Product Listing

Digital Speech Level Analyser II

Download the DSLA II Brochure for a detailed performance description.

The new Digital Speech Level Analyser II extends the performance of the earlier DSLA. It predicts a speech quality score and measures the characteristics of the speech channel. It generates a linguistically motivated artificial speech test signal and evaluates the resultant signal after it has been processed by any network or network element, even those subject to packet loss or variations in delay. The Phonytalk™ process can use any user-defined signal as stimulus. This versatile instrument is essential equipment for developers and operators of voice activated devices, codecs, echo-cancellers, transcoders, VoIP gateways, telephones and networks.

• Measure the quality of the speech transmission in terms of listening quality on the MOS scale of one to five
• Generate the Phonytalk™ linguistically motivated artificial speech test signal in different languages
• Measure, repeatably, the mean active speech level in accordance with ITU-T Recommendation P.56 Method B
• Measure delay, echo, noise
• Measure Voice Activity Detector operation
• Measure Comfort Noise Level in the presence of speech
• Record the received signal for later analysis
• Create your own complex signals using a wide range of programmable sources and an optional Equaliser
• Optional DTMF Analysis
• Optional ISDN Basic Rate Interface
• Optional VoIP Virtual Terminal to set up calls over SIP or H.323
• Optional Codec Library supporting G.729A, G.729B and G.723.1
• Make test calls, measure and record the results as well as the signal from the network element under test
• Windows operating environment makes test set up simple
• Bring the test process to your laptop or desktop by controlling the DSLA locally over Ethernet or remotely over the Internet
• Log GPS data with results

Using a Digital Speech Level Analyser II to determine the characteristics of a speech path stimulated with the Phonytalk™ test signal will provide a useful insight into the path performance with real speech. Both the PAMS and the PESQ processes compare the original and received signal and determine Mean Opinion Score predictions on a scale of 1-5. The linguistically motivated artificial speech signal contains a representative selection of sounds put together in legal strings to provide a brief speech signal that does not contain the redundancy of real speech.

The Phonytalk™ test signal speech level and duration can be defined entirely by the user. In addition to the linguistically motivated artificial speech test signal, the user can include any other speech material of any duration and add noise or other interfering signals to build up complex test processes.

The Digital Speech Level Analyser II comprises a signal processor connected to the serial port or Ethernet port of a host computer.

A Windows user interface is provided for an optimum environment so that test processes can be quickly defined and implemented. The user has options to align the signal levels to the system under evaluation, to generate the Phonytalk™ test signal in a variety of modes, to record the results of measurements on the received signal for subsequent analysis and to record the received signal. The two channel Digital Speech Level Analyser operates in 4-wire mode, at line level or microphone level, or as a telephone device.

Language
American English, British English.
Voice type - Male, Female
Speech type - Single Dual Conversational
Bandwidth - 8kHz or 4kHz flat or Modified IRS Send Characteristic

Speech Test Signals
User defined 16k or 8k sample rate .WAV
Optional Equaliser defines any spectral characteristic for Test Signal
Use any Speech Level from -60 to 0dBm
Test Signal of any duration streamed from mass storage device
Conversation with/without double-talk
Store up to 4 minutes of speech in DSLA II memory

Other Output Signals
Tone, Swept Tone, Tone burst, DTMF sequences, white noise, noise burst

Speech Measurements
Mean Active Speech Level, Peak, Activity Factor, Doubletalk measure

Signal Analysis
Record input signal
Measurement Period - 1ms to 4,200,000 seconds with 1ms resolution
View perceptual analysis waveforms
Detailed time offset (delay) data
Measure the proportion of lost speech
Detect all 16 DTMF digits and tones
Optional DTMF Analysis measures twist and frequencies of up to 20 digits

• Utterance and Frame time offset statistics
• Frame by frame score data
• Voice activity and clipping analysis
• Spectral data
• Voicing analysis

Speech Quality Measurements
Mean Opinion Score prediction on a scale of 1-5 for Listening Effort and Listening
Quality from PAMS or PESQ Score, PESQ-LQ, PESQ-P.862.1 and Ie from PESQ.
PSQM/PSQM+ option

Signal Delay
Measure difference between start of output signal and start of input signal to 1msec resolution

Input Connections
2 x Signal Input Balanced 600ohms or 2Mohms CMR>60dB 10Hz to 15kHz AC coupled via 4mm sockets
2 x RJ22 handset connector provides input connection to an IP, digital or cellular phone

Speech Measurements
Speech Level Measurement Range -65dBV to +5dBV with overload for less than 1% of the time
Maximum Measurable Input +17dBV or 7Vrms
Minimum Measurable Input -85dBV less than 1dB error, -65dBV for Speech Level measurement
Noise in speech to 20dB below mean active speech level

Output Connections
2 x Signal Output Balanced 600ohms or 25ohms coupled via 4mm sockets
Maximum Output Level +12dBV
2 x RJ22 handset connector provides attenuated output to handset port
Maximum Output Level -14dBV

Phone Connection
2 x US Modular Jack, DTMF or Pulse dialling

Other Specifications
Power supply - external 9-18VDC or universal mains power cube 90-264VAC 50/60Hz supplied
Monitor Output - 13dB to 25dB below input, +4dBV maximum output, 25ohm output impedance
10/100Mbps Ethernet or RS232 control
Calibration Cycle - 3 years
Temperature Range - 2 to 40 deg C
Size - 220mm wide x 200mm deep x 75mm high
Weight - 3kg