacy 2024 Challenge participants.
Submition link: https://vpc-2024.sciencesconf.org
Deadline to submit system descriptions (an initial submit should contain the title, authors, team name, abstract and an expected number of systems to be submitted (can be updated later)): June 15th 2024 (AoE).
Deadline for participants to update system descriptions (with complete information about all the submitted systems): June 18th 2024 (AoE).
The system descriptions should not be anonymized, please specify the author names in the submission files.
Format is the same as for the Interspeech 2024 paper: template, 2 to 6 pages long excluding references.
Submission of the anonymized data and corresponding results is required for all VoicePrivacy 2024 Challenge participants.
Data and results submission should be done using dropbox file request links from the organizers (a dropbox account is not required). Submission links and corresponding passwords will be sent individually to each team upon receiving their system description.
Extended deadline: June 18th 2024 (AoE).
In case of any technical problems that may cause delay in uploading the data, participants can contact the organizers by email organisers@lists.voiceprivacychallenge.org before June 18th 2024 specifying their team name and send the full list of the expected submission files with the corresponding md5sum values for each archive file to be submitted. Upon request, the deadline for data upload can be extended for the team, but only files with the provided (before the deadline) md5sum will be taken into consideration in the challenge ranking.
Please check Section 8.2 Submission of results of the evaluation plan.
In addition to these rules:
There is no limit on the number of submitted systems. The number of submitted systems should correspond to the number of submitted compressed archive files.
Each single system submission should include a compressed archive (around 35-36 GB) containing:
Directories with the result files, the corresponding cosine similarity scores (saved in exp/asv_orig/cosine_out and exp/asv_anon/cosine_out), and additional information generated by the evaluation scripts: • exp/results_summary • exp/asv_orig • exp/asv_anon • exp/asr • exp/ser/*csv.
The corresponding anonymized speech data (wav files, 16 kHz, with the same names as in the original corpus) generated from the development and evaluation sets and from the LibriSpeech-train-clean-360 dataset used to train the ASV evaluation model. For evaluation, the wav files will be converted to 16-bit signed integer PCM format, and this format is recommended for submission. Directories with anonymized audio files include:
IEMOCAP_dev_<system_name>
IEMOCAP_test_<system_name>
libri_dev_enrolls_<system_name>
libri_dev_trials_f_<system_name>
libri_dev_trials_m_<system_name>
libri_test_enrolls_<system_name>
libri_test_trials_f_<system_name>
libri_test_trials_m_<system_name>
train-clean-360_<system_name>
<Team name>_<System name>.tar.gz (or zip).Team name - team name used in the registration form. (Please also specify your team name in the system description.)
System name - system name or id - should be exactly the same in the submission file archive and in the system description.
All the submitted systems must be described in detail in the system descriptions.
The challenge ranking will include only those systems which are compliant with the challenge rules (described in the evaluation plan) and for which we receive before the deadlines:
System description should contain all the results for the submitted systems.
Challenge papers and system descriptions may also include additional results without data submissions, but these results will not be included in the official ranking.
We also recommend to add tables (similar to Tables 6-8 in the evaluation plan that provide the summary descriptions of each submitted system (modules/components with their descriptions, input/output features, training data and open-source models used in training/development)).
System descriptions and results submitted after the deadlines will be marked as late submission in results ranking.