IMPORTANT NOTICE. As of December 2016, iSTI is no longer supported. We decided to focus completely on our Bedrock product line. While we understand that this is a disappointment to many loyal iSTI users, we have come to the conclusion that through our hardware-based product line we are capable of providing a professional, reliable and certifiable solution that simply cannot be achieved with mobile apps - at least not efficiently and cost-effectively. We apologize for any inconvenience.
iSTI Professional walkthrough
This is a screen-by-screen walkthrough of the app. This should answer most questions that may arise when exploring the possibilities of iSTI Professional. If you are interested in a specific step in setting up your measurement, or a specific view of the app, then you can jump ahead using the menu below.
iSTI at first glance
Shortly upon launching the app for the first time, your screen should look something like this.
Nothing much to see yet -- and only three things you can tap on: a button marked "signal," one marked "measure" and the ">" button on the top right. The button marked "store" (for saving measurement results) remains disabled until you have actually carried out a measurement.
As usual, the ">" button gives access to a settings menu. Anything you might want to configure in iSTI is accessible after selecting this menu.
The vertical scale on the left is a decibel scale, which is used for presenting input sound levels (A-weighted level and seven octave bands). The center part of the display is used for displaying the octave band spectrum of the current input signal. Upon initial startup, there is nothing there to see yet - no measurement is taking place. The label "STI" at the top indicates where the actual Speech Transmission Index will be displayed during measurement. The label "no project" indicates the position where iSTI will show the project name you have selected for your measurements; the '0' is the number of measurements saved sofar. The characters 'v4' indicate that iSTI is currently set up to measure according to the fourth edition (version) of the STI standard, which is the latest edition published in 2011. We will go in to all of this later.
Main measurement view
Let's assume that you are in a hurry to hear the STIPA test signal, and get started with your first measurement. We will skip all of the configuration and calibration options for the moment, although you may need to use these later on to get accurate results.
Tap on "signal." You will now hear the STIPA test signal through your internal loudspeaker, or perhaps through your headset (if you have one connected). Now press "measure." iSTI professional will attempt a first measurement. It may not be a very meaningful measurement: you will measure the speech transmission index that quantifies the channel that happens to be between your signal source. Unless you have a headset connected, this is the transmission path between your internal loudspeaker and your internal microphone.
A progress bar near the bottom will start moving, indicating that your first measurement is progressing. Your screen should look something like this:
The number above the graph is the Speech Transmission Index itself: a number between 0 and 1. In this example (which happens to be an iPhone 3GS), it seems that we are measuring an excellent transmission path between the internal loudspeaker and the internal microphone, featuring an STI of 0.90. However, the measurement had just been running a few seconds at the time of this snapshot. This can be seen from the position of the progress bar. The STI value is still fluctuating, and will stabilize as the measurement progresses. The measurement is only considered valid if completed, and the progressbar has been filled all the way to the right. This takes 20 seconds to happen.
The bar plot that is being displayed in this view represents the octave spectrum (125 Hz - 8 kHz) in blue bars, and the A-weighted sound pressure level in the red bar. These values will not be accurate until you calibrate them through the settings menu.
You may find different STI values if you repeat this example measurement for your own iOS device's internal transducers; this depends on your device and volume settings, but also on the noise levels around you at the time of measurement. Try interfering with the measurement by introducing some (stationary) noise; you will see the STI decrease, as it should under the influence of interfering noise.
Upon completion of the measurement, you may want to save the results. This is where the button marked "store" comes in. Or perhaps you will want to redo the measurement, in which case you simply press "restart." But before you begin storing measurement results, you should configure a project file in which to save your date. This can be done through the settings menu.
The settings menu
Let's tap the ">" button at the top right of the main measurement view. You will now enter the settings menu. The pictures below are snapshots of the settings menu (top and bottom half; you can reach the options in the 2nd snapshot by scrolling down).
The settings menu allows you to configure and calibrate iSTI Professional for optimal results. It is the starting point for the following actions:
- Define a new project in which to save your measurement data, or load an existing project to continue with.
- Export your measurement results (by email)
- Provide details about the current user (operator) of iSTI Professional
- Calibrate the sound level for the iOS device and microphone you are using
- Calibrate the sample rate of your STIPA test signal generator
- Select which edition of the STI standard (IEC-60268-16) to use
- Disable level dependency of the STI method
- Get information on the current version of iSTI Professional and access to the online support pages
We will go into all of these options below.
Configuring your project
Once you are in the settings menu, you can assign a name to your project after clicking on the ">" sign besides "Project settings." If you have not configured any projects yet, then you will see the following screen.
After tapping "New project" you will be asked to provide a name for your project - we choose "Test" as our project name for now. The project screen will now look like this:
The screen shows some basic information about your current project (which you just created). If you tap on the ">" besides "Browse stored projects," you will enter the project browser. It only shows your brand new project for now, but if you define multiple projects, the project browser will enable you to switch back and forth between different projects. You can also delete projects from the project browser (use the swipe gesture on the project name to reveal the "Delete" button).
Browsing measurement data
If you are following this walkthrough for the first time (to get a feel for the "flow" of the app), you may want to skip this section.
Once you have started a new project, and have stored measurements into it, you will want to browse through past measurements from time to time. From the settings menu, you can access your project through "Project Settings." Now tap on the ">" next to "Number of measurements" to see a list of measurements in your current project. Choose one to see a screen with all details of that particular project. In this example, we have not stored any measurements yet - but let us look at this feature anyway, pretending that we did store measurements before. After selecting a measurement from a list, you will see all details of that particular measurement.
For an explanation of these measurement details, please refer to the section on the detailed measurement view (below). This view is essentially the same - the difference being that you are now looking at historical measurement data rather than the current measurement. If you are interested in the STI value only: you find it in the middle of the bar at the top of your screen (0.86 in this example).
Export measurement data
You export your measurement data from iSTI Professional by simply emailing your current project to yourself. This option is available in the Settings menu as well as the project screen. All data in the current project will be emailed to the email address of the current operator, in comma-delimited (.CSV) format. This format is easily read by most spreadsheet software.
Providing user (operator) details
The option "user" in the settings menu can be applied to pick an operator name from your address book, or (alternatively) to provide details about the iSTI operator (you) manually. This feature can be used to distinguish measurements between different users, in case multiple people make use of the same iOS device for measurements with iSTI. The name of the operator is stored with each measurement, to improve traceability of you measurement data.
You can also assign an email address for the operator. This is vital information, since sending measurement (project) data by email is the easiest way to export your measurement data to work on outside of your iOS device. Unless you provide an email address, iSTI Professional will not be able to export data.
Calibrating the sound level
There are two parameters that can be calibrated: the overall sound level and the sample rate of your STIPA generator. We will first go into level calibration. From the settings menu, under the heading "Calibration," select sound level. If you have not previously saved any calibrations, this is what you will see:
Now select "New calibration." You will now see a screen marked "Sound level," featuring a button labeled "Calibrate." Tap on that button, and this is what you will see:
In order to calibrate your iOS device, you need to present a signal to your microphone for which the A-weighted sound pressure level is known. In practice, this requires that you have access to a Sound Pressure Level (SPL) meter or an external microphone suitable for calibration through a dedicated calibration device, such as a pistonphone.
Upon tapping "calibrate" iSTI will start acquiring the signal of your microphone. It will display an 1/1 octave spectrum similar to the view in the STI measuring screen. The number on the top right (and the vertical red bar) indicates the A-weighted sound pressure level according to iSTI's current calibration. Tap "calibrate" again to end data acquisition, or better yet: wait until iSTI finishes calibrating on its own after 16 seconds. Now input the A-weighted sound pressure level that was actually present at the microphone position. The current input source is shown at the right side of the horizontal red bar at the top; in this case, the iPhone microphone. Once the calibration is finished, the calibration level is shown on the left side of this bar.
Now tap on "sound level" (top left) to go back to the sound level calibration menu.
Each calibration is automatically assigned an identifier, so you can find it back if you need to re-use the same calibration later on. The default identifier consists of the current project name and the calibration level (in this example: Test: 80.0 dB(A)). You can edit this identifier by hand by tapping ">."
NB: when using external connections or microphones, please note that calibration depends on the microphone configuration that you are using. You will need to recalibrate every time you change this. Also, it is recomended that you verify your calibration settings regularly.
Calibrating the sample rate
If you are generating the STIPA test signal from the same device you are running your analysis on (pressing "signal" and "measure" on the same iOS device), you need not worry about sample rate calibration. However, it wil quite often be practical to generate your test signal from a separate iPod. You can also play back the signal from a CD or PC sound card, using the test signal that you can download from this site. If so, you need to consider the possibility that there may be a small mismatch in clock frequencies between your generator and your analyser. Differences in clock frequency translate into differences in audio sample rates that may have noticeable impact on your measuring accuracy.
In our experience, the majority of cases in which people have problems with (hardware or software) STIPA measurements, these problems lie in improperly adjusted clock sample rates. Other devices currently do not offer an option to calibrate this. Fortunately, iSTI does - based on a proprietary algorithm developed by Embedded Acoustics.
From the settings menu, under the heading "Calibration," select "Sample rate ratio." The screens are similar to the level calibration screens. Once the sample rate ratio calibration has ended, a dialog will appear that tells you the sample rate mismatch. You will be asked to provide a name to identify the device you have just calibrated.
After you have given your calibration a name, the screen looks as shown above. The horizontal red bar near the top indicates the sample rate deviation (left) and the name you gave this calibration on the right (in this example, "internal"). The deviation is expressed as ‰ (sample rate ratio x 10E-3). The margin of error of calibration algorithm is up to 0.2 ‰. Deviations greater than 0.5 ‰ require sample rate correction in order to obtain accurate STI measurements. If the calibration algorithm shows such values, then sample rate calibration was indeed necessary. If deviations above 10 ‰ are found (1 percent of the sample rate), then you need to consider finding a more accurate test signal source - or perhaps check if there was a problem with the test signal during calibration.
NB: a valid, IEC-compliant STIPA test signal must be received at the input of your iOS device in order to be able to carry out sample rate calibration.
Select which edition of the STI standard to use
The international standard which defines the Speech Transmission Index is IEC 60268-16. The first and second edition are now considered obsolete. Supported editions are 3rd edition (2003) and 4th edition (2011). The differences between both editions are minor; your choice depends on what version of the standard is required for your application. You can toggle between versions from the settings menu. The main measurement view will show a label to indicate which edition of IEC 60268-16 you are currently using ('v3' or 'v4').
Disabling sound level dependency
The STI depends on the absolute sound level, to a certain degree. If the sound level is very low (close to the hearing threshold), then intelligibility is reduced. If the sound level increases above comfortable listening levels, then intelligibility also decreases, through the inherent non-linearities of human hearing (flattening out of the auditory masking curves). The STI was designed to incorporate these effects. However, for some applications this level dependent behaviour of the STI is undesired or confusing. The transmission channel under test may, for instance, be a purely electric / electronic transmission channel; the acoustic levels that the channel produces may be unknown or varying. For such (rare) applications, the level dependent behaviour of iSTI may be disabled. Please be advised that when disabling level dependency, your measurement no longer complies to IEC standard requirements. Your resulting STI measurements will overestimate intelligibility at high or very low speech levels. Sound level dependency is enabled by default. Disabling it will result in a warning label (!level) being displayed in your main measurement view.
Measurement details view
We will assume that your project has been configured and your device is calibrated, we will return to the main view of the app.
We have named our project "test." This project name now appears in the top left of the screen. The number "0" on the top right indicates the number of stored measurements in this file. Every time you tap "store," the current measurement number is incremented by 1, giving each measurement a unique label. Once you are finished measuring, you can send your data home by email through the project browser. Using these measurement numbers, you will be able to match up the individual measurements with any notes and drawings you choose to make.
Perhaps you require a greater level of detail than just the STI value, A-weighted sound level and octave spectrum. At any time during or after a measurement, you can access a detailed view using the swipe gesture. This detailed view is positioned to the right of the main view.
A full explanation of how the data on this screen can be used to make use of the powerful diagnostic capabilities of the STI framework is beyond the scope of this manual. Suffice it to say that these data can be used to determine not just to what degree speech intelligibility is affected, but also through what causes: be it noise, reverberation or nonlinear distortion.
We will simply provide an overall legenda to the screen.
- Oct: Octave band center frequency
- Leq: Equivalent sound pressure level (per octave band)
- Rel: Octave band spectrum relative to the source signal spectrum (overall frequency transfer)
- Mf: Modulation frequency used in this octave band
- Mr: Raw modulation index
- Mc: Corrected modulation frequency (corrected for the effects of auditory masking and reception threshold)
- TI: Transmission Index (one value per octave band, per modulation frequency)
- MTI: Modulation Transmission Index (one value per octave band)
NB: since STIPA uses two modulation frequencies per octave band, there are two values for Mf, Mr ,Mc and TI per octave band.