Adapted from the ZynAddSubFX Equalizer.

Gain: Overall output level

Q: Resonance of individual filters. Generally helps smooth extreme settings...or make extreme settings yet more extreme.

Originally adapted from ArtsCompressor. DSP routine has been rewritten entirely as the code had been adapted at a stage in ArtsCompressor development prior to many bug fixes that occurred later in its evolution which unfortunately removed calculation of a soft-knee compression characteristic. This is mentioned to give hope to those who were aware of the bugs in compressor in Rakarrack 0.3.0. These bugs have been fixed, and this compressor has been traced using test signals to confirm it does what we think it does.

A. Time: Attack Time. Time in milliseconds for attack to settle to 64% of final amount of compression.

R. Time: Release Time. Time in milliseconds for gain to return to 64% of it's initial setting.

Note: For the curious, the strange number, 64%, relates to the RC time constant (RC means Resistor-Capacitor), which follows a natural exponential curve. The use of this behavior will make this compressor feel more natural to one who is most familiar with rack-mounted analog compressors.

Ratio: For every Ratio (dB) that the input exceeds the Threshold, the output will be allowed to increase by 1 dB. For example, using compression ratio of 2, if the signal gets 2dB above the threshold, the output will only go 1dB above the threshold.

Knee: Percentage of the region in dB of space between Threshold and 0dB (basically is Knee X -Threshold). Within this region, the ratio increases from 1 to log2(ratio). Threshold + Knee marks the point of full compression onset. For example, if knee is set to 100%, a very gradual compression characteristic will be obtained, but the maximum compression available is log2(ratio). A ratio of 32 will result in a real ratio ranging from 1 to 5. This is useful for processing cymbals as it does not crunch the dynamic attack quite so badly, but helps sustain the trailing resonance.

Thrhold: Threshold. Defines the onset of compression.

Output: How much gain to remove from the final stage of the compressor.

Peek: Enable/Disable peek compression.

Auto Output: Automatically calculate makeup gain. If this is unchecked, then adjustments of Ratio, Knee, and Thrhold will make changes to the overall output level. Unchecking may be useful to interpret Thrhold as a limiting threshold, and one would likely use a high ratio in this case.

Stereo: Process each channel separately.

Adapted from the ZynAddSubFX 'Distortion'

This is a waveshaper, and not particularly an amp or stompbox modeling effect. This must be used judiciously with EQ's as well as the LPF and HPF settings (Low Pass Filter, High Pass Filter).

Tip: Your classic green stompbox has a Pre Filter curve with cutoff at 720Hz. This corresponds to approximately 51 on the HPF setting. If trying to emulate stompbox or amp sounds, it is highly advisable to use the Pre Filter option with HPF higher than 30 and LPF lower than 70. As you increase drive, decrease LPF.

Rakarrack attempts to be friendly to computers limited in both RAM and CPU. Because of this, the waveshaper does not use oversampling to reduce the effects of digital aliasing. Without going deep into DSP theory, this means that hard clipping with a lot of drive will create harmonics that are greater than half the sampling rate. These harmonics get mirrored back into the audio range. If they are significant in magnitude, then there will be a "grainy" non-musical sound. This is what most guitarists are talking about when making reference to "digital distortion". Functions like Sine, Pow and Atan generate a lot of harmonics, and even at lower signal levels. Lmt, Clip and Zigzag are among the worst, but these are generally used for nasty noisy sounds, so a little digital aliasing may not be a bad thing.

Fortunately, for instruments such as guitars, you can significantly limit the bandwidth of the input signal without catastrophically discoloring the timbre of the instrument by applying Low Pass Pre-filtering. By reducing the high frequency content of the input, you can greatly minimize the level of aliasing harmonics present in the output.

The main point is that great sounding distortions can be obtained from Rakarrack, but the distortion module is not meant to emulate your favorite stompbox, but to off the flexibility create these sounds as well as unique flavors of distortion.

The Crunch waveshaping type is the most physically informed of the waveshaping functions currently used. The clipping characteristic is most closely related to curve produced by a JFET amplifier stage. With proper EQ settings, a sound reminiscent of high gain British stacks can be obtained.

