Multi-Sub Optimizer Reference Manual (page 3)

Taking Measurements With Your Measurement Software

First Clear Out DSP Filters, Delays, Gains and Polarity Inversions

It's important to realize that the filter parameters, polarity inversions and delay and gain values used by MSO represent changes relative to the conditions at which the measurements were performed. If MSO recommends a delay of e.g. 7.4 msec for a particular sub, and the measurements were performed with that sub's delay set to 2 msec, the correct final delay would be 9.4 msec, not 7.4 msec. It should be clear that performing the measurements with this sub's delay set to a non-zero value has caused at least two problems. First, the non-zero initial delay requires tedious and error-prone manual calculation of the final delay value. Second, it makes the task of preventing MSO from exceeding the hardware limits of your DSP for e.g. maximum delay more difficult. Assume for this example that a non-HD miniDSP 2x4 device (which has a maximum delay of 7.5 msec) is being used. Setting the maximum allowable delay of the applicable MSO delay block to 7.5 msec would be incorrect for this example, as that would allow for a final delay of up to 9.5 msec in the DSP device. The maximum allowable delay that you would specify for the delay block in MSO for this case would need to be 5.5 msec, as the delay was 2 msec when the measurement was performed. To avoid such problems, it's much easier to simply zero out all DSP delay and gain/attenuation parameters prior to performing the measurements.

A similar, but even worse situation can occur if the measurements were performed with DSP filters set to a non-flat condition. In the case of miniDSP hardware, exporting a biquad text file from MSO and importing it into a DSP channel that had some filters enabled during the measurements will cause those filters to be replaced, not appended to. This guarantees an incorrect result, not just an inconvenience as with the above delay example.

Subs+Mains Configurations: AVR Main Speaker and Sub Distances Before Measurement

When using sub-only configurations in MSO and considering sub delays, only the relative delays between subs matter with regard to the integration of the subs with one another. But when using subs+mains configurations, delay differences between main speakers and subs become important. This means special consideration must be given to both main speaker and sub distance settings in your AVR. MSO does not allow delay blocks in its main speaker channels. This means MSO cannot be used to determine AVR main speaker distances. Therefore AVR main speaker distances must be established prior to performing measurements for MSO.

The best way to establish main speaker distances in your AVR is to run its room correction software. Room correction systems use impulse response measurements to set main speaker and sub distances. This works quite well for main speakers, as the impulse response of a full-range speaker is narrow and sharp, such that delay (and therefore distance) can be easily and accurately measured by determining the peak of the impulse response. This delay calculation generally works poorly for determining subwoofer distances though, as a typical subwoofer impulse response is very spread out in time due to the limited bandwidth of subs. For best results integrating subs and mains using the sub distance adjustment, frequency-domain techniques must be used, as is done by systems such as MSO and Dirac, or by manual measurement and adjustment techniques such as the "sub distance tweak".

When relating speaker distances to delays, it's important to consider the following.

It follows that setting a speaker or sub distance to zero maximizes its delay relative to other speakers whose distance in the AVR is not set to zero. Internal to an AVR, the speaker or sub with the largest distance gets a delay of zero, while delays of all other speakers or subs are set relative to the speaker or sub with the largest distance (thanks to Roger Dressler for this observation).

Prior to performing measurements for an MSO subs+mains configuration, the following steps should be performed.

  1. Run your AVR's room correction program.
  2. Take note of the speaker distances it computes.
  3. Turn off your AVR's room correction.
  4. If turning off room correction resets speaker distances, set them back to the values determined by the room correction system.
  5. Set the AVR sub distance for your measurements as follows:
    • If the sub distance calculated by your AVR's room correction software is too close to zero or to its maximum distance, set the sub distance to be roughly equal to the Left, Right, or Center speaker distance.
    • Otherwise, use the sub distance calculated by your AVR's room correction software.

See Using Normalize Delays to Constrain AVR Sub Distance With Subs+Mains Configurations for information about constraining the shared sub delay in MSO. This will ensure that MSO's calculated sub distance will be within your AVR's sub distance limits.

