The Impulse Response module for AudioTools provides an easy way to capture an impulse response audio file on iOS devices, and also calculates the most-needed metrics from the data. In seconds, a complete set of measurements is made that describe the acoustic characteristics of a room.
Make sure to check out our Impulse Response demo videos located in the Support > Video Center section of AudioTools.
Impulse response analysis is a way of assessing the acoustical parameters of an enclosure, be it classroom, auditorium, concert hall, or place of worship. Impulse responses can also be used to characterize the frequency response of loudspeakers or how speakers behave in a room. By measuring the transfer function in a space and extracting time-based (temporal) or frequency-based (spectral) parameters, the acoustical engineer, architect, or recording engineer can understand how sound behaves and decays in a room.
This module is also equally effective for home studio, recording studio, home theatre and car audio use, where highly detailed, professional results are required.
The impulse response is measured by either exciting the room directly with an impulsive sound source (like a balloon pop or starter pistol) or by capturing how a reference signal (logarithmic sinusoidal chirp) excites the enclosure. Through a process of deconvolution, the known reference signal is subtracted from the recorded signal (in the frequency domain) within a space to extract the underlying transfer function. This is the same exact procedure that is used in professional sound level meters and laptop-based measurement systems.
Impulse Response will give you very good results even with the internal mic of your iOS device. Of course, you are limited to moderate SPL levels, and the lowest frequency bands will not be as accurate as when used with the iAudioInterface2, iTestMic, or iP{recionsMic, but Impulse Response will give you accurate reverb decay times and other measurements. You should also be aware that the frequency response recorded by the built-in microphone will be dependent on your specific device due to the diffractive effects of the individual hardware.
This section describes how to use the graphic user interface to operate Impulse Response.
Split-Screen -- (iPad only) Use the single/split control (lower left-hand corner 1 | 2 buttons) to select from a single screen or split screen. You can use the graph type buttons on the top and the bottom of the screen to select which graphs you would like to see in either window.
Gestures -- Use the pinch-zoom horizontally to move the time axis on the energy time curve (ETC) screen to select the part of the IR that you'd like to focus in on (editing features covered in detail in a later section). Using this gesture on the FFT screen will zoom the range of frequencies being displayed in this analysis. Use the pinch-zoom vertically to select an area on the ETC curve to use for local broadband reverberation calculation (RT60). A single finger drag will control the cursor on the screen to see detailed output of the currently selected graph or plot. The cursor can also be used to the set the left and right window time points (windowing covered in detail in a later section). If you have zoomed into the time axis horizontally on an impulse response viewing the ETC, you can double tap to return to the full response. Triple tapping on the ETC screen will zoom the time axis to correspond to the current window length.
You can record either an actual impulse, or a swept sine wave excitation signal (chirp) that the app will deconvolve into an impulse. If you plan on recording a swept sine wave, you can either play the signal directly from within the Impulse Response module or play the signal externally by downloading our excitation signal files, and burning them to a CD or by putting them on another iOS device or laptop.
To record your impulse response, tap the Sine Wave icon, which will bring up the recording screen. The first time that you access this screen, the reference swept sine signal files will be downloaded from our server and saved to your device. You can always re-download them by tapping the Setup (wrench) icon in the main screen and then tapping the Re-Download Chirps text. This could become necessary if we change the chirp files in the future. You will also see options for downloading a set of demo IR files, and to import an IR (this procedure will be covered in a later section).
To get started, select the type of excitation signal you wish to record: An impulse, 3, 7, or 14s chirp. Longer chirps will give you better resolution and a higher signal to noise ratio at the cost of longer calculation time. Current generation iOS devices are powerful enough for you to use any length chirp without much calculation time, so feel free to choose the chirp length that makes the most sense for your current measurement task.
Pick the expected reverberation time rounded UP to the next second. A good starting point for a large room or hall is 3 seconds, a small room, 2 second and a large church, 10 seconds. Longer expected reverberation time selections will result in longer processing times, and longer .WAV files increasing memory usage on your iOS device.
Tapping on the Test Tone text will emit a 400 Hz test tone from your device. This is useful for setting the level of your playback speaker in order to have the right amount of gain, avoiding clipping of your input microphone.
