transfer function

The Studio Six Digital Transfer function includes the most-used Transfer Function features in an easy-to-use GUI package that runs on iOS devices and Macs with M chips.

With Transfer Function, you can make changes to a sound system without having to run pink noise or swept sine waves through the system. This allows you to monitor system performance, or make adjustments, while the system is in use and the crowd is in the room.

Here’s a video of AUDIX demonstrating the use of their in-ear monitor test fixture with Transfer Function

Available in AudioTools on the App Store

Transfer Function is by definition a dual-channel measurement, and so to use it you need to either use a two-channel input device, or you will use the signal generator as the reference signal. iAudioInterface2 was designed specifically with Transfer Function in mind, since it includes a microphone input, and a second line input for the reference signal.

You can also use other two-channel interfaces, such as the original iAudioInterface or many of the generic 2-channel audio interfaces that work with iOS.

Using Transfer Function With A Single-Channel Input Device

Transfer Function can be used with a single-channel input device, such as the built-in iOS device microphone, iTestMic, or iPrecisionMic. In this case, you have to use the signal generator to make the reference signal.

To do this, you will need to use a cable that has an 1/8″ TRS 3-conductor plug on the end that plugs into the iOS device. If the plug has 4 conductors, the iOS device will turn off the built-in mic and look for a mic on that cable. The other end of the cable might be 2 RCA connectors, or it could be wired to an XLR or 1/4″ TRS plug. If you are using 2 RCA connectors, make sure you select “mono” for the generator output. If you are wiring to a balanced connector, like an XLR or 1/4″ TRS, make sure you select “balanced” for the generator output. Selecting the wrong mode will result in attenuated low frequency readings.

Next, make sure that the output cable is the last thing that you plug into the iOS device. Plug in the iTestMic or iPrecisionMic first, and then plug in the output cable.

Now open the generator panel by tapping the sine wave icon, select pink noise, and turn the generator on.

Next, start the Transfer Function running, using the Play/Pause button on the toolbar. You should start to get results at this time. Continue to the other steps described below, including setting the delay, to get accurate results.

Using Transfer Function With A Multi-Channel USB Audio Input Device

You can use Transfer Function with any USB audio device that the iOS device supports. For example, on the newer iPhones and most iPads, you can plug a USB audio interface in directly, although in many cases you will need to supply external power to the interface.

Once connected, you just need to assign the desired channels on the interface to the Measure Source and Reference Source in Transfer function. Typically the Measure Source will simply be a microphone that is set up in the room to pick up the sound system signal, and the Reference Source will be connected to the signal that is feeding the sound system. 

By default, AudioTools will use the first input (left) as the Measure Source and the next input (right) as the Measure Source. You can reassign the input channels if you need to. To use a channel other than channel one for the microphone, connect your multichannel interface, and navigate to Settings->Channel Mapper. From here, you can select the channels from your interface for both the microphone source and for the line input source.

Using Transfer Function With iAudioInterface2

iAudioInterface2 is designed to work as an input device for Transfer function. Here are two common use cases.

We are assuming that you will be using a microphone, so in both cases you would have the mic plugged into the XLR input, as the Measure source.

Using the Internal Pink Noise Generator

For pink noise, on the Transfer Function Setup page, select Ref Source, and on the top, select Mic/Gen. Then just run a cable from the Line Output / Headphones jack to your system input. Turn on or off the pink noise using the sine wave icon on the main screen. Note that pink noise will only run when either the Delay Finder is on, or when the measurement is running.

Using Program Material as the Reference

You may be in a situation, such as during a live sound event, when it is not possible to run pink noise. In this case, you can use the actual sound from the event as the reference signal.

To set up for this, select Mic / Line on the Reference source page, and run an output from the sound system that is a mono mix of the audio going out to the room into the Line Input jack. Probably Mid gain will work, just check the little M R meters on the main screen for the levels. Now set the delay, and you are all ready to measure the system frequency response and phase while the music is playing.

Using Transfer Function With SurroundPod

SurroundPod is a wireless reference-signal source that streams pink noise from your iOS device and outputs it through HDMI to your AVR or processor. The AVR decodes the stream and routes the signal to the channel you select in AudioTools, so you can run Transfer Function measurements on any channel of an immersive sound system – including height channels – without rerouting cables.

To set up, connect the SurroundPod’s HDMI output to an available HDMI input on the AVR and switch the AVR to that input. On the Transfer Function Setup page, select Ref Source, and on the top, select Mic/Gen so that pink noise is generated internally and streamed out through SurroundPod. Choose the channel you want to test from the channel selector in the SurroundPod controls.

Run the Delay Finder and tap Set as Baseline. Measured delays with SurroundPod are typically substantial – 500 ms is not uncommon – because the reference signal traverses Wi-Fi from the iOS device to the SurroundPod before reaching the AVR. This is normal and does not affect measurement accuracy. The Delay Finder captures and compensates for the full path, so the magnitude and phase results are correct as long as the network connection remains stable through the measurement.

If the coherence trace drops out or becomes erratic mid-measurement, suspect Wi-Fi congestion rather than an acoustic or system problem.