Imported Measurements Must Be Time-Synchronized

MSO requires that the measurements you import be time-synchronized with one another. This means that the actual relative delays between all loudspeakers measured at a given listening position, including mains and subs, will be preserved in the measurements. While it might appear that this would not be a problem and would be accomplished automatically by measurement software, this feature is not automatic and special attention is required in order to ensure time synchronization. The method for ensuring that relative delays are preserved involves the use of a timing reference. When using a microphone with an XLR output (an analog output), the technique used is called a "loopback timing reference". When using a microphone with USB output (a digital output), the technique used is called an "acoustic timing reference" in the case of Room EQ Wizard. Accomplishing time synchronization of measurements is described in detail in connection with specific software in the sections below.

Caution: Measurements Taken With USB Microphones Require An Acoustic Timing Reference

Time-synchronized measurements with a USB microphone can only be done using Room EQ Wizard version 5.15 beta 6 or greater, with its "acoustic timing reference" feature. For all other known measurement software, time-synchronized measurements with a USB microphone are not possible.

Those familiar with HolmImpulse might be tempted to try HolmImpulse's "time locking" feature to work around this problem, but that workaround fails. The reason is described in the HolmImpulse section below.

Caution: Making Time-Synchronized Measurements With HDMI Requires Special Precautions

Special care must be taken when making time-synchronized measurements using HDMI. This is described further in the HDMI measurement section below.

Room EQ Wizard (REW) Considerations

REW is a popular choice for measurement software, and is recommended as the first choice for use with MSO on all platforms.

Using a USB Microphone: Acoustic Timing Reference

When using a USB microphone with REW, time-synchronized measurements must be performed using REW's "acoustic timing reference" feature. When using an acoustic timing reference, the same speaker must be used as a timing reference for all measurements. This speaker must be one with a high-frequency driver that can properly reproduce a sweep whose lowest frequency is 5 kHz. A subwoofer cannot be used for this purpose.

Using a Timing Reference With an XLR Microphone (Uncommon)

To obtain time-synchronized measurements with REW and an XLR microphone such as the Parts Express EMM-6, you must use a separate sound device such as an external USB sound device with a microphone input. Measurements may be performed using either a loopback (hardware) timing reference if your hardware supports it, or an acoustic timing reference.

XLR-to-USB Adapters

When using an XLR microphone in conjunction with a so-called "XLR to USB adapter", this "adapter" is really an external USB sound device with no analog input available for a hardware loopback connection. Therefore such a combination should be considered as the equivalent of a USB microphone, and can only be used in conjunction with an acoustic timing reference to achieve time-synchronized measurements.

General REW Timing Reference Considerations

For instructions for how to measure with a timing reference, see the REW documentation on the topic. Notice especially the following statement about subwoofer delay measurements:

"For speakers the delay estimate is based on the location of the peak of the impulse response. Subwoofers have a broad peak and a delayed response due to their limited bandwidth so the delay is instead measured relative to the start of the impulse response. The start of the impulse response cannot be located as precisely as the peak, however, so delay values are less accurate for subwoofer measurements."

Do Not Shift Any Impulse Responses in REW

MSO requires that the actual relative delays between all loudspeakers measured at a given listening position, including mains and subs, be preserved in the measurements. Even if a timing reference is used in the REW measurements, certain manipulations you can do in REW will shift the impulse responses and break these relationships. These include the following:

If you perform either of these operations before exporting the data, you must redo all the measurements. If you wish to do such experiments in REW, it's best to make a copy of your .mdat file and do the operations on the copy.

Smoothing in REW Should be Disabled

For best accuracy of MSO's predicted response, the measurements exported from REW and imported to MSO must be unsmoothed. This is normally accomplished by choosing Remove Smoothing from the REW Graph menu when REW is set to display All SPL. However, under certain conditions, REW can still apply smoothing to frequency response measurements even after you choose this menu option. This can happen if Allow 96 PPO log spacing was enabled in the Analysis tab of the Preferences dialog at the time the measurement was performed, or at the time the impulse response measurement window was last applied. If this is the case, choosing Graph, Remove Smoothing will not remove all the smoothing, as the 96 PPO log spacing will remain, and is implicitly associated with 1/48 octave smoothing by REW.

