Filter generation options

General settings

Frequency range

This is the range of frequencies where a flat room response is intended. Do not extend the range far beyond the capabilities of your speakers. Especially speakers with vented enclosures dislike lower frequencies and may be destroyed this way.
This also controls the X axis of the graphs.

Frequency bin size

The resulting filter kernel is an interpolating function rather than exact values. The next frequency is calculated by adding a constant and a relative factor to the current frequency.
There is normally no need to change this option.

Do not process phase response

This is mainly a work around for insufficiently accurate measurements. If you turn on this option then the resulting filter will not compensate for the group delay but only for the frequency response. This results in a symmetric, linear phase filter. Note that this is a significant cutback. You will not get the crisp results than from a full compensation.

Normalization settings

DRC123 must normalize the different measurements to a common level before combining them. This is done by calculating an average gain and an average group delay.

Frequency range

Range of frequencies that contribute to the normalization.
There is normally no need to change this option.

Normalization mode

When normalizing the gain you can chose between quadratic (energy) and logarithmic (dB) mode.
Normalization by energy results in the same amount of energy transferred if white noise is passed through the filter. In this mode +3 dB and -3 dB do not cancel since the first causes twice as much energy while the second causes only half of the energy, in average you are still at 125%. This mode lowers the gain if the response is changing a lot. It usually prevents the audio DAC from overdriving.
Normalization by dB causes the total loudness to be kept approximately. While this sounds useful there are serious drawbacks. First of all some frequencies have to be amplified by a significant amount. If your samples have been normalized to 0 dB FSR (full conduction) it is likely that your sound device overdrive with the corrected response. Taking back the master volume will not prevent from that on most sound hardware. Secondly an audible resonance usually sounds much worse than a frequency notch. So in general normalization by energy should be preferred unless you have professional studio equipment.

Limiter

To get reasonable results extreme corrections have to be avoided. First of all to prevent damage from your speakers but also to avoid audible artifacts.

Limit gain

Maximum positive gain factor for the correction. If your room response falls below this amount no further compensation takes place. Do not chose unreasonable high limits to protect your speakers.
Additionally the gain change rate is smoothed by an IIR filter.

Invert high gain values

This option cause DRC123 to decrease the gain of frequencies that need an amplification beyond the gain limit instead of simply clipping to the maximum value.
The idea is that a speaker should be protected from frequencies that it cannot reproduce anyway rather that putting more and more energy into it which might cause damage.

Limit delay

Minimum and maximum delay that is compensated by DRC123. This is mainly to restrict the required length of the resulting filter kernel.
Additionally the change rate of the group delay is restricted by an IIR filter. Fast changes in the group delay are likely to produce audible artifacts.

Display settings

Gain range

Range of the Y axis of the frequency response graphs.

Group delay range

Range of the Y axis of the group delay graphs.