Using Transfer Function With HAA Link

HAA Link is the API-based integration with the AudioControl Hyperion 16-channel processor, available in HAA CinemaTools. With HAA Link, the Hyperion generates pink noise internally on the channel you select, and no audio is sent from the iOS device – only control messages. The measurement microphone connects to the iOS device input, or to an iAudioInterface2.

To set up, connect to the Hyperion from the app, select the channel you want to test in the signal generator panel, and start pink noise on the Hyperion. Run the Delay Finder and tap Set as Baseline. The measured delay includes the Hyperion’s output path and the acoustic path to the microphone.

HAA Link is particularly clean for tasks like time-aligning loudspeakers or aligning a subwoofer, because the signal routing and any DSP changes you make both live inside the same device. Adjustments take effect immediately and there is no wireless link in the reference path, so coherence is typically excellent.

Subwoofer Alignment with Transfer Function

One of the most valuable uses of Transfer Function in a cinema or home theater system is aligning a subwoofer in time to the main loudspeakers. When the sub and a main speaker both reproduce energy in the crossover region, any time offset between them will cause cancellations or peaks at and around the crossover frequency. Getting this right is the difference between a subwoofer that sounds tacked on and one that blends seamlessly with the mains.

The method described here is the industry-standard approach used by tools like Smaart: aligning the phase traces of the sub and the main speaker in the crossover region by adjusting the sub’s delay.

Why Phase Alignment Works

Magnitude response alone will not tell you if a sub is aligned in time with a main speaker. Two drivers can have a reasonable summed magnitude response and still be misaligned, because a delay offset of a fraction of a cycle at the crossover can still produce an acceptable-looking magnitude trace while causing deep cancellations at other mic positions or with small crossover changes.

Phase, by contrast, is very sensitive to time. Near the crossover frequency, both the main (through its high-pass) and the sub (through its low-pass) go through a rapid phase rotation – often steep enough that the trace sweeps from near the top of the phase plot to near the bottom in a narrow frequency band, wrapping from -180 to +180 in the process. Both traces have the same characteristic shape through this region.

If the sub and the main arrive at the mic at the same time, these two steep phase features sit at the same frequency and the traces overlay. If there is a delay offset between them, the two features have the same shape but are shifted left or right relative to each other on the frequency axis. Adjusting the sub’s delay slides its whole phase curve left or right on the plot, and the goal is to slide it until its steep feature lines up with the main’s.

The Workflow

Position the measurement microphone at the primary listening position on a stand. It must not move for the remainder of the procedure. In your processor, add a modest amount of delay (for example, 10 ms) to both the main and sub outputs before you start. This is arbitrary padding, but it gives you headroom to move the sub either earlier or later during alignment; subtract this amount from both at the end.

Route the reference signal to a single main speaker – typically the Center, or the Left channel if there is no center. Run the Delay Finder and tap Set as Baseline. Take a measurement of the main speaker with averaging on, let the traces settle, and save the plot using the plot storage controls. Recall the saved plot so it is displayed behind the live trace.

Do not re-run the Delay Finder, do not move the mic, and do not change the Transfer Function delay value. The delay you measured for the main speaker is now your fixed reference.

Mute the main speaker and route the same reference signal to the subwoofer instead. Identify the spectral crossover – the frequency range where the two magnitudes are within about 10 dB of each other. Pick a frequency inside that region with good coherence and read off the phase difference (ΔPhase) between the sub and the main. Adjust the sub’s delay in your processor (not the Transfer Function delay) to close the phase gap. The delay required, in ms, is approximately:

    delay (ms) = ΔPhase / 360 / frequency (Hz) * 1000

As you add or remove sub delay, you will see the live sub phase curve slide left or right on the plot. The target is for its steep feature to land on top of the main’s steep feature. Unmute the main, verify the summation in the crossover region looks smooth, and remember to subtract the padding delay from step one when you finish.

Half-Cycle Alternatives

Once you have a match at your chosen frequency, it is worth checking alternatives at half-cycle and whole-cycle offsets. If the alignment frequency is 80 Hz, one full cycle is 12.5 ms (1 / 80). Try the current setting, plus and minus 6.25 ms with the sub polarity inverted, and plus and minus 12.5 ms with polarity normal. Pick the variant that gives the smoothest summation and the best listening result. Sometimes a half-cycle offset plus a polarity flip delivers a better real-world outcome than the nominally-closer delay value.

Hardware Scenarios

There are three practical reference-source paths for sub alignment, each with different implications for where the reference comes from and how delay is handled. The diagram below shows all three.

SurroundPod (AudioTools) – Pink noise is streamed wirelessly from the iOS device through SurroundPod and out via HDMI to your AVR. The AVR decodes the stream and routes the signal to the channel selected in AudioTools. Because the reference signal travels over Wi-Fi, the measured delay is typically large but stable, and the Delay Finder handles it correctly.