To completely clear out any smoothing effect caused by having had Allow 96 PPO log spacing enabled when the measurements were done or when the impulse response measurement window was last applied, perform the following steps:

The above steps need only be done once, as Apply Windows To All, Keep Ref Time applies the window to all measurements simultaneously. Then smoothing can be removed by first selecting All SPL in REW to show all the measurements, then performing the Remove Smoothing command from the Graph menu. This will remove smoothing from all measurements at once.

When exporting unsmoothed data as a text file, the frequency resolution of the exported data will be very high. MSO will internally generate its own list of frequencies using a log sequence whose frequency resolution will not exceed a set limit. This is to ensure that all the measurements share a common set of frequencies so they can be added together, taking phase into account. This addition is performed to calculate the combined output of all the subs and main speakers at each listening position over frequency.

Using HolmImpulse With The Time-Locking Feature (Windows Only)

Another option for measurements using Windows-based systems is the HolmImpulse freeware. Unfortunately, the HolmImpulse documentation tends to be theoretical in nature and doesn't include much information about how to actually use the software. A worthwhile source of information for learning HolmImpulse is the HolmImpulse tutorial at the Parts Express forum.

HolmImpulse uses a single-input, single-output approach that nonetheless can achieve time-synchronized measurements using the so-called "time locking" feature. The diysubwoofers.org site has a very good step-by-step procedure for using the time-locking feature. A word of caution is in order regarding HolmImpulse and time-locking. When used with sound cards having clocks that are not locked to one another on the record and play sides, the timing reference will shift from one measurement to the next, causing them to be non-time-synchronized and thus unusable with MSO. To find out more about how time-locking can fail in HolmImpulse, see the "Holm Impulse - Timelock, creeping offset?" thread at the Parts Express forum. That thread contains a description of a test that can be run to determine if your sound device is compatible with time-locking. It also contains an explanation by Bill Waslo, author of the Dayton OmniMic measurement software, of the clocking issue as it relates to the potential failure of time-locking in HolmImpulse.

If the clocks on the record-side and playback-side sound devices are not locked to one another, then as Bill Waslo states in the "creeping offset" thread, the time locking cannot work. With that information, it can be seen why using a USB microphone will cause HolmImpulse time-locking to fail. The USB microphone contains an internal A/D converter, so the converter's clock must be contained within the microphone itself. An analog or digital signal from the computer's sound device supplies the input signal to the device under test in this configuration. This signal's clock is internal to the computer itself or in an attached external sound device, so the record and playback clocks are physically different and can't possibly be locked to one another. This will cause time-locking to fail. A similar situation will occur if one attempts to use HolmImpulse time-locking with HDMI supplying the measurement source signal.

It should also be pointed out that these failures of time-locking won't be flagged by the software. In order to find them, you'll need to look for them by performing multiple measurements and checking for consistency.

FuzzMeasure (Mac Only)

FuzzMeasure has a loopback timing reference capability and is therefore recommended for use with MSO, provided a USB microphone is not used. For more information about using loopback with FuzzMeasure, see this Youtube FuzzMeasure tutorial video.

Dayton OmniMic

Dayton OmniMic cannot be used with MSO, as it has no timing reference capability.

General Considerations

When performing subwoofer measurements for use with MSO, it's essential that only one subwoofer be energized at a time for a given subwoofer measurement. Assuming you have a DSP device for performing individual EQ of each sub, this means muting all outputs of the DSP device except the one associated with the sub being measured. If you have more than one sub connected to a given DSP channel, these count as a single "logical sub" for the purposes of MSO, so there is no need to individually energize the physical subs. Multiple subs assigned to a single DSP channel should be physically close to one another.