The Play chirp file while recording switch will toggle whether or not the measurement chirp will be emitted from the IR application or if you wish to use an external device for the playback of the excitation signal.
To record an impulse, just start recording, and create the impulse. Make sure that the Clip indicator does not appear. You can also listen to the impulse to check it by taping the Play button. When you are happy with the recorded impulse, tap Done to return to the calculation screen.
When you are ready to start recording, tap the Record button on the Recording screen, and then start the chirp playing. You will hear a short sync tone, followed a period of silence, followed by the chirp. Do not click stop until the minimum time displayed on the screen is reached. Watch the level meter as the chirp runs.
Check that Clip does not appear on the screen, and that the chirp got more than half-way up the level meter. The yellow area is perfectly fine. If you like, you can play back the chirp and listen to it. This will let you know if there is too much background noise, any distortion in the recording or any impulsive noises that could affect the subsequent calculations.
At the end of the recording, a message box will appear asking whether you want to deconvolve the chirp, or cancel. If you are happy with the recording, choose deconvolve. A series of calculations will begin, that make take up to a minute, depending on the speed of your device. The new models are much faster.
If all goes well, you will see the word "Done" in green, and you can tap Done on the recording screen and go on to the calculation screen. If there is a problem, it will be shown on the Record IR screen in red text. The most likely problem is that the sync pulse may not be found. The sync pulse must occur at least one second after starting the recording, and must be within 20dB from the loudest part of the chirp. Also, the recording must be long enough to include at least more than the “expected reverberation time” selection in silence or the process will fail. Normally, just waiting for the minimum amount of time displayed to the right of the running clock will avoid this problem.
When you go on to the calculation screen, the energy time curve (ETC) will be computed and shown on the screen. From here you can look at other parameters and metrics, or trim (edit) the IR.
You may want to start with a 3-second chirp as you get your setup working properly, and then switch to a longer file.
Setup your playback system with the CD or iPod playlist that you’ve created with the measurement signal that has been downloaded from our website. Connect to a system with only one speaker active, and try playing the chirp. Listen for obvious distortion or other problems. Then record the signal in the IR module of AudioTools.
For the best frequency resolution, use a dedicated subwoofer and dodecahedron loudspeaker specifically made for acoustical measurement if available. In general, best results are found when using a loudspeaker that has the flattest frequency response from 20 Hz - 20 kHz and is able to reproduce high SPLs without distortion.
The internal microphone in iOS devices are very sensitive and distorts easily even before the digital audio convertor. You will get the best results if your peak dBFS in the recording screen does not go above -40 dB. With a dedicated measurement microphone, you will be able to record much higher levels without distortion.
If you are getting unusual results, or audible distortion in your impulse responses, recording at a slightly lower level is usually a good first step to improving the quality of your recorded IRs.
If you are using your own microphone to record the measurements, that microphone should be omni-directional and have a flat frequency response from 20 Hz - 20 kHz. Impulse Response does not utilize the microphone calibration files, so uneven microphone frequency responses cannot be corrected with a text file. This process does work with the deconvolution algorithm utilized in Impulse Response.
Back to topOften you can get better results by trimming the start and end of an impulse response. This also lets you focus on just the reverberant decay in the room and not the noise floor before and after the impulse. If you have recorded an actual impulse, such as the popping of a balloon, it is highly recommended that you trim the result. If you have recorded a sine chirp signal, sometimes the resulting IR will need some trimming before the impulse, and the noise after the decay curve can be trimmed to shorten the screen window for a better look at the data.
When running a swept sine on an external device, if the sample rate of the playback device is slightly different than the sample rate of the device that the app is running on, even by just a few Hz, you may see a short, false peak, or ramp-up before the actual impulse. If the sample rate of your devices cannot be matched exactly, and you still wish to use the playback device, it is important to trim this false peak for the best analysis results.
To edit an IR, go to the ETC screen, and use the pinch/zoom gesture to expand the display. Your goal is to remove extraneous sound before the actual impulse begins, without cutting into the impulse itself, and to remove noise after the decay ends, without cutting into the decay curve itself.
