Cavern QuickEQ setup guide

What is QuickEQ?

Cavern QuickEQ is Cavern's advanced and customizable room correction software, bundled with the Cavern Driver. QuickEQ works with various devices, but uses Equalizer APO for calibrating the sound of PCs. This section of the documentation deals with the configuration of QuickEQ and exporting its results. On this page, in the other tabs, you can find each of QuickEQ's configuration panels' description. Export and other guides are available on the sidebar.

Before you start, please note that by default, all settings are set to the values that are correct in most cases. If you simply select your microphone with its configuration files and press Start, the results will most likely be the best you can get out of your system, and you just have to export it. Maybe you should try changing the target curve or doing a multi-mic measurement, but in most cases, the default is fine.

Minimal setup

QuickEQ can be a one-click experience, but a single thing has to be set up first: the microphone. The automatically selected system default is most likely the one you'd want to use, but double checking is always a good safety measure. Depending on your equipment, follow one of the following:

Using a regular microphone: When the microphone is not calibrated, use the Smoother target curve under the Setup tab. All other options are for calibrated microphones only.
Using a calibrated measurement microphone: At least one calibration file has to be loaded on the Microphone tab for the calibration to properly work.

After that, you could start the measurement, the default settings are based on what most people like and what works on most systems. If you don't want to tweak the settings, look at one of the export documentations on the sidebar. This page and its tabs deal with everything you can modify about EQ generation. All settings can be changed after measurement, nothing requires measuring again.

Controls

Use these buttons to start a measurement, interrupt an ongoing measurement, or exit QuickEQ and return to the Cavern Driver.

90° calibration

Open the microphone's 90° calibration file here. Calibration microphones should be pointed upwards and calibrated to sound coming from around it. If only this calibration is given, it will be used for all speakers, otherwise the curve will be interpolated with the 0° calibration.

0° calibration

Open the microphone's 0° calibration file here. This is the calibration for sounds that directly hit the microphone membrane (from the top center speaker if the microphone is rotated correctly). If only this calibration is given, it will be used for all speakers, otherwise the curve will be interpolated with the 90° calibration.

Microphone

Select your microphone here. The system should default to occasionally connected measurement microphones, but if you don't know if this is the case, make sure it's the active one by selecting your microphone in this panel.
To select multiple microphones, hold the Ctrl key while selecting a device. The reference microphone is marked blue, the other selected microphones are green. For correct microphone positioning, refer to the Multi-mic page. When multiple microphones are selected, the following menu appears when you press the Start button:

The options are:

Use reference mic only: Make a measurement only with the reference microphone, while not adding it to the averages. This is the same as pressing Start with a single microphone selected.
Reference + average (all): Uses all selected microphones for measurement. The reference microphone will be added as reference to the Multi-mic list, the others will be added to the average.
Average only (all selected): Uses all selected microphones for measurement. All microphones will be added to the Multi-mic average.

For the last two options, averages are immediately calculated.

Settings related to the measaurement and correction target.

FFT size

The length of the measurement in samples. The longer the measurement, the more precise and less noisy it will be. Recommended sizes are:

32768 - The minimal FFT size that supplies acceptable resolution for room correction. High frequencies can stilly be noisy.
65536 - For low volume measurements. In low noise environments (basically in most home theaters and even living rooms) higher values won't make a difference.
131072 - Compensation for high noise or low volume. If the measurement is still inconsistent, go higher.

Gain

Measurement signal volume relative to system output. The reference cinema level is -20 dB, home theater equipments are mostly using -30 dB. When the noise is low enough (two measurements after each other has the same result), this value doesn't matter, and choose the lower, don't damage your hearing.

Apply subsonic filtering

Remove low frequency signals from the speakers they are unable to reproduce efficiently. This can help protect the speaker, but might reduce audio quality. The feature is also useful for most cheap AVRs, which can't properly produce power on all channels simultaneously, as subsonic filtering reduces power draw.

Correct delays

Apply delays in the exported configurations to counteract different speaker distances and correct relative phases.

Warm-up mode

Waits a sweep's time before the actual measurement. This helps for measuring with microphones that click when the system turns them on. Use this feature when the left channel is drastically different from the others.

Gain standard

Sets the volume difference between channels. Cavern sets all channels to the same volume, while DCI (Digital Cinema Initiatives) applies -3 dB to the surrounds. DCI is the actual cinema standard, but it's not suitable for many games, Cavernize, and is not preferred by many users. The Custom option is for manually setting the differences.