When performing mains measurements, you have more than one choice. You can choose to integrate center channel plus subs, or left and right plus subs, but not both. Suppose you have a four-channel DSP, a total of four subwoofers, and each DSP channel individually drives a single subwoofer via its associated power amp channel. If you choose to integrate center channel plus subs, each listening position would consist of five measurements as follows.

If you choose to integrate left and right front channels with the subs, and are energizing the left and right front channels simultaneously, each listening position would consist of the following five measurements.

If you choose to integrate left and right front channels plus subs, and you are unable to energize left and right simultaneously, each listening position would consist of the following six measurements.

Combined frequency response of mains and subs can be optimized at multiple listening positions, or you can choose to optimize just the subs.

How Many Positions Should Be Measured?

The algorithm improvements introduced into version 1.44, when Minimize seat-to-seat variations was introduced, caused a change in guidelines for the number of positions that should be measured. The recommendations now depend on which optimization method you've chosen in the Method property page of the Optimization Options dialog.

Minimize Seat-to-Seat Variations Mode

This is the recommended optimization mode for the majority of applications. When using this mode, the error of the main listening position from the target curve is always weighted to be one-fourth of the total error. This makes the relative contribution of the error of the MLP from the target curve to the total error independent of the number of positions measured. Therefore when using this mode, it's recommended that you measure as many listening positions within the listening area as is practical.

As Flat as Possible Without Additional Global EQ Mode

When using As flat as possible without additional global EQ mode, it is recommended that the number of listening positions measured be the same as or a little greater than the number of independently-controlled subs. When the number of listening positions begins to exceed the number of independently-controlled subs by a significant amount, the ability to get response that's as close to the target curve as possible at any given position becomes degraded. This mode is usually reserved for specialized applications.

HDMI Considerations

Do Not Use HDMI Channel 4 (LFE) for Subwoofer Measurements When Using MSO to Integrate Mains and Subs

It might seem that using HDMI channel 4 (the LFE channel) to energize the subs would be the correct way to perform the sub measurements when using a typical AVR or pre-pro. However, if you are using MSO to integrate your main speakers and subs, using the LFE channel to energize the subs would be incorrect. The LFE channel has a 10 dB higher gain than all other channels, so the result will be sub measurements that are in error by 10 dB from the correct values. In addition, the low-pass filter assigned to the LFE input, sometimes called the "LPF of LFE", will often have a different cutoff frequency from the one used in the bass management, and may even have a different slope and alignment as well. For these reasons, channel 4 should never be used to energize the subs for measurements to be used with MSO for integration of subs with main speakers. Instead, energize the subs by using the center-, left-, or right-channel signal, forcing the AVR's bass management to generate the subwoofer signal.

It's Okay to Use HDMI Channel 4 (LFE) for Subwoofer Measurements When Using MSO to Optimize Subs Only

For sub-only optimization, using HDMI channel 4 is okay, as the 10 dB level increase of the LFE channel relative to the main speakers is not a problem when you're measuring only subs. When optimizing only subs, it's best to set the LPF of LFE to the highest allowable setting to get the flattest possible sub response up to the highest measurement frequency.

Using HDMI With the REW Acoustic Timing Reference and USB Microphone

If you're a Windows user, it's a good idea to read AustinJerry's guide (in PDF form), called "Getting Started With REW: A Step-by-Step Guide" before doing measurements with HDMI and a USB microphone (which requires an acoustic timing reference). This document has many details and step-by-step instructions for how to get everything set up and running. Even experienced users will likely find something useful that they may not have known before.

Mac users should check out the Mac REW tutorial thread on AVS forum by Enrico Castagnetti from Rythmik Audio.

HDMI Channel Numbers on Windows

The table below shows the HDMI channel numbers to use with REW on Windows to energize the corresponding channels in a 7.1 setup. When using a Windows computer with an HDMI interface with REW, use a universal ASIO driver such as FlexASIO or ASIO4ALL to enable the measurement of all available channels if the hardware supports it. FlexASIO is the more robust of these alternatives, but requires using the latest REW beta version to avoid the need for manual editing of its configuration file.