Sub Octv: Allows you to mix some sub-octave rumble into the output. Technically speaking, this modulates the output with a square wave at half the fundamental frequency (sub octave).

Adapted from the ZynAddSubFX Distortion

Same as Distortion, but without sub-octave.

Adapted from the ZynAddSubFX Echo

Wet/Dry: Mix level of echo with original.

Reverse: Mixes reverse delay with forward delay.

More about reverse: Echo works by storing audio samples in memory and playing back samples that were stored a while ago. At a sample rate of 48kHz, the number of samples stored in memory for a 1 second delay is 48,000. To play back audio from 1 second ago, there is a reader that increments through, right in front of the write. The reader reads a sample from a second ago, and the writer writes that memory location with a new sample from the present.

Now, what if we could read this block of memory backwards? Then everything stored in memory comes out in reverse, and has an interesting reverse envelop sound. In the digital word, we can read it forward as easily as backwards, so why not keep track of both and let you mix them together?

Pan: Sends the delay more to the right or left. Zero balances it in the middle.

Delay: Amount of time before you hear the echo.

Lrdl.: Left/Right delay difference. A setting of 64 means that the delay time is the same for left and right channels. If larger than 64, the right channel is delayed longer than the left channel. Less than 64 the reverse is true. This allows you to achieve the stereo 'Ping Pong' effect.

L/R Cr: Left/Right Crossing. This mixes left and right channels.

Fb: Feedback. How much of the delayed output to add to the input. This makes the echo regenerate, like it's bouncing around in a canyon.

Direct: When is on, the effect only play the echoes.

Damp: Low pass filter on the delayed signal. This rolls off the high frequencies on every regeneration. Setting to a higher level will make the echo sound more natural.

Adapted from the ZynAddSubFX Chorus

Tips

Tempo: LFO Rate

Rnd: Add noise to LFO to make it more natural sounding.

Intense: Increase effect intensity. A digital Flanger or Chorus uses a delay line with several hundred to several thousand taps. When it changes taps to modulate the delay time, there is a discontinuous change in the signal called "zipper noise". To eliminate this, fractional delay times need to be calculated in order to smoothly transition from one delay line tap to the next. The default Rakarrack Phaser and Chorus use linear interpolation. This mode makes use of a Lagrange interpolation polynomial to estimate fractional delay times and some other tricks to maximize the flanger depth and flatten the frequency response on the chorus. Adding the mode switch adds the feature without breaking old presets.

St.df : Stereo difference of LFO.

L/R Cr. : Mix left into right, Right into left. At maximum level, left and right channels are swapped.

Adapted from the ZynAddSubFX Phaser

More mild digital phaser with exponential LFO sweep.

Phase: This is the offset for the "center" of the sweep.

Depth: LFO deviation or some may like to think of it as LFO amplitude. This is the same type of function that has been labeled 'width' in Analog Phaser.

Fb: Feedback. Some phasers title this "Regen". Feedback is negative as this moves toward zero, positive from 64+. 64, strangely, corresponds to zero feedback.

ZynAddSubFX Phaser Seriously Hacked

Overhauled DSP engine in Zyn phaser and used shell to create a "Physically Informed Digital Model of an Analog FET Phaser". Don't forget to spit after chewing that mouthful.

When should you want to use Phaser instead?

A) Next to the pitch shifter, Analog Phaser is one of the most CPU intensive FX in Rakarrack. If you find your computer can't handle it, then Phaser is much more friendly to slow processors.

B) You just plain like the original Phaser Effect and want to use it.

C) When you want to use two phasers at the same time, you can chain Analog Phaser and Phaser together for some unique filtering sounds.

A Phaser works by adding a delay to a certain group of frequencies and shifting the filter that selects the group of frequencies up and down the spectrum. When mixed with the original input, some of the delayed parts of the signal cancel with the the original (destructive interference). The technical term for this is 'notch filter'. For every two phase stages, a new notch appears in the spectrum, so for 4 stages and up we have a 'comb filter' because a plot of the frequency response begins to look like something you use to brush your hair as the number of phase stages increases.