HAA Link (HAA CinemaTools) – The Hyperion processor generates pink noise internally on the requested channel; only control messages are sent from the iOS device. There is no wireless link in the reference path. Particularly clean for sub alignment because routing and DSP changes both happen inside the same device.

iAudioInterface2 with internal loopback – The reference signal is generated inside the iAudioInterface2 and presented to Transfer Function via internal loopback, while the same signal is sent out the line output to an analog input on the processor. This is the most accurate electrical reference because the loopback happens inside the interface before any analog output, but it requires an available analog line input on your device under test.

Verifying the Result

After alignment, unmute the main speaker so both the main and the sub are playing pink noise. Switch to a magnitude display and look at the crossover region. You should see a smooth summation with no deep notch at the crossover frequency. A deep, narrow notch centered on the crossover is usually polarity – invert the sub and re-check. A notch slightly above or below the crossover usually means the delay is close but not exact – return to the phase view and refine. A smooth, broad bump or dip of a couple of dB is normal and is generally addressed with level trim or mild EQ, not delay.

Once the summation looks good, move the microphone to a second seat and verify the alignment still holds. If the crossover region summation falls apart dramatically at other seats, you may need to revisit sub placement or consider multi-sub optimization rather than a single delay value.

Transfer Function on iPad

Here is Transfer Function on iPad, showing a running trace (green and blue), and a recalled trace (yellow and magenta). The coherence plots for both measurements are shown, in red and dark red:

Included in Transfer Function are:

Magnitude Graph

The magnitude graph is used to show the difference between the reference signal and the test signal. Typically, the reference signal is pink noise, although you can also use mono program material (music) and still get good results.

The graph auto-centers around the 0 dB axis, so it’s easy to see how the test signal differs from the reference signal.

Phase Graph

The phase graph shows the relative phase of the signal, and is useful for setting crossover points.

Coherence Graph

The coherence graph is used as a measure of the reliability of a set of averaged transfer function measurements. 100% would mean that you are seeing highly correlated graphs, and thus the information on the magnitude and phase graphs would be accurate.

Coherence Blanking

It may be useful to only display transfer function plot data when the coherence is above a certain threshold. This helps you know that the information being shown is reliable. To do this, you can set a coherence blanking threshold in percent, on the setup screen.

When the blanking threshold is greater than 0, plot data will only be shown that has a corresponding coherence greater than the threshold.

Typical coherence blanking values are 20-50%.

Delay Finder

Transfer Function includes a sample-accurate, impulse-response based delay finder. This is a very powerful feature, and is used to set the delay for the most accurate Transfer Function measurements, and it can also be used as a standalone delay finder, to accurately measure the time arrival of a signal, for setting driver delays times.

Using Delay Finder to Set Up Transfer Function

Since it is critical that the time offset between the reference and test signal is known, there has to be a way to accurately compute the time delay between these signals. Our impulse response based delay finder does just this.

Just tap the delay value to bring up the Delay Finder panel. Select an FFT size, which sets the longest delay that you can measure, and tap Find. The delay will be computed, and shown on the screen. To average more than one FFT, enter a number larger than 1 in the Averages field. Averaging several FFTs can increase the accuracy of the results.

To calibrate the system to run the Transfer Function, place the microphone at the location that you will be testing, run the delay finder, and tap Set as Baseline. This is important, because the phase plot will not be correct if you have not time-aligned the output to the input.

Using Delay Finder to Time-Align Loudspeakers

To calibrate the system to compute relative delays, place the microphone at your zero distance. This may be the speaker grill, or the plane of the front of the driver that you are testing. Run the Delay Finder, and tap Set as Baseline. Now, you can move the microphone to a different location, run the delay finder again, and see the exact delay in time and distance. You can also enter the temperature of the air in the room, to get the most accurate distance calculations.

You can also change to another driver, and compare the delay for this driver to your reference driver. If you get negative delays, that means that the driver you are testing has less delay than the one you used to set the baseline. Just reset the baseline to zero out this driver, and go back to the other one and remeasure it.

Note on Using The Internal Generator

When you are using the internal generator, the first time that compute the delay you may notice a very large delay time and distance being computed. This is normal when using the internal generator, and the built-in mic or an external microphone. It is due to the iPhone audio latency between the iOS audio output and input that varies greatly, from device to device, and even from one run to the next. Follow the instructions above to correct this delay.

Setup

Select the FFT size and windowing, display type, turn Peak Tracking on or off. Select the SPL readout and decay speed.

The Measure Source and Reference Source are selected from this screen.

The Coherence Blanking percent will blank out parts of the magnitude and phase plot that have coherence lower than the value. The switch turns the Coherence plot on or off.

You can also engage Difference Plot Mode, control audio pass-through, turn on Difference mode.

The graph scales may be set manually, and you can lock the graph scale so that the touch-GUI is disabled.

The Phase plot may be smoothed, or you can leave it unsmoothed to see the full phase detail. You can also select whether or not you want to wrap the phase plot when it exceeds the plot bounds.

Using Transfer Function with iTunes

When you are using iAudioInterface2, it’s easy to use music as the test signal source for Transfer function.

To do this, you will have to run at 44.1kHz sample rate, because iTunes will not run at 48kHz, which is the sample rate that we

For the full user guide, Click Here