When you have decided on your trim, tap the Edit button, which will bring up a dialog box giving you the option to Trim the IR File to Screen. This will create a new IR that has been destructively trimmed to your selection. You can then save this new impulse response by tapping the Save/Recall folder icon. For additional editing of the impulse response, fading in and out from the noise floor smoothly, or just focusing on a portion of the response, the windowing functions can be used. This is also known at gating the impulse response, and has advantages for certain applications. This will be explained in a later section.
Back to topRT60 -- You have the ability to measure the broadband reverberation time (RT60) directly on the ETC screen. To do this, perform a two-finger vertical pinch gesture on the decay curve between the low and high limits that you wish to calculate the decay rate. Be careful to avoid the noise floor, as this will artificially lengthen your calculated reverberation time. Note that this is a local calculation and does not affect the calculation of other measures of reverberation time in Impulse Response.
Early Reflections -- The Early Reflection screen allows you to zoom in (in the time domain) to the early part of the IR to determine the magnitude and delay of individual reflections within a selected time window. Time windows of 25ms - 400ms can be chosen. Scrub through an IR with your cursor to see the delay in both ms and distance, as well as the individual energy level of various reflections in your response. This analysis is particularly useful in seeing where an audible echo occurs in both time and distance.
After editing your impulse response and removing extraneous noise, you can then select one of these calculations to display in a bar chart, either by octave band or 1/3 octave band. Calculations are done as needed when you change graph selections.
T30 (RT60) -- T30 is the decay rate of the impulse response or reverberation time (RT60), the time that is takes for a sound to decay by 60 dB in a space. By using a noise compensation technique to avoid inaccurate representations of reverberation time, the decay rate is calculated by determining the slope of the decay function from -5 dB to -35 dB, and extrapolating the time to a decay of 60 dB.
T20 (RT60) -- T20 is the decay rate of the impulse response or reverberation time (RT60), the time that is takes for a sound to decay by 60 dB in a space. The decay rate is calculated by determining the slope of the decay function from -5 dB to -25 dB, and extrapolating the time to a decay of 60 dB.
C50 -- The clarity parameter (50ms), expressed in decibels, is the ratio of the early energy (0-50 ms after the direct sound) to the late reverberant energy (50-end of the decay of the IR). C50 is often correlated with speech intelligibility in a enclosure. Ideally C50 calculations should be from IRs that are measured with a loudspeaker that has the directional characteristics of the human voice.
C80 -- The clarity parameter (80ms), expressed in decibels, is the ratio of the early energy (0-80 ms after the direct sound) to the late reverberant energy (80-end of the decay of the IR). C80 is often correlated with music perception within an enclosure.
EDT (RT60) -- Early decay time represents the reverberation time (RT60) of the IR based on the slope of the early part of the energy decay curve. It is the slope of the curve limited from 0 dB to -10 dB extrapolated to a decay of 60 dB below the stopping of the direct sound energy. EDT is typically associated with the perceived reverberance in a room.
CT -- Center time or Ts is the time after the onset of the direct sound to the time where half of the energy has decayed in the IR. It is the time corresponding to the "balancing point" or center of gravity of the squared IR. Center time is correlated with reverberation time, so center time increases as a function of reverberation time.
Definition -- Similar to the C50 metric, definition represents the ratio of sound arriving in the first 50 ms of the IR compared to the rest of the IR. It is expressed as a percentage and correlates to speech perception in a room. For rooms that primarily are to be optimized for speech, D50 should be > 50 percent.
Signal/Noise -- Represents the difference between the measured peak signal captured by Impulse Response versus the noise floor present in the room under test. It is a good tool to verify that the measurement signal that is being used will yield quality results. Aim for a signal/noise ratio above 50 dB whenever possible or the reliability of the derived metrics of the IR will be reduced, especially reverberation time. If the signal/noise ratio is too low, some calculations will not be available. Improve the measurement setup by increasing the level of the playback chirp (without distorting the signal), increase the gain to the recording microphone (without distorting or clipping the input), or reduce background noise in the space (turn off HVAC systems if possible, close doors, or record when the space is unoccupied and is the quietest). The chirp technique is also very sensitive to impulsive noises created during the recording process (like footsteps, claps, or doors shutting). If these sounds occur during the recording process, it is possible to have a good signal/noise ratio impulse response that gives poor quality or inaccurate results. If such impulsive sounds occur during the recording process, stop the recording and start the chirp again before deconvolving and analyzing the response.