If you don't care about what's under the hood, skip this paragraph -- To get the basic behavior of an Analog Phaser, the transfer function of a real analog phaser all-pass filter stage was computed and transformed into a discrete-time equivalent transfer function, then broken down into the resulting numerical computation algorithm. These all-pass filter stages were chained together in sequence of the overall analog circuit. Five or six different schematics were referenced in the design of this model to see what different units do to achieve a certain sound. The resulting 'Virtual Circuit' is an original creation. Finally, some of the non-perfect physical components and distortion were modeled in a simplistic way to add a certain warmth to the effect without sinking your CPU into an ice age.

Wet/Dry: At zero, both the phase shifted signal and input signal are mixed equally. This creates the deepest notch filter. Why would you want to mix all Wet? Modulating the all-pass filter stages can create frequency shifting at faster modulation rates. This produces a chorus-like effect, and you may want the frequency bending without the notch filtering. Think UniVibe.

LFO Type: More or less self explanatory. The Barber Pole setting will disappoint you if you're looking for a true barber pole phaser. This is simply an arrangement of multiple ramps modulating the phaser, so you hear the thump every time the ramp starts over. It's a pleasant enough effect at really slow rates, and creates something interesting at very high rates. In between it's annoying :) .

Tempo: LFO frequency.

Depth: How deep the phaser can sweep. It's an LFO offset. 64 is dead center. If you want it to stay in the high range, set this higher than 64. If you want the phaser to spend more time carving out the lower frequency range, set this to some thing less than 64.

Width: This is how far the LFO travels. It's the LFO amplitude.

Fb: Feedback. Usually named 'Regen' on a phaser stompbox. For deeper notches keep Fb negative for even number stages and positive for odd number stages.

Mismatch: FET Phasers suffer from the manufacturing process of JFET transistors. These things generally vary over a very wide range of properties per batch. What does this mean to a Phaser? It means that the frequency where 45 degrees phase shift occurs varies from stage to stage. This mismatch makes the notches wider, but less deep. This parameter best aligns with reality at a setting of 5 to 10. Large settings may be used to obtain a better 'Vibe sound.

Distort: FETs used as variable resistors are only linear over a certain range. This nonlinearity adds harmonics to the processed signal and somewhat warps the frequency response. This is a very subtle effect, but worthwhile to be able to zero it if you don't want it modeled.

St. Diff: Stereo Difference. Delay the LFO in the right or left channel. This combined with the panner effect can make it sound like something is twirling in an elliptical orbit around your head.

Stages: First order all-pass filter networks to be chained together. Set to one for a high pass filter (or low pass filter with subtract checked). 4 stages is very typical in the average stompbox. Bi-mode and Tri-mode phasers include switchable 6 and 8 stage filters. It's rare to see an analog phaser with more than 8 stages in stompbox form because parts get expensive, and noise becomes an expensive design problem to mitigate. In the digital world, adding more stages is just another time through the filter loop if you can spare the CPU time.

Subtract: Subtract wet from dry instead of add.

Hyper: Flattens out the lower end of the LFO. Used with a Tri wave, this emulates the "Hyper Sine" found in some analog phaser pedals. It was a clever way that analog filter designers contrived to make the LFO behave according to the human's perception of frequency. To our ears, musical pitch increases exponentially with frequency.

Adapted from the ZynAddSubFX Chorus

Intense: Increase effect intensity. A digital Flanger or Chorus uses a delay line with several hundred to several thousand taps. When it changes taps to modulate the delay time, there is a discontinuous change in the signal called "zipper noise". To eliminate this, fractional delay times need to be calculated in order to smoothly transition from one delay line tap to the next. The default Rakarrack Phaser and Chorus use linear interpolation. This mode makes use of a Lagrange interpolation polynomial to estimate fractional delay times and some other tricks to maximize the flanger depth and flatten the frequency response on the chorus. Adding the mode switch adds the feature without breaking old presets.

Adapted from the ZynAddSubFX Reverb

Adapted from the ZynAddSubFX Equalizer

Effect using the ZynAddSubFx Equalizer engine

This applies an approximation of certain speakers, cabinets,speaker/cabinet combinations according to publicly available frequency plots. For those wanting convolution-based cabinet modeling you can use Convolotron, but this cost a little bit of CPU resources.

