Lab

Are you developing a new voice device, determining how a product inter-operates with your network, or investigating a customer complaint at one of the major service providers in the world?

MultiDSLA has been providing performance analysis for engineers in the lab with incredible resolution, clarity and flexibility for many years. Things we hear as important from our customers are:

Depth of Analysis

We are regularly told by our customers that only MultiDSLA allows them to really look into what causes a particular MOS.  Shown in the thumbnail image on the left is the 2D error view. This gives a visual indicator of error locations in a recording, incorporates synchronised audio playback, area selectors, user controllable time ruler, detail zoom, utterance markers and much more.  In addition to the 2D error surface view, we have developed extensive visualisations which help in the analysis of many voice-service problems.

Quick Intuitive Scripting

The MultiDSLA test script engine has been developed with input from our users over many years. Scripts are created using an integrated graphical editor, or they can be edited directly in a text editor. MultiDSLA is shipped with example scripts. But real power comes in being able to create the test the way you see fit.

Test Automation

We understand that speech quality performance testing is just one part of a product test system. We provide an extensive programming API, with specific support for Python, Perl & TCL as well as a mode for any general purpose programming language including C# and C++. Alternatively, you can use MultiDSLA to control other systems while running a test.

Many access options

Access options include 2-wire 600 Ohm, 2-wire complex impedance, balanced 600 Ohm, balanced high impedance and handset analogue (these are all on our main hardware device). We also provide SIP, ISDN BRI, ISDN PRI, E1 and T1 interface options. MultiDSLA was designed to make testing between any 2 interfaces as simple as joining two dots.

Easy integration with other systems

Use MultiDSLA in conjunction with other test devices. Work alongside a Bulk Call Generator to monitor the effects of load on speech quality. Add speech quality measurement to a VoLTE signalling tester.

Generate the right tone

view in MultiDSLA of DTMF tone generated with some out of spec tonesIn addition to replaying speech at any level, many other test signals are available. Sine wave single tones, swept or stepped, dual tones, multi-tones, white and pink noise, DTMF with user definable frequency offset and twist - all accurately delivered at any level for any duration. Combine any two signals using the DSLA’s two tracks. Apply real-time equalisation to shape the spectrum of the speech file to the system under test while it is being played out.

Share results with colleagues

Test results are stored in a database. The data can be exported to various file formats and an individual test result can be saved in a Malden result file. This means you can send all the details of a test result to a colleague, where they can be viewed using the free Speech Performance Viewer application. The colleague may be the developer or an expert in a specific area. Being able to package all the result data and graphics into a single file makes communications simpler.

Open directly in Excel

The test results generated by MultiDSLA can be viewed in numerous ways within the application. At some point, however, it is useful to view and manipulate the results in Excel. This is supported with a single click in the report generator.  MultiDLSA profiles help you extract different slices through your result data. Determine that certain measurements help you identify a particular problem and simply save these in a profile to make future report generation quick and easy.

Sound File Viewer

Open and view sound file recordings; according to a customer survey, MultiDSLA provides 95% of the functionality they currently use in sound file editors such as Adobe Audition. In fact, MultiDSLA is optimised for analysing speech recordings simultaneously in both the time and frequency domains, combined with a tracking playback facility so you can track visually what you are listening to.

Automated Call Control & Analysis

The first step in measuring speech quality, or any other aspect of call quality, is the establishment of a phone call. In MultiDSLA test design and call-setup configuration are purposely separated.  Create a set of standard tests and re-use them on any test setup – standard tests mean you avoid adding unexpected changes when editing scripts.  Call setup properties, such as telephone numbers, are specified in node configurations.   In addition to making call setup more manageable, MultiDSLA also integrates advanced call setup analysis features, to help on those occasions when you need to figure out why a call setup failed.