Target curve

Target curve to equalize the system to:

Flat - No additional EQ is applied, each frequency response is flattened as much as Target EQ resolution allows.
Room Curve - Frequently used home theater target curve with amplified lows and suppressed highs. A healthy mix of Depth and X-curve.
Punch - Adds an emphasis on the chest slam range with noticeable falloff so compatible effects rise to a higher amplitude in the same time span.
Depth - Adds a ramp to the subwoofer that keeps increasing the gain down to the end of the human hearing range. This can help increase the perceived low extension of the speaker.
X-curve - The standard cinema curve, nearly all movies were mixed to this target and all cinemas are configured to this curve. Greatly suppresses high frequencies.
Bandpass - Recommended for stage use, where the subwoofer level is changed all the time. This curve makes the transition to the subwoofers seamless at any volume difference if crossovers are not used. This curve is only applied to LFE channels.
Smoother - This option does not calibrate the system to any standard curve, only differences between channels are removed, making them sound uniform to each other, creating a good sound stage. This makes the Smoother curve work without a calibrated microphone.
Custom curve - Advanced users might load custom curves from TXT, FRD, or TARGETCURVE files. These are formatted the same way as calibration files, each line containing frequencies and their respective gains, separated by a single space.

The following example somewhat resembles the Depth curve for low extension and has a moderate high frequency rolloff based on X-curve:

20 6 120 0 2000 0 10000 -3 20000 -6

The file could be saved as TXT and loaded into Cavern, where the result is:

Regain volume

If the maximum system volume is still not enough after calibration and other volume options, this is the last resort for increasing it. The safeness of preset values show the likeliness of running into the 0 dB limit, thus the system compressor. The Auto option tries to determine the highest positive gain with the resulting EQs which won't hit the system compressor, making this the recommended option.

Target EQ resolution

Determines how strongly the curves are smoothed and how many EQ bands will be used.

Low - Bands are 1/3 octaves from each other, LFE bands are 1/12 octaves from each other. Curves are smoothed to 1/3 octaves and LFE curves are smoothed to 1/12 octaves.
Var - Bands are 1/12 octaves from each other, LFE bands are 1/48 octaves from each other. Smoothing starts as the High option and ends as the Low option.
High - Bands are 1/12 octaves from each other, LFE bands are 1/48 octaves from each other. Curves are smoothed to 1/12 octaves and LFE curves are smoothed to 1/48 octaves.
Full - The EQ is completely adaptive, bands are inserted wherever they are needed, at each local minimum and maximum. Curves are smoothed to 1/12 octaves and LFE curves are smoothed to 1/48 octaves.

Advanced features

These features could be inaccurate and might require several tries to get them right or they might even require advanced audio knowledge. Use them at your own risk, and disable them if the results don't match expectations.

Correction frequency range

Limit the generated EQ bands to this range. Contrary to popular beliefs, using limits like this will most likely result in worse sound. This feature should only be used in special cases, following tips about limiting the effect of the calibration is strongly not recommended.

Experimental features

Features in this box are not considered accurate, you should double-check their results. More information about them is visible as tooltips on mouse hover. For example, the Correct polarities option might mess up systems close to 90-degree polarity, so impulses should be manually checked.

Cavern Clear Center

This feature is designed to counter diffraction from centers with two low frequency drivers horizontally and/or poor acoustic treatment by simulating a worse acoustic environment. This should result in more human-like speech, less distortion, and a more uniform soundstage on the front. However, this feature will degrade well treated rooms with high quality equipment.

This feature requires:

PC used for correction
48 kHz system sample rate
checked "Correct delays" option
an extra WAV file next to the export

Cavern Sealing

Cavern Sealing simulates the sound character of a sealed subwoofer on a ported subwoofer. This does not work the other way around!

This feature requires a PC used for correction with 48 kHz system sample rate.

The impulse response of a sealed subwoofer decays way slower than a ported one, in an oscillating manner. Cavern Sealing emulates this behavior on any ported subwoofer. Before applying Sealing, check the difference it makes:

Optimized amplifier use - Since this feature spreads the required impulse power in time, any impulse effect applies way less strain on the amplifier: Light mode cuts the required power in half, while Heavy mode makes a tenfold difference. This reduces port noise a bit, but it can't be eliminated.
Feeling changes - Only fast ported subs are capable of chest slams, sealed cabinets can only massage. Hard chest slams might fully disappear.
Resonance prevention - Since the character has an oscillating element, pure sine waves will get amplitude modulated, reducing the hum that is commonly present in ported subs, even when calibrated properly.
Musical character - A good ported subwoofer can barely be heard, only felt. This makes subwoofers only designed as LFE to have very poor quality when listening to music. Sealed subwoofers have good, warm musical character.