For the stompbox minded people, this is a stereo tremolo effect. Extra Stereo setting creates a more spacious stereo imaging effect.

Intelligent Harmonizer Explained

Rakarrack harmonizer effect use the audio engine of the smbPtichShifter.cpp located at http://www.dspdimension.com. In order to save CPU use only a mono pitch-shifter in the lowest quality available, you can change this on the program Settings window, but only for a low quality values because high quality ones use too much CPU. The audio signal converted to mono is send it to the pitch-shifter and returned to the two pole Peak filter, panned and send it to both channels L/R. The rakarrack Harmonizer has 3 modes:

Normal Mode

Is how a normal pitchshifter run, the pitch ratio is fixed and can be selected in the Interval effect parameter of the effect.

Select Mode

In this mode the pitch ratio is variable, this value depends of the selected Interval parameter, and the Note and Chord effect parameters. Rakarrack then recognize the audio note played (Only "melodies" monophonic data can be played in this mode) and modifies the pitch ratio in order to do a musical harmonization with the Tonality/Chord selected in the parameters, of course the twelve tonality's are available and 33 chords:

,6,Maj7,lyd,Maj(9),Maj7(9),6/9,+,m,m6,m7,m7(b5),m9,m7(9),m7(11)
,m(Maj7),m(Maj7)(9),dim,dim7,7,7(Sus4),7(b5),7(9),7(#11),7(13),7(b9)
,7(b13),7(#9),+Maj7,+7,1+8,1+5,(Sus4),2

MIDI Mode

This mode run in the same way as Select Mode but the Tonality/Chord is recognized via MIDI notes, the MIDI chord recognizer recognizes all the above chords plus all the inversions and bass changed chords, also has memory, they use the chord just another chord is send it and recognized. The MIDI channel can be selected in the Settings window, then the Harmonizer adjust the pitch ratio with the audio note recognized, played by the user, and the Tonality/Chord received via MIDI. (Sequencer track ... )

Musical Delay Explained

The musical delay effect is a dual line delay, the word "musical" is due that you can adjust the delay time in both lines in a musical terms, with Tempo effect parameter, and the Delay1, Delay2, Delay3 effect parameters.

The Delay lines are measured in fractions of quarter notes at the Tempo selected. That means 1/2 is an Eighth note and 1/4 is a Sixteenth note. (1,1/2,1/3,1/4,1/5,1/6) are the possible values, that include eighth triplets, etc.

The center delay parameter, is the delay between the two delay lines, and is the only one can be set equal to zero.

The Tempo effect parameter value range is big (10~480) that's for admit half or double song Tempo in order to obtain largest or shortest delays.

Off course you have Gain, Pan and Feedback parameters for each delay line in order to adjust the desired effect.

Distortion using resonant filter engine and standard Distortion waveshaping functions. Controls are mostly like Distortion module, but includes "Color" to add resonance to the waveshaper signal. This allows for a "bigger" sound that is useful for fuzzes and so on. Your imagination will take it where you want. Addition of this module also allows the user to have up to 3 simultaneous distortion modules, 2 of with can have sub-octave modulation.

New Effect based on Gate, Steve Harris LADSPA plugin. Only use the noise gate when you really need it. This can be a life saver, but it can also be a source of great frustration if improperly configured.

Adapted from the ZynAddSubFX DynamicFilter

By now you're saying, "Ok, I get the Pan, Freq, Rnd, St. Dif, and Depth thing, but what the..."

In addition to use as Auto-wah (LFO-modulated Wah-Wah), this is also an envelope filter.

Currently, here are the three parameters related to envelope control:

Amp S.: Amplitude sensitivity. This is like the sweep range. If sensitivity is set low, you have to pluck really hard to make the filter move. If set high, then the filter will sweep to the limit.

Amp. S.I.: This stands for "Amplitude Sensitivity Inverse", but is applied in a somewhat interesting way. This slider offsets the base frequency for the wah wah filter resonance. At the same time, it acts as a logical check for which direction to sweep. When this is set less than 64, the filter responds to plucking by sweeping upward in frequency. When this is set higher than 64, then the filter responds to plucking by sweeping downward in frequency. Tip: If you want to control wah-wah with a MIDI foot pedal, this is the parameter you will want to map. For pure pedal-controlled wah-wah, set Amp. S and Depth to zero.