Schroeder Plot -- The backwards integration of the decay function of the impulse response. Broadband reverberation times are also calculated across the energy ranges of 0 dB to -10 dB (Early Decay Time, EDT), -5 dB to -15 dB (T10), -5 dB to -25 dB (T20) and -5 dB to -35 dB (T30). The corresponding linear regressions are plotted on the Schroeder plot.
ETC -- This is the time-domain plot of the decay response of the IR. From the ETC, you can measure RT60 (described in the previous section), set left and right window points (explained in a later section) or trim the response to a horizontally zoomed section of the full recording.
FFT -- The frequency response of the impulse response results in the frequency response graph for the signal. You can select anywhere from no smoothing to octave smoothing with choices of 1/1, 1/3, 1/6, 1/12, and 1/24th octave smoothing. In the FFT screen, you can use a single finger drag to move the dB axis up and down without zooming. A two-finger pinch gesture in the X or Y direction will zoom in and out the frequency and dB scales, respectively. Double tapping the screen at any time will fit the entire graph on the screen. In the Setup screen, there is also the option to set the Graph Scales manually. The dB Scale fields will let you set the minimum and maximum dB values. The Frequency Scale fields will let you set the minimum and maximum frequencies displayed on the graph. The Lock Scales switch will lock these ranges.
Also, in the Setup screen, there is a switch to turn on Sum FFTs. When activated this will sum the dB values of the two FFTs arithmetically (See Multiple IR Mode section below for working with two IRs). This will generate a third FFT trace, which will be drawn in green.
Reference curves are also available from within IR. Reference curves let you use a reference frequency response function created outside of IR for comparing to test measurements made in this app. This is available as an in-app purchase, which then shows up in all AudioTools modules that support this functionality. Complete documentation of Reference Curves are available when the Reference Curves text is tapped. One good use of the reference curve is to see if you IR measurement falls within the range of a target response (or house curve), as show in the figure below.
Waterfall -- The waterfall analysis is a way of looking at energy over frequency and time in an impulse response. The analysis works by taking a small slice of the impulse response starting at the beginning of the decay, windowing it in the time domain, and then plotting the FFT of the slice. Then the window is moved in time and another slice is analyzed and plotted. This is repeated a number of times and creates a 3D graph, which helps to look at how energy present at the beginning of an impulse response decays over time. The 3D plot can be rotated by using a single finger gesture to interact with the plot.
Options for the waterfall analysis can be adjusted in the Setup screen in the 3d Waterfall Settings.
* Number of Plots determines the number of slices for the graph. This affects how many total slices are taken in time of the impulse response. More slices means greater resolution over time at the expense of processing time and speed at which the graph can be rotated. This is reflected in the time resolution. The time for each slice equals the total graph time, divided by the number of slices
* Smoothing adjusts how the frequency response is smoothed within each slice. Options for full octave, 1/3, 1/6 and 1/12th octave are available.
* dB Scale adjusts the lower and upper limits for amplitude on the plot.
* Time Range adjusts the time before and after the direct sound onset for use in analysis. The slices work by taking the total time before the direct sound + the time after the direct sound. The first field is actually negative (because it comes before the direct sound), so it is actually -time before + time after. The time per slice is the total time / number of slices. Then for the FFT, the overlap time is added (which can be set by the slider). So, if there is a slice time of 100 ms (4800 samples), that would be multiplied by 1.0 + overlap, or FFT time = total time * (1.0 + overlap / 100), since the slider is shown as a percent.
* The Lines / Surface control switches the way the plot is displayed. With lines, each slice is shown as a line graph. With surface selected, the plot is shown as a filled surface, which connects the slices.
