Skip to content

Typical STI Applications

Two examples are provided on this page:

Open plan office measurements also include STI measurements, described here.

Tip

Please refer to Section 7 of IEC 60268-161 for more application examples.

Unamplified Speech

Typical unamplified speech applications include a teacher in a classroom or lecture theatre, or an unamplified actor in a drama theatre. IRIS will measure the impulse response and assist with calculating the speech level at the receiver position, but the background noise should be measured with a calibrated sound level meter.

Sound Source Requirements

The source loudspeaker should model the directivity of a human talker, such as a mouth simulator or suitable directional loudspeaker (see Section 7.2 of1, and2 for directivity specifications).

The loudspeaker should be equalised to have a flat frequency response (±1 dB from 88 Hz to 11.3 kHz).

System Calibration

The complete IRIS system including equalised sound source should be calibrated to measure sound strength (G). Since the source is directional, it must be calibrated on-axis in a free field. Please see here for details.

Speech Level

IRIS can determine the speech level at the receiver position from the measured G value and an appropriate anechoic speech effort (see Speech Efforts for details).

For unamplified speech, we recommend a suitable anechoic vocal effort from Cushing et al. or IEC 60268-16.

The gender of the resulting STI value should match the gender of the speech effort. For example, the STI Female value should be evaluated with a female speech spectrum.

Steps

  1. Calibrate your IRIS system with the sound source in order to enable measurement of G.

  2. Place the sound source in an appropriate position and aim it in a sensible direction.

    Note

    The STI standard 1 provides little guidance on aiming the source in unamplified situations. Often there isn't one specific talking direction so it may be helpful to evaluate STI with the source pointed in multiple directions. For example, in a lecture theatre, the source could be placed at a typical teaching position and aimed to the back of the room (teacher facing students) and also the front of the room (teacher's back to students) to evaluate a best and worst case STI.

  3. Place the receiver at an appropriate position.

  4. Record the impulse response.

    Important

    The level of the excitation signal should be set to achieve a high quality impulse response. Remember to only adjust the level using the output gain slider in IRIS. Any hardware settings affecting the output level must be set as they were during system calibration.

  5. Measure the background noise at the receiver position with a sound level meter.

  6. In the SPEECH page, create a new speech/noise scenario.

  7. Select one of the anechoic speech efforts as described above and IRIS will populate the speech level values.

  8. Enter the background noise values into the speech/noise scenario.

Sound Systems

Sound Source Requirements

  • For testing a sound system with no specific input or a pre-recorded message, electrically inject the sine sweep excitation into a line input.

  • If the input to the sound system is acoustical (i.e. a microphone), use a mouth simulator or suitable directional loudspeaker (see Section 7.2 of 1, and 2 for directivity specifications).

    The loudspeaker should be equalised to have a flat frequency response (±1 dB from 88 Hz to 11.3 kHz).

Important

The sound system must be configured so that any test signals follow the same signal path used in normal operation. Any non-linear processing should be disabled (see Non-Linear Processing).

System Calibration

IRIS system calibration is not applicable to measuring the STI of sound systems. The speech level at the receiver must be determined outside of IRIS.

Speech and Background Noise Level

The effect of the speech-to-noise ratio on STI will be negligible and can be ignored in most situations where it easily exceeds 15 dB.

However, if the speech-to-noise ratio is poor, it is important to calculate the speech and noise level at the receiver position and add a speech/noise scenario in IRIS with this data. The speech and background noise levels must be determined with other equipment (tips below).

Steps

  1. Either connect the stimulus output from IRIS to a line input on the sound system, or setup a directional loudspeaker at a talker position on-axis to a microphone.

  2. Place the receiver at a suitable position.

  3. Record the impulse response.

    Important

    The level of the excitation signal should be set to achieve a high quality impulse response.

  4. You may need to measure the speech and background noise level at the receiver position and include this as a speech/noise scenario in IRIS, especially if the normal operating level of the sound system does not exceed the background noise by 15 dB.

Tip

One way to measure the speech level at the receiver:

  1. A pink noise signal could be used as the base stimulus which is then filtered according to one of the IEC 60268-16 spectra.

    A STIPA signal (which is derived from pink noise) could also be used. By default this is encoded with the IEC 60268-16 male spectrum.

  2. Play this back through the sound system at the typical operating volume and measure the level at the receiver position.

  3. This measurement will include background noise, so make a separate measurement of background noise and remove this from the speech level.

Alternatively, you may use real voice recordings, and this also applies to systems with pre-recorded messages, such as emergency announcement systems. Since there are gaps in voice recordings, follow the instructions in Annex J of IEC 60268-16 1 for how determine the average speech level.


  1. International Standard IEC 60268-16 Ed. 4.0, Sound system equipment - Part 16: Objective rating of speech Intelligibility by speech transmission index (2011). 

  2. ITU-T P.51, Telephone transmission quality - Objective measuring apparatus - Artificial mouth