Smooth: How much to smooth the signal envelope. At a low level, the sound is like a bubble maker. At a high level, the filter responds very slowly to signal attack.

AlienWah - Created by our hero, Paul Nasca Octavian originally for the ZynAddSubFX synthesizer.
AlienWah fits into the comb filter category, similar to phasers and flangers.  AlienWah amplitude modulates two delay lines each with a phase separation of "Phase", then feeds them into each other in a somewhat circular fashion making each delay line a cascade of low order digital filters. In this manner it is similar to a Phaser.  It's sort of like a digital low-pass filter that spirals down a pair of delay lines causing an interesting interference pattern.  Some people may think of this effect as a "Flaser".

For the less technically minded, the AlienWah configuration produces mostly vocal sounds with a strange timbre sounding much like sci-fi depictions of a language spoken by invading space aliens, or some strange creature found near the center of the earth.

Even though some of the more extreme settings create weird alien sounds, it can also be configured for interesting subtle comb filtering sounds for shaping a rather beautiful timbre onto an instrument.  It is a truly amazing effect.

Here is a description of the parameters:

Wet/Dry: Being a delay-based effect, Wet/Dry does more than simply mix Wet/Dry sounds.  Destructive interference occurs between the effected signal and clean so a proper blend can be used also to voice the effect.

Pan: Put effect more to the left or right channel.  Used in conjunction with Wet/Dry mix, one can voice the effect differently between left right.

Tempo: LFO rate

Rnd: Add random noise to LFO to make it sound somewhat more "analog".

LFO Type: Self explanatory

Phase: This is a very interesting parameter.  It sets the modulation offset between the two delay lines.  On either extreme it has a lower "closed mouth" sound, while in the middle it creates more of an "open mouth" sound.  This parameter "centers" the LFO around the point it is set.

St. df: Stereo Difference between left & right LFO.  A setting of 64 has both LFOs in phase between left/right.  This parameter used with L/R Cr. is like having a second "Phase" parameter, so more interesting interaction can be achieved by using this control.  It's more than what meets the eye.

Depth: This would be better termed "width" as this parameter adjusts the amplitude of the LFO.

Delay: Length of the delay line in samples.  In a sense, this also is the filter order.  The longer the delay time, the deeper the sound of the filter.

Fb: Feedback.  Less than 64 is negative feedback, more than 64 is positive feedback.  64 corresponds to zero feedback. Without feedback, the effect is rather dull.  It's the perpetual regeneration along the delay line that creates  the interesting phase cancellations and resonances that sound alien.

L/R Cr.: Left/Right Crossing.  Amount of left to mix into right and right into left.  Less than zero subtracts left from right, right from left, while greater than zero adds them.

New Effect based on Valve, Steve Harris LADSPA plugin, filters, harmonic enhancer and some extra distortion where added.

Another Flanger alternative.
Basic operation of a flanger: Delay the incoming signal and add it to the input signal. This creates destructive interference in a similar manner to a Phaser, but the notch filter frequencies are evenly spaced from the lowest notch up to infinity (in theory). Dual Flange picks two different delays and adds them to the input signal. The two separately delayed paths have a spacing relationship that follows the other by a percentage, and not a constant offset. At maximum wet, these two delay "taps" are added equally, thus the notches between the two are infinitely deep, and the effect is at its maximum severity. Mixing in the clean signal reduces the depth of the notches, but also creates a complex and interesting frequency response. The result is a unique color to the tone not available by a standard flanger.

Wet/Dry: Mix delayed signal with Dry

Pan:

L/R.Cr:

Depth: Frequency (Hz) of the lowest notch frequency on lowest end of sweep range

Width: LFO sweep width, deviation measured in Hz.

Offset: Percent difference in delay between delay A and delay B.

Fb: Feedback. 0 is none, negative and positive.

LPF: Damp the delay line. This makes the high frequency notches more shallow, thus the sound approaches that of a Phaser or other mellow sweeping comb filter

Subtract