The Only seal for crossover feature does not alter the LFE channel, only applies Sealing on sounds moved from the main channels with the crossover. This creates a different low frequency character for the mains than for the LFE. This is the way all commercial cinemas operate, but it's nonexistent in the home theater industry.
If you set any of the following modes, the feature will be enabled:

Light - Tries to be a middle ground, by barely changing the sound character. This might get you some of the pros of a sealed subwoofer, like the musical character, but barely.
Heavy - Aggressive, but not complete character transition, elongates impulses as long as it can without huge delays. Resembles a midrange sealed subwoofer. Simulating high-end sealed subs is possible, but it was not included in Cavern because of the huge delay that simulation requires.

Arbitrary measurement signal

Any signal can be used for measurement, even noise or songs. The default measurement signal, the swept sine wave is used because it has the highest possible signal-to-noise ratio, and changing it can drastically reduce the accuracy of the measurement. Only RIFF WAVE (.wav) files can be used for this feature, and a mono mix will be played on each channel. The file length will be truncated to the nearest lower power of 2, for example, a file with 50000 samples will result in only the first 32768 samples being used. This behavior actually helps minimizing noise.

Impulse response filtering. For multi-mic measurements, this feature is not available.

Show impulse responses

Change the displayed graphs from frequency response to impulse response.

Show verbose measurements

Display both the frequency and impulse response for each channel on the same panel with extra informations such as low frequency extension, phase, and RT60.

Left IR window

If you enable this options, you can window out samples to the left of the largest impulse spike. Only noise and harmonic distortion is present in this region unless heavy processing is applied on the system, so you can usually safely cut most of this region.

Right IR window

If you enable this options, you can window out samples to the right of the largest impulse spike. Reflections, low frequency details, and most noise are contained in this region, so expect a more precise but less detailed curve if you cut more samples.

Offset

Cavern's crossovers are only available on PC, when exporting to Equalizer APO. For dedicated AV equipment, only the internal crossover can be used. Internal crossovers must be bypassed if you use any crossover from this panel.

Crossover frequencies

Crossover frequencies for screen and surround channels, sounds below these frequencies will be moved to the subwoofer(s). Try to set them to the lowest frequencies your main channels can drive without any issues, based on the measured curves. Move the sliders all the way to the left to disable crossovers (which is their default state, as most systems already have crossovers).

Clipping protection

Adjusts gains for the entire system to make sure the sounds moved to subwoofers can't hit 0 dB FS.

Use CUCC

Use Cavern Ultimate Crossover Chain, an experimental bass management system dividing low frequencies across all channels. This feature requires completely disabled crossovers (other than Cavern's own) in the audio chain and high power amplifiers on most channels.

Used crossover

Crossover method. Use Cavern for maximum quality, instant crossover, and linear phase in small rooms, but that requires perfectly set delays. In other cases, Biquad will sound more uniform, but breaks phases. Synthetic biquad is the best of both worlds for large rooms: it's phase-perfect with the same frequency response as Biquad.

Crossover visualizations

This is a debug interface to visualize how crossovers based on biquad filters behave. Their impact on the curve will not be exported as any crossover's spectrum distortion is measured and already corrected.

Use the multi-microphone measurement feature to create a more uniform listening space, which means the experience of multiple seats are closer to each other. To add a position, start a normal measurement, and use one of the Add this result... buttons after:

Add this result to average - Used for every position except the reference detailed below. These results are only used for increasing the EQ precision.
Add this result as reference - Used for the reference position, which is the center seat in home theatres, and the center seat 2/3 of the way to the back in the cinema. Delays, gains, and polarities are calculated from the reference position.

After all positions were measured, press Calculate average, and the averaged result of multiple measurements will appear.

Recommended measurement positions up to 5 points

How to measure at over 5 points

Measure at the position of a single mic measurement first. Every other position should be randomly selected around it, with the following considerations:

further than 50 cm from the nearest wall,
not at 1/2, 1/3, 1/4... position on any axis,
in line of sight from each tweeter.

Cavern's multi-sub correction is only available on PC, when exporting to Equalizer APO.

7.1 downmix

When your system has more than 1 subwoofer, this panel will be active after measurement. Since multi-sub layouts are not available in any operating system, if you set up any output channel as a subwoofer, you need to downmix the system output after measurement. Cavern can do this for you, if you select one of the available downmix patterns. Any unmentioned channel for each option is a valid subwoofer output.

6.1 - 7.1 will be downmixed to 6.1.
Stereo - 7.1 will be downmixed to front left/right.
Quadraphonic - 7.1 will be downmixed to front and rear left/right. The first 4 channels will be used for this.
5.1 side - 7.1 will be downmixed to 5.1 with side surrounds.
5.1 rear - 7.1 will be downmixed to 5.1 with rear surrounds.

All downmix patterns are normalized. This means that they result in additional volume loss. The more channels you keep, the less volume is lost. Increasing Regain volume under Setup might be required to push the system peaks back to 0 dB.