The Analysis Preferences alter the way REW carries out some of its calculations.
The Left Side and Right Side window selectors offer a choice of window types to be applied to the impulse response data before and after the peak. These are the defaults applied to new measurements, window types for existing measurement can be altered via the IR Windows toolbar button. By default REW will set the widths of the windows automatically to show the whole room response, with a 500 ms right side window and a 125 ms left side window if the end frequency of the sweep is above 200 Hz, or a larger left side window for end frequencies below 200 Hz. To override this uncheck the Set window widths automatically box and set the default widths you wish to be applied to new measurements.
If Add frequency dependent window is selected a frequency dependent window (FDW) will be applied to the measurement after the left and right windows have been applied. The width of the FDW is set by the controls to the right and can be specified in cycles or in octaves. If the width is in cycles a 15 cycle window (for example) would have a width of 150 ms at 100 Hz (15 times 10 ms), 15 ms at 1 kHz (15 times 1 ms) and 1.5 ms at 10 kHz (15 times 0.1 ms).
The IR Decay/Waterfall Left and IR Decay/Waterfall Right window selections are applied to the impulse response data when generating Spectral Decay and Waterfall plots.
The Waterfall (audio data) window selection is applied when generating waterfall plots for imported audio data. This is the default, the window type used for any particular measurement can be altered via the selector in the graph controls.
The Spectrogram window selection is applied to the response data when generating Fourier Spectrogram plots. This is the default, the window type used for any particular measurement can be altered via the selector in the graph controls.
The timing reference selection controls whether REW uses a loopback on the soundcard as a timing reference, or an acoustic timing reference, or no reference. Using a timing reference allows REW to eliminate the variable propagation delays within the computer and soundcard so that separate measurements have the same absolute timing. If a loopback is selected the reference channel signal must be looped back from output to input on the soundcard and measurements will be relative to the loopback timing, usually this means measurements will have a time delay that corresponds to the time it takes sound to travel from the speaker being measured to the microphone. If an acoustic timing reference is used REW will generate a timing signal on the output that has been selected to act as the reference before it generates measurement sweeps on the channels being measured. The timing signal is a high frequency sweep to allow accurate timing, a subwoofer cannot be used as the reference channel. Measurements will have a time delay that corresponds to the difference in their distance from the microphone compared to the distance of the reference speaker - if the reference speaker is further away the delay would be negative. Note that multiple sweeps cannot be used when using an acoustic timing reference.
If using a timing reference REW can calculate the delay through the system being measured and show it in the measurement Info panel as "System Delay" in milliseconds, with the equivalent distance in feet and metres shown in brackets. Note that delay values are not accurate for subwoofer measurements due to the limited bandwidth of the subwoofer response, the delay estimate is based on the location of the peak of the impulse response and subwoofers have a broad peak and a delayed response.
If a timing reference is not being used there are 3 options for setting the t=0 time of the impulse response: the time at which the peak occurs, the time at which the IR starts or the time after removing the estimated time delay. The IR start time is based on the point the response first exceeds 10% of the peak level. Using the option to Set t=0 at IR start may produce unexpected high frequency phase behaviour if the system being measured has a bandwidth that is close to or exceeds the measurement bandwidth, such as soundcard loopback measurements. That is because such measurements will exhibit ringing around the peak at close to the Nyquist frequency, meaning the IR start will be assigned to a position several samples before the peak where the ringing first exceeds the 10% threshold.
The Decimate IR selection controls whether REW reduces the sample rate of the impulse response to correspond to the range of frequencies in the measurement. Selecting this option greatly reduces the impulse response size for low frequency measurements and speeds up processing of the data.
When impulse responses are imported the t=0 position can be set to either the first sample in the imported data or the location of the peak of the impulse response.
After REW has made a measurement it can truncate the derived impulse response to preserve the important information while minimising the storage required for the measurement file. A 1 second period is retained before the peak, and by default a 1.7 second period is retained after the peak (this varies a little depending on the sample rate, at 44.1k (or multiples) it is approx 2 seconds, at 48k 1.7 seconds). In some cases it may be useful to retain more of the impulse response, such as measurements in very large spaces which have very long impulse responses. REW provides options to truncate the response after approx 4.4 seconds, or 9.9 seconds, or to retain the entire impulse response. It may also be necessary to use a longer sweep (with a 256k sweep there is about 6 seconds of impulse response data available after the peak). Note that retaining the entire impulse response will produce much larger measurement files, especially if long measurement sweeps are used. If the entire response is retained the peak will be centred within the response.
When REW generates in impulse response it is oversampled to provide data for graph rendering. The method used to generate that oversampled data is set by the IR oversampler setting. Windowed sinc uses a bandlimited interpolator, FFT time shift uses fractional delays implemented via the time shift property of the FFT, None disables IR oversampling (in which case Hermite interpolation is used in the graph plot generation between the original sample points). Windowed sinc is the default setting.
The Allow 96 PPO log spacing selection controls whether REW is permitted to convert frequency responses from linearly spaced data to logarithmically spaced data at 96 points per octave. The FFT that calculates the frequency responses produces data that is linearly spaced in frequency, i.e. there is a constant frequency step from each value to the next. For the high frequency parts of responses this means there are a very large number of points, using a lot of memory but not contributing anything useful to the displayed data. When this option is selected (it is on by default) REW will automatically convert frequency responses to more efficient logarithmic spacing with 96 data points in each octave of the response if this will reduce memory usage (which is usually the case for sweeps that end above 300 Hz or so). As part of the conversion process REW first applies a 1/48th octave smoothing filter to the data to remove any high frequency combing from the response. The conversion takes place on any new measurement or when an IR Window is applied. Whether a measurement is log spaced or linearly spaced can be seen by bringing up the measurement info window by clicking the info button in the toolbar.
Note that conversion to 96 PPO is inhibited if the impulse peak is far from the impulse zero time, where "far" means the peak is offset from zero by a time that corresponds to more than 90 degrees of phase shift between samples at the measurement end frequency. This is to prevent aliased phase data at high frequencies which would lead to incorrect group delay figures.
The Show response below window limit selection controls whether REW displays the frequency and phase responses at frequencies lower than those valid for the current impulse response window width. For example, if the window width were 10 ms frequencies below 100 Hz would not be valid and are normally not displayed. There are circumstances in which it may be helpful to see that data, which this option allows, but the responses are drawn dashed to indicate they lie below the window cutoff.
The default smoothing to apply to new measurements can be selected from the drop-down box next to the log spacing check box.
If Limit cal data boost to 20 dB is selected REW will limit the total gain applied to compensate for calibration data attenuation to 20 dB. This prevents excessive boosting of the noise floor in areas where the combined mic/meter and soundcard response is more than 20 dB down. This setting affects the frequency response, RTA trace and SPL meter readings and is also applied when carrying out trace arithmetic.