Channel Number Channel Name
1 Left
2 Right
3 Center
4 LFE
5 Rear left
6 Rear right
7 Side left
8 Side right

Channel number vs. physical channel for Windows 7.1 audio

When using the REW acoustic timing reference feature, it's essential that the speaker you use as a timing reference not be a subwoofer. The timing reference speaker must be able to reproduce a frequency sweep above 5 kHz. It's common to use a surround speaker for this purpose.

The example below demonstrates the use of the acoustic timing reference when measuring a system having a Left channel, Right channel, and four subs. It's also essential that you use the same channel as the acoustic timing reference for all measurements. The example below uses the right channel for this purpose. This example assumes that you'll be using a "subs + mains" configuration to integrate the subs with the main speakers. The usage of a conventional AVR or pre-pro, together with a DSP device that supports 4 subs connected to the AVR's sub out is also assumed. Notice that for the subwoofer measurements, the REW output is set to the left channel, but the left speaker is muted or disconnected, so that no sound comes from it. In this case, the left channel energizes the subs through the AVR's bass management. These measurements need to be repeated at each listening position.

Left Channel, Right Channel and Subs

For the arrangement above, each measurement position becomes an MSO measurement group with the above six measurements. MSO performs a complex summation of these six measurements to obtain the data both for this measurement group's graphs and for its optimization. MSO will sense that there are two main measurements in the measurement group, and will boost all subwoofer data by 6 dB before summation to take into account the fact that both main speakers are summed together, but the subs were only energized by a single main channel (not both main channels simultaneously).

The following example is for a center channel and four subs, using the right channel as the acoustic timing reference. These measurements need to be repeated at each listening position.

Center Channel and Subs

Hardware Loopback Measurements Using HDMI and Non-USB Microphones

In order to do hardware loopback measurements with HDMI, you'll need to use either a pre-pro or an AVR with preamp outputs. You'll need to take the loopback signal from the preamp output corresponding to the HDMI output used for the timing reference signal. You'll also need to avoid configurations that cause a 6 dB error in the sub measurement level.

This technique does not work with USB microphones.

Avoiding Configurations That Cause 6 dB Sub Level Measurement Errors

A 6 dB sub level measurement error will occur when the following conditions occur simultaneously.

The only way bass redirection in the loopback channel is acceptable is if the left and right main speakers are energized and measured simultaneously, and the loopback channel is either the left or right main channel. The workaround examples below should clarify this idea.

Workaround 1: Measure Left and Right Main Speakers Simultaneously

In this workaround, both left and right main speakers are measured at once. The timing reference signal also serves as the stimulus for the left or right channel. The following example uses the right channel as the timing reference. For this specific case, a y-cable is required at the right-channel preamp output so the signal can be routed to both the right-channel power amp input and the timing reference input of the computer's sound device.

Workaround 2: Disable Bass Redirection in Loopback Channel

This workaround measures a single main speaker at a time. It requires that bass redirection be disabled in the loopback channel by configuring its respective speaker to "large" in the AVR or pre-pro, and ensuring that no special mode is enabled, such as "LFE+Main", "Double bass" and so on. These modes redirect bass even when the main speakers are set to "large". Note that this means that if you're measuring left and right main speakers individually, you must use some channel other than left or right for the loopback. Here's an example of measuring left- and right-channel speakers individually.

Note: Some AVRs and pre-pros won't let you set the center channel and/or the satellite channels to large if the left and right front channels are set to large.

Avoiding Other Problematic Configurations

Problems associated with hardware loopback can also occur even when measuring the subs through the LFE channel (HDMI channel 4) of the AVR for use with MSO "sub-only" configurations. For this case, the fix is to set the speaker type of the loopback channel to "large" and avoid "Double Bass"-type modes in the AVR or pre-pro. Bass redirection in the loopback path must always be avoided in this case. If it is not, some really strange effects can occur. This is because for such arrangements the signal driving the subs will consist of the sum of two signals filtered by two different low-pass filters: the so-called "LPF of LFE" and the normal crossover LPF. This results in comb-filtering effects that might even be mistaken for room mode effects in the sub measurements.