--- title: "Calculating STI" slug: "calculating-sti" updated: 2026-09-01T17:06:40Z published: 2026-09-01T17:06:40Z canonical: "doc360.rationalacoustics.com/calculating-sti" --- > ## Documentation Index > Fetch the complete documentation index at: https://doc360.rationalacoustics.com/llms.txt > Use this file to discover all available pages before exploring further. # Calculating STI ## IR Measurement Considerations for STI Smaart can calculate Speech Transmission Index (STI) indirectly from an Impulse Response (IR) measurement provided that certain conditions are met. In general, there are several "moving parts" one needs to keep in mind when performing STI or STIPA measurements using *any* methodology. Ambient noise levels and the operational speech level of the system under test (SUT) are important factors in obtaining accurate estimations of intelligibility. For systems that include a live microphone, the microphone and its acoustical environment along with the acoustical properties of the sound source used to excite the SUT are also important considerations, regardless of the measurement technique. For indirect measurements specifically, where the Modulation Transfer Function (MTF) of the SUT is derived from an impulse response rather than measured directly using specialized test signals, additional requirements are: - A deterministic test Signal (period-matched noise or log sweep) must be used without a data window to measure the IR. - The measurement duration (the FFT Time Constant in our case), per IEC 60268-16:2020, must be at least 1.6 seconds. The latter requirement, however, seems to overlook an important property of Fourier analysis using linearly spaced frequencies. It is true that a conventional Discrete Fourier Transform (DFT) (or FFT) with a 1.59 second time constant would have its first (non-zero) frequency bin at 0.63 Hz (the lowest modulation frequency evaluated in the MTF for STI). But since FFT bins are linearly spaced, the *second* bin is always an octave above the first, whereas modulation frequencies for STI are on 1/3-octave center frequencies. Calculating STI from a 1.6 second IR measurement would therefore require somehow interpolating the missing frequencies in the lowest octaves, and the standard does not address that requirement. Another option would be to simply measure over a longer period. A 5-second DFT time constant, for example, would provide frequency resolution of 0.2 Hz, resulting in frequency bins very close to all of the STI modulation frequencies. For this reason, Smaart offers a DFT size equating to exactly 5 seconds (5000 ms) for all supported sampling rates ≥ 44100 samples/sec. ## Calculating STI from an IR To calculate STI from an impulse response measurement, make sure the IR measurement that you want to analyze – whether it is a new measurement or an IR stored in a .wav file – is visible in the graph in the main window and is the front trace on the graph (at the top of the list in the [Graph Legend](/v1/docs/ir-graph-legend)) if multiple traces are displayed. Click the [All Bands](/v1/docs/all-bands-table) button at the bottom of the [Control Bar](/v1/docs/ir-mode-control-bar) to open the All Bands table and then click the Calculate STI button in the *All Bands* Table to open the *Calculate STI* dialog. ![](https://cdn.document360.io/a725c8f7-3a1a-4c09-9518-5d15d09a2ae2/Images/Documentation/image(44).png?sv=2026-02-06&spr=https&st=2026-11-18T17%3A42%3A10Z&se=2026-11-18T18%3A42%3A10Z&sr=b&sp=r&sig=sy%2Fv3lfKoYswopY7DoBiyu1MUsiUA7oZ3vczYR9VD%2BA%3D) The first thing you will need to tell Smaart is whether the measurement is “Noiseless” or made in the presence of normal ambient background noise levels (*Noise Present*). - *Noise present* measurements require the use of a period-matched, speech-weighted noise signal when capturing the IR. They are measured at the operational sound level of the system, without averaging, in the presence of typical in-service ambient noise levels. The measurement signal input for the transfer function signal pair used to capture the IR measurement must be [calibrated](/v1/docs/sound-level-spl-leq-calibration) for accurate sound level measurement. - *Noiseless* measurements are typically made during quiet times, when ambient noise levels are lower than they would normally be when the system is in use and/or at a higher output level than the normal operating level of the SUT. They require separate estimation of operational speech level and typical in-service ambient noise levels. IR measurements can be made using period-matched Pink Noise exponential sweeps (a.k.a., log sweeps or “pink” sweeps). The signal-to-noise ratio in each octave band from 125 Hz to 8 kHz must be at least 20 dB. IR measurements may be averaged to improve the signal-to-noise ratio. Selecting *Noise Present* is fairly automatic, so long as the IR measurement being analyzed was captured: - Using Smaart 8.5 or later, - With a calibrated measurement signal input, and - With no averaging. Once *Noise Present* is selected, Smaart calculates sound level for the measurement channel while capturing the IR and it is assumed that typical ambient noise levels are included in the measurement. STI, STIPA(IR), and equivalent CIS figures along with overall qualitative assessment (Excellent, Good, Fair, or Poor) and letter grades appear in the *Results* section of the dialog window. For *Noiseless* measurements, you first need to provide the operational Speech Level for the SUT as an A-weighed sound level figure. You will also need unweighted (Lz) noise levels for each octave band that are typical of ambient noise levels present when the system is in normal use. The latter can be imported from a calibrated Smaart RTA measurement by clicking the Import Noise button and selecting the appropriate .srf file, or by manually entering Decibel levels for each octave directly into the table's Noise dB row. You also have the option to estimate the effect of equalization on the STI figure by entering +/- dB values in the EQ dB row of the table. Values in the Results section will update as you make changes. The *Clear EQ* and *Clear Noise* buttons reset the corresponding rows of the table to all zeroes. The *Copy* button copies MTI, EQ, and Noise values from the table along with the MTF “m” values for all modulation frequencies in all octaves to the operating system's clipboard in tab-delimited ASCII format, suitable for pasting into a spreadsheet, text file, or any other program that accepts ASCII text. When you have made your selections you can save your work by clicking the *Save* button. If you are working with a new live measurement, Smaart will prompt you for a file name and then write the IR measurement to a .wav file with the STI calculation details included as metadata in the file header. If you are working with an IR measurement already stored in a .wav file, Smaart will rewrite the file with a new header that includes the STI metadata. In that case, you will be asked to confirm that you want to over-write the existing file. > **Note that the audio sample data in an existing .wav file is untouched when you save STI calculation details, but some programs (particularly older or very basic audio editing/analysis software) may be unable to open the file after saving it.