Pass Filter CHOP
Summary
The Pass Filter CHOP filters audio input using one of four different filter types.
Contents |
Some Filter Terminology
A Low pass filter removes the higher frequencies of a sound, while a high pass filter reduces the bass of the sound. A band pass filter is used to extract a frequency range (i.e. extracting a person's voice from background clutter) and a band stop filter is used to cut out a frequency range.
If a certain frequency lies outside the pass band, sounds at that frequency will be reduced in magnitude. The farther outside the pass band the frequency is, the more it will be reduced.
The Cutoff frequency is also known as the "half-power" frequency. A wave at the cutoff frequency will be reduced to half power.
The Rolloff Factor of a filter determines how quickly the drop occurs at its Cutoff frequencies. A low rolloff will produce a gradual filter falloff (more of the sounds outside the frequency range are heard), and a high rolloff will produce a sharp filter falloff.
Parameters - Pass Page
Filter
- Low Pass Filter - All frequencies below the High Cutoff are passed through the filter (the "pass band").
- High Pass Filter - All frequencies above the Low Cutoff are passed through.
- Band Pass Filter - All frequencies between the Low and High Cutoff are passed through.
- Band Stop Filter - All frequencies above the High Cutoff and below the Low Cutoff are passed though.
Low Cutoff /cutofflow - The frequency (in Hertz) of the lower cutoff. This value is not used by a low pass filter, since it has only an upper cutoff.
High Cutoff /cutoffhigh - The frequency (in Hertz) of the upper cutoff. This value is not used by a high pass filter, since it has only a lower cutoff.
Pass Gain (dB) /gain - Defines the gain of passed frequencies, specified in dB (decibels). Every increase of 20dB corresponds to a 10 times power increase (and decreases by 20dB similarly reduces power to 0.1 the original).
Rolloff Factor /rolloff - Defines how the filter drops off at the cutoff frequencies. Low values (less than one) produce more gradual rolloff, and higher values produce sharper filters.
Also Filter Phase - Normally, the magnitude of the signal is filtered. You can optionally filter the phase of the signal as well. This takes about twice as long, and normally doesn't produce audible differences. It is included for special cases.
Filter Animation Channels Input
The second input is the Filter Animation Channels, which allows the filter parameters to be changed over the CHOP's interval.
Parameters - Digital page
Digital Filter Overview
Digital filters use a Discrete Fourier Transform (DFT) to filter the sampled data. A digital filter processes chunks of data, unlike an analog filter, which continuously processes data. The digital filter divides the sample into chunks which are each individually filtered. This causes discontinuities in the filtered sample at each division, which must be removed.
By discarding the extreme ends of a chunk, and then slightly overlapping and blending it with the previous chunk, these discontinuities can be eliminated.
The chunk size of a digital filter affects its characteristics. Smaller chunks allow more control of the filter when filter parameters are animated, but have coarser frequency resolution. At very low frequency resolution, the filter will have a grainy sound.
When filter parameters are animated, they are only evaluated once per chunk. If the chunk is large, then actual steps between chunks may be heard, rather than a smooth transition (assuming the animation was meant to be smooth!).
The default values of this page are good for general use, and normally do not need to be changed.
Filter Chunk - This parameter allows you to select the chunk size, to balance the demands of animating filter parameters with the quality of sound.
Chunk Overlap /overlap - How much of the current chunk is overlapped and blended with the previous chunk. It can range from zero (no overlap and blending) to 0.95 (95% overlapped). For practical use, keep it between 0.05 to 0.3 . Values close to one will produce long cook times, and values too close to zero will not entirely remove the discontinuity.
Chunk Discard /disgard - How much of the chunk to throw away. Since the middle section is most accurate, the end sections are discarded. This parameter can also range from zero (don't discard anything) to 0.95 (keep only the middle 5%). Normally, discard between 0.05 and 0.2 of the chunk. Larger values will produce long cook times.
Standard Options and Local Variables
Local Variables: $I $C, $NC and $N - The current digital chunk being processed.
| ||||||||||||||
