AQuA Online voice quality testing service

  • Posted:
  • Proposals: 2
  • Remote
  • #128497
  • Completed
Zev T.Alex J. have already sent a proposal.
  • 2

Description

Experience Level: Intermediate
AQuA Online – voice quality testing service requirements


Online service is devoted to provide its users with a possibility to evaluate voice quality by comparing reference audio with audio under test. Typical procedure of working with the service is the following:

user uploads uncompressed reference audio file in .wav format (limited by 500K)
user uploads uncompressed degraded (under test) audio file in .wav format (limited by 500K)
user selects what file(s) to test against what file(s) from available uploads
user checks default set of AQuA parameters and applies changes by turning some switches on/off or providing numerical values for where appropriate (parameters are described in AQuA Manual and AQuA DLL files, hints for each parameter are taken from AQuA DLL document)
system invokes AQuA command line application and makes on the fly test by comparing two audio files.
After test execution system creates correspondent records and test report (similar to http://www.sevana.fi/test/index. php): files that were tested with a possibility to download and listen to them (embedded player with a waveform view), parameters used for the test, AQuA version that was used for the test, contents of report.txt (reasons for audio quality loss), spectrum graph with a possibility to download it as a .csv file.
Each test consumes 0.01 EUR from user account
Payments arranged by PayPal (web service integration)
Basic protection from multiple registration of the same user under different accounts, proposal only
Tests are allowed only if user has positive account balance
GUI of the front page should show number of tests performed


Online service requires the following:

user registration (each new user gets 1 EUR on his account, but this feature may be turned off, so that user gets 0 on his account and online payment is required to work with the system)
user dashboard (simplified file manager view)
HTTP based API for audio quality testing (POST method, simple, return parameters are MOS or % and delivered as text, but test results are stored under user account, so that the user can check them by logging into the web site)
Besides regular data provided for test execution there must be a hidden filed that is used for different type of users tracking (web, mobile, other)


There is one type of user: premium rate

max. file size to upload 500K – controlled by the system
audio files must be 8kHz, Mono, 16 bit, uncompressed .wav (pcm inside .wav)
single file upload – no multiple files, no archives
unlimited file storage and system keeps information on all tests made by the user
simple search or filter to search for tests in the system



Main emphasis is to simplicity and usability, so graphical user interface preferably similar to Asterisk VQM or sevana.fi

The system does require integration with online payment systems, PayPal, also suggest other

Typical scenario is when a user registers into the system to evaluate voice quality of two files: reference (original) and file containing audio quality impairment (degraded). After registration user may be willing to use the system via HTTP API only, however, simple dashboard with recorded tests should be available upon user's login.

Note: uncompressed voice files used in tests are typically of a low size and therefore the idea is to upload uncompressed audio

GUI requirements are not different to what Asterisk VQM has already now. http://www.sevana.fi/AsteriskVQM-slow2.avi contains video of Asterisk VQM Framework a very poor graphical design, but the idea is that design is light and easy, other design proposal is welcome.

Some free player should be embedded in the system, so that users can listen to audio files. This gives good visibility on the wave form and perceptual quality.

New Proposal

Create an account now and send a proposal now to get this project.

Sign up

Clarification Board Ask a Question

    There are no clarification messages.