STI/PA Analyzer

About this tool

SignalScope's STI/PA Analyzer tool supports measurements that can be used to determine the intelligibility of speech when broadcast through a speech transmission channel, such as a public address or sound reinforcement system. A speech transmission channel may be electronic, electroacoustic, or purely acoustic. SignalScope supports measurements of each type of channel or system.

The level of intelligibility is quantified with a value called the Speech Transmission Index (STI).

Traditional full STI measurements can be time consuming due to the need to measure the response of the system in 7 different octave frequency bands at 14 different modulation frequencies, which requires 98 separate measurements. For this purpose, the Speech Transmission Index for public address systems, or STIPA, was developed. STIPA measurements rely on a specially designed test signal that allows measurement of the system's response to 14 different modulation frequencies (2 in each octave band), simultaneously, which drastically reduces measurement time.

It is also now possible to perform a faster version of the full STI by using test signals similar to that for STIPA to measure all modulation frequencies in all octave bands with 7 measurements instead of 98.

With the STI/PA tool, you can perform direct-method STIPA measurements using an external STIPA signal generator or the tool's built-in generator. Accelerated full STI measurements (direct method) are also supported using the built-in test signal generator. The STI/PA tool's full STI measurements rely on the internal generator to properly synchronize the modulation frequencies and octave bands being analyzed with those being generated and broadcast through the system under test.

This tool was designed to meet or exceed the requirements of IEC standard 60268-16, Edition 5.0 (2020-09). The user is encouraged to acquire and become familiar with this standard, its benefits and limitations, and its best practices for ensuring accurate results and proper interpretation of those results.

Getting Started

The STI/PA Analyzer tool employs the concept of a test which can include several measurements. In general, when determining the STI or STIPA value of a system under test, at least 3 measurements of the system are recommended, from which the STI/PA tool will calculate a final averaged result.

Please note the following in regard to tests in the STI/PA tool:

• Test results and data are stored in human-readable JSON files.
• A test may either be "live" or "loaded."
  • A live, or active, test can have more measurements added just by pressing the start button in the app's top toolbar. A new test can be created by tapping or clicking on the New Test button. If too few measurements exist in the current live test, or if there is too much variation in the individual test results, you may be prompted and given the option to add more measurements to the current test before creating a new one.
  • A test can be loaded from an existing file at any time.
  • Only a single test (live or loaded) may be open at a time.
• Each time a measurement completes by running for the user-selected duration, it will automatically be saved along with existing test data.
• If a measurement is manually stopped early (before the selected measurement duration has been reached), it will not be saved. This allows quick level tests or troubleshooting to be performed without adding an invalid measurement to the overall test results.
• The last test that was open when SignalScope exited will be reloaded the next time the app is launched.
• The user interface presents a list of live, average, and individual measurements in the current test.
  • The live measurement will show data from the last measurement made in a live test.
  • The average measurement ("Avg") will show the average results from all selected measurements in the current test.
  • Any measurement in the list may be selected for display by tapping or clicking on its name.
  • Individual measurements can be selected or de-selected for inclusion in the average results by long-pressing/clicking on the name of the individual measurement (e.g. "M1"). A green dot will appear in the upper righthand corner of the box containing the measurement's name when that measurement is included in the average.
  • If the input signal is clipped (overloaded) at any point during a measurement, the outline around the name of the measurement in the list will turn red.
• Below the list of measurements, you can select from 3 informational displays:
  • Levels
    Levels will show a bar graph of sound levels in the 7 relevant octave frequency bands.
  • MTF
    MTF will show a table of modulation transfer ratios ("m-values") for all modulation frequencies in all octave bands. This represents the modulation transfer function (MTF) matrix from which the STI is calculated.
    Correction factors applied to the results are shown for each frequency band in the row just below the MTF matrix.
    Modulation transfer index (MTI) values are shown for each frequency band in the bottom row of the table.
  • TI
    TI will show a table of transmission index (TI) values for all modulation frequencies in all octave bands. The transmission index represents an effective signal-to-noise ratio (SNR) scaled to a value between 0 and 1.
    Correction factors applied to the results are shown for each frequency band in the row just below the MTF matrix.
    Modulation transfer index (MTI) values are shown for each frequency band in the bottom row of the table.