** > > > > **Smaart 8.5 (or later) uses the BWAV header specification for IR data files, which most modern (and probably some not-so-modern) software that works with audio files should be able to parse but it is a good idea to test that assumption before overwriting your only copy of a file created by another program (including older versions of Smaart).** An objective estimate of the intelligibility of human speech transmitted through a transmission medium or system under test. It is based on the relative loss of very low frequency modulation in higher frequency carrier waves. STI is stated as a fractional value between from 0 to 1, where 0 represents a completely unintelligible result and 1 denotes excellent intelligibility with no loss of modulation in transmission. STI is calculated from the modulation transfer functions (MTF) for 7 octave bands from 125 Hz to 8 kHz, and evaluates 13 modulation frequencies in each band. The results for all modulation frequencies in each band are combined into a band-by-band Modulation Transfer Index (MTI) then weighted and summed to produce a single-number STI figure. Separate estimates for the male and female speech may be obtained using different weighting tables when calculating STI. Unless otherwise stated, weighting for male speakers is presumed. The signal that describes the response of a system to an ideal impulsive stimulus in the time domain. The impulse response of linear time invariant (LTI) system is the inverse Fourier transform of its frequency-domain transfer function. (STI for public address systems) is a less rigorous variant of STI (see above) intended specifically for use in measuring public address systems. STIPA has been validated only for male speech. The only functional difference in how STI and STIPA are measured is that instead of evaluating 13 modulation frequencies in each octave band, STIPA uses only two frequencies per band to cut down the time required for direct measurement of the modulation transfer functions (MTF). Since Smaart calculates the MTF of a system indirectly from its impulse response and all 13 modulation frequencies required for STI are also used for STIPA (just not in every band) there is no particular advantage to using STIPA in Smaart The frequency (magnitude and phase) response of a system, function or network. The transfer function of the linear time invariant (LTI) system can be measured directly, using techniques such as dual-FFT transfer function measurements that compare the output of the system to its input signal in the frequency domain, or by taking the Fourier transform of the system’s impulse response. Strictly speaking a signal can be any set of values that depends on some other set of values. In signal processing, the independent variable, e.g., time or frequency, is said to be the domain of the signal. In audio and acoustics, the things we most commonly think of as signals are time domain signals, where voltages or numeric values representing amplitude (the dependent variable) vary over time (the independent variable). But by strict definition, most of the things we see in Smaart could also be called signals, including transfer function Magnitude and Phase and RTA displays where relative energy or phase shift are presented as a function of frequency, rather than time. In physics and engineering the term time constant is most commonly used to denote a time span between reference or threshold points in continuous processes, such as rise or decay time in the step response of filters, heating and cooling times in thermal systems and lag times in mechanical systems. In the context of acoustic measurement we typically use (or perhaps misuse) the term to mean the total time required for discrete processes, such as the time it takes to collect enough samples for an FFT and/or the time span of an impulse response measurement. This is to say that we tend to use the term interchangeably with Time Window. A mathematical technique for determining the spectral content of complex waveforms. The DFT essentially compares the signal being analyzed to a series of sine and cosine waves at regularly spaced (harmonic) intervals to determine how much energy is present at each harmonic frequency. The spacing between frequencies or frequency resolution of the DFT is a function of its size in samples and the sampling rate used to record the signal being analyzed. Plotting the amplitudes of the energy found at each frequency an x/y graph produces a visual representation of the spectral content of the original time-domain signal. A random (or pseudorandom) signal in which, over a given averaging period, each Octaveband (or other logarithmically spaced interval) contains an equal amount of energy. The decibel, often abbreviated as dB, is a logarithmic ratio between two values. In electronics and acoustics, decibels most commonly refer to the ratio between a given amplitude value and the number 1, where some reference value such as the maximum output of an A/D converter (dB Full Scale or dBFS) or the threshold of audibility for human hearing (for dB SPL) is scaled to equal 1. The decibel value for an amplitude is then calculated as: dB = 20·Log10(A) = 10·Log10(A 2 ), where A is linear amplitude. In this case, amplitude values greater than one yield positive decibel values and numbers smaller than one become negative dB values. This is why dB FS values are most often negative and dB values in sound level measurements are nearly always positive.