* The Low Hz / Full Hz frequency range buttons adjust the range for the frequency axis used for analysis and plotting on the waterfall. When Low Hz is selected, the frequency range is from 20 Hz - 200 Hz. This is especially useful for investigating room modes that may be present in a room, which are excited in a measurement. Full Hz shows the response of the impulse across the full range from 20 Hz to 20 kHz. The screenshot above shows the Full Hz waterfall with a surface selected. The screenshot below shows the Low Hz waterfall with the line plot option selected.
* Overlap controls how much the FFTs overlap in the calculations. An overlap of 0% means that each slice starts at the point that the previous slice ends. As the overlap increases, each slices takes some data from the previous and next slice. This makes for a smoother decay plot, where the room modes will be more easily seen.
The Frequency Resolution and Time Resolution fields will update as you adjust the parameters that affect them, which are the Number of Plots, Time Range, and Overlap.
Impulse Screen -- You can see the full .wav file of the recorded IR and play it back by tapping the Play button.
T30/EDT/C50/C80/CT Mid Values -- This is the average value for the reverberation time, clarity and center time in the measured room taken at the 500 Hz and 1 kHz octave bands. T30 at 1 kHz is also reported.
SN Average -- The overall signal/noise ratio, expressed in dB of the measured IR.
Bass Ratio -- The metric that correlates to the “warmth” of a room. It is calculated by adding the T30 values for the 125 Hz and 250 Hz octave-bands and dividing them by the sum of the T30 values for the 500 Hz and 1 kHz octave bands. For rooms that primarily are concerned with music, a BR from approximately 1.0 to 1.3 is desirable. For speech rooms, a BR from 0.9 to 1.0 is preferred.
Back to topYou can save measured results using the Save/Recall screen that is common to all modules in AudioTools. The Save/Recall feature is accessed by tapping the Folder icon on the bottom of the screen. Enter the name that you wish to give the file. Fully detailed Save/Recall instructions are available by tapping on the i icon in the lower right hand corner of the screen.
You can save the analyzed data for all screens in IR as well as the .wav audio files of your measurements. Analysis results are saved to a .XLS file, optionally storing GPS coordinates, a photo if applicable, and a notepad. Saved data can be synced to a computer over WiFi, by using the Utilities > Files function, iTunes sharing or through Dropbox automatically.
When
recalled, the loaded measurement can be analyzed just as if you've
recorded a new IR. You can import a .wav impulse response from another
system or software (must be a mono 16-bit 48 kHz .wav file) by going to
Utilities > Files and importing the files from your Web browser. In Impulse Response tapping on the Import IR text in the Setup
menu will bring up a browser where you can select the file that you
downloaded in the Files module. This IR will now show up in the Save /
Recall screen for analysis. You can also use Dropbox or iCloud to accomplish this
task. You can setup up Dropbox or iCloud sync in Setting > General.
To import .wav IRs to be analyzed in IR, place the audio file in the import folder in your Dropbox or iCloud directory for AudioTools on your computer. If this folder isn't there, create it. Then you can tap the Dropbox icon in the Import IR screen, or just select the file if you are using iCloud. Your IR file name will show up. Tap it, and then tap the Import IR text on this screen.
Back to topYou can use IR to compare and contrast a recorded impulse response to a stored impulse response using the dual IR mode. In order to access dual IR mode, record an IR normally and then load an IR from memory using the Save/Recall screen. The current IR will be shown in yellow whereas the comparison IR will be shown in orange.
There are two primary ways of reviewing double IRs and what you see is dependent on the specific analysis screen you are looking at. The "impulse", "ETC", "Schroeder", and "Early Reflections" screens will allow you to toggle between two IRs by tapping on the IR 1 or IR 2 buttons on the menu bar. The "T30", "C50", "C80", "Definition", "CT", "Signal/Noise", and "FFT" screens will display both IR's data at the same time. Additional, in the FFT screen, you can choose to Sum the FFTs, by selecting Sum FFTs in the Setup menu. You may choose how you want the two IRs to interact by selecting from the picker on the menu screen. Selecting Both IRs will display alternating bars for IR 1 and IR 2 in yellow and orange colors respectively. Average will take the mean values from each octave or third-octave band from IR 1 and IR 2. IR 1 - IR 2 will take the different of IR 2 from IR 1, and IR 2 - IR 1 will take the difference of IR 1 from IR 2. The FFT analysis will display the frequency response of IR 1 in yellow and IR 2 in orange.
You may clear the screen of both IRs at any point by going into the Save/Recall menu and tapping the Clear Recalled Graphs text. This is how IR looks when in dual IR mode:
On the ETC screen, you can put windows on the left and right side of the impulse response using the windowing functions. This lets you analyze a gated impulse response. This is useful when you are measuring a loudspeaker, for example, and just want to get the frequency response of the speaker alone, and not the impact of any reflections. This is impractical in most indoor spaces, as the physical distance between the direct sound and floor or ceiling first reflection is too short to extract much meaningful information from the IR. The frequency resolution you are able to use with the technique corresponds to 1/length of the window in ms. Therefore, shorter windows give you much less resolution in low frequencies. In order to be able to get a quasi-anechoic analysis of a loudspeaker it is often advisable to do outdoor gated measurements (unless you have access to an anechoic chamber). Bring you loudspeaker outside away from reflecting surfaces and elevate it and the microphone as high as practical (about 6 feet is good), with the microphone about 3.5 feet away from the transducer. If available, place absorbing materials on the ground between the speaker and microphone to attenuate the first reflection. Then record the IR as normal.
In order to just look at the response of the loudspeaker, we will gate the response using a window before the onset of the first reflection. Windows are accessed by tapping the Setup icon. When turned on, the data is windowed in the time domain before being processed in IR.
Windows are used as a time-weight function, focusing on specific parts of the IR, and to reduce spectral splatter cause by sudden and sharp on- and off-set ramps. These windows variably weight the time before the IR's peak and the time after the peak to be zero and the middle of the window (corresponding to the peak of the IR) to be weighted towards unity gain. For more general information, read about windows for spectral analysis. There are multiple options for windowing: Rectangular, Hann, Hamming, Tukey 10ms, Tukey 0.25 and Blackman. Rectangular is the default used. For detail about the algorithms used, read more about Rectangular, Hann, Hamming and Blackman windows. The two Tukey windows are what is most often used for a gated IR analysis, and correspond to adding Cosine^2 ramps to the beginning and end of the windows for either the first 10ms (in the case of the Tukey 10ms), or 25 percent of the duration of the window. Since the left and right windows are set independently in IR, each of the window selections is actually a half-window. Selecting the same window type for both left and right will yield a full window.
The onset of the direct sound is the 0 ms point for the widows as a default. If you want to window a different section of the IR, you can change the Reference Time in the Setup menu. This will let you look at the frequency response of a single reflection, for example. You can either set the time for the left and right widows manually from the setup menu, or use the vertical cursor to set the window. The default left window is 10 ms before the peak of the IR and the right window is default to your selection of Expected Maximum Decay Time in the Record IR screen. To use the vertical cursor to set the window, simply scroll to the point in the IR where you would like the window to start, enter the setup screen and tap Set From Cursor to set the window. You can also turn on or off the window globally from this setup menu. In the ETC screen, you can see the envelope of the window function you've created by viewing the blue trace superimposed on top of the ETC. This corresponds to the amplitude function that you have chosen for your window in the setup menu.
Note:When using windowing functions on an iPad in split screen mode, the curse and window functions will follow whatever is on the lower pane. So if you want to set windows for IR1, turn the lower pane to the ETC view, set the windows. Then switch to IR2, set the widows in the same way. The Set From Cursor in the Setup menu will follow the cursor of the Impulse, ETC or Early Reflections views, and will always give priority to the cursor on the lower pane of the screen, when using dual view mode.
This section describes in a bit more detail how you would use IR for making gated impulse response measurements of loudspeakers. When building speakers or testing out a new speaker you have in your recording studio, home theater or listening room, you often want to be able to just measure the frequency response of the loudspeaker without the impact of your listening room. The way that this is done by speaker manufacturers is in an anechoic chamber. The anechoic chamber has no reflections so the engineer can just measure the direct sound from the loudspeaker and extract the frequency response. Most individuals do not have access to an anechoic chamber, but we can get close by using gated outdoor measurements.