Additional Specifications

Modulation depth

• STIPA and STI (Fast) test signals employ a modulation depth of 0.55. A perfect STI of 1.0 requires that the received signal also exhibit a modulation depth of 0.55, which would result in a modulation transfer ratio (m-value) of 1.0.
• IEC 60268 requires that the absolute value of any error in m-value determination not exceed 0.05 and that all m-value errors not yield an error in the overall STI value greater than 0.01.
• In our testing of STIPA measurements made with SignalScope, using standard verification signals available from www.stipa.info, we observed differences between measured m-values and expected m-values of no greater than 0.011 across all frequency bands and modulation frequencies.
• Our tests were performed by analyzing the digital signals directly--no analog circuitry was included in the signal path. The STIPA measurement ran with the standard's recommended averaging time of 18 seconds.

Crosstalk between octave-band filters

• Poor design of the octave-band filters can lead to crosstalk which effectively increases the noise level in adjacent frequency bands, causing the m-values to be underestimated. A STIPA measurement device can be tested for this using the standard verification signals avilable from www.stipa.info.
• When testing filter performance using standard verification signals, the m-value in each octave band can be observed with a high-amplitude signal present in an adjacent octave band. IEC 60268 requires that the m-value in the observed band be at least 0.5 (+/-0.05).
• In our testing with SignalScope, the minimum m-value in any observed band with any modulation frequency across all test cases was 0.607.
• Our tests were performed by analyzing the digital signals directly--no analog circuitry was included in the signal path. The STIPA measurement ran with the standard's recommended averaging time of 18 seconds.

Supported octave frequency bands and modulation frequencies:

• STIPA
  

  Octave Band, Hz	125	250	500	1k	2k	4k	8k
  Modulation Freq (Fm)	1.60 & 8.00	1.00 & 5.00	0.63 & 3.15	2.00 & 10.0	1.25 & 6.25	0.80 & 4.00	2.50 & 12.5

• Full STI (Fast)
  

  Octave Band, Hz	125	250	500	1k	2k	4k	8k
  Fm, Meas. Part 1	1.60 & 8.00	1.00 & 5.00	0.63 & 3.15	2.00 & 10.0	1.25 & 6.25	0.80 & 4.00	2.50 & 12.5
  Fm, Meas. Part 2	1.00 & 5.00	0.63 & 3.15	2.00 & 10.0	1.25 & 6.25	0.80 & 4.00	2.50 & 12.5	1.60 & 8.00
  Fm, Meas. Part 3	0.63 & 3.15	2.00 & 10.0	1.25 & 6.25	0.80 & 4.00	2.50 & 12.5	1.60 & 8.00	1.00 & 5.00
  Fm, Meas. Part 4	2.00 & 10.0	1.25 & 6.25	0.80 & 4.00	2.50 & 12.5	1.60 & 8.00	1.00 & 5.00	0.63 & 3.15
  Fm, Meas. Part 5	1.25 & 6.25	0.80 & 4.00	2.50 & 12.5	1.60 & 8.00	1.00 & 5.00	0.63 & 3.15	2.00 & 10.0
  Fm, Meas. Part 6	0.80 & 4.00	2.50 & 12.5	1.60 & 8.00	1.00 & 5.00	0.63 & 3.15	2.00 & 10.0	1.25 & 6.25
  Fm, Meas. Part 7	2.50 & 12.5	1.60 & 8.00	1.00 & 5.00	0.63 & 3.15	2.00 & 10.0	1.25 & 6.25	0.80 & 4.00