The figure above shows the way that you would set up your loudspeaker and microphone to perform a gated outdoor measurement.
*Bring your measurement gear outdoors as far as possible from any
reflecting surfaces such as walls, buildings or trees. An open backyard
or field is ideal.
*Elevate your loudspeaker about 6 feet off the ground. The goal is to
record the direct sound and have the floor reflection delayed in time
so we can window it out. The longer the delay between the direct sound
and first reflection, the lower the frequency we can reliably measure
with this technique.
*Place you microphone about 3.5 feet from the loudspeaker on the same plane as the midrange driver.
*If you have absorbing materials available, place them on the ground
between the loudspeaker and microphone to attenuate the first
reflection.
*Record the IR as normal, using the techniques described in the previous sections of this manual.
Above is an image showing this measurement setup. The way to elevate the speaker doesn't have to be pretty, and an ironing board works well for this task. This technique works well for midrange and full-range loudspeakers. Woofers and subwoofers are impractical to measure this way as there is rarely enough of a delay between the direct sound and first reflection to have sufficient resolution in low frequencies. Subwoofers are often measured near-field with a microphone close to the loudspeaker. This way the level of the reflections is very low compared to the direct sound.
Back in AudioTools, window out the direct sound by using the ETC Windowing functions in the Setup screen. The ideal window set in Impulse Response is to have the ETC on the top of the screen and the FFT window on the bottom of the screen on split-screen mode on the iPad. use the horizontal cursor to scroll to a short period before the direct sound (about 5-10 ms). Then, in the Setup screen, tap the Set From Cursor text for the Left Window. Repeat with the horizontal cursor set to just before the first reflection (you should be able to see it easily if you zoom into the response, it will be the first peak after the direct sound). Use this second time to set the Right Window. As a good starting point, use the Tukey 0.25 window type for both left and right. The typical selections are shown below.
Turn on the Enable Windowing switch to gate your measured IR. Below shows the resulting window and frequency analysis for a midrange speaker on a flat open baffle. The first reflection arrives 6 ms after the direct sound, so we've set the right window to just before 6 ms. The blue trace shows the amplitude function of the window. At 5.6 ms the response has been completely faded out, allowing us to just analyze the frequency response of the direct sound, free of any reflections. Again, this technique is not without compromise. Because we only have a few ms of audio to analyze, the FFT block size is about 156 Hz. This large bin size limits the ability for this technique to be used on low frequencies reliably. We can use this to analyze the midrange performance of this speaker, see the presence of a dip at about 8 kHz, likely due to a cone breakup and a peak between 8 and 16 kHz where the whizzer cone takes over.
All of the details and best practices for this measurement technique is beyond the scope of this manual. For further information, you may want to read this article: http://www.brentbutterworth.com/PDFs/Speaker-Measurements-101.pdf, which describes in greater detail several different loudspeaker measurement techniques and caveats.
Back to topVisit our downloads page on the Studio Six Digital website to find our IR chirp test signals. They are archived in a .zip file called IR Signals. The files include two 3s, 7s, and 14s chirps. Each chirp starts with a sync pulse 1 second after the file starts, a period of silence, the chirp, and another period of silence. When you are recording the chirp, make sure to start recording first, then start the chirp file playing, and let the entire file play before stopping the recording.
There are two versions of the sync files. Use the "syncxxx10.wav" if you need to use the 10-second expected maximum decay, otherwise use the "syncxxx.wav" files. The files that have a 10 at the end of the name include a sync pulse 1 second after the file starts, 10 seconds of silence, the chirp, and another 10 seconds of silence. The standard chirps have 5 seconds of silence at the beginning and end of the audio file. Please do not use the 10 second files if anything other than a 10 second Expected Maximum Reverb Decay Time is selected in the Record IR screen, as the deconvolution will fail. Similarly, you must use the 10 second files if the 10 second Expected Maximum Reverb Time is selected, otherwise the deconvolution will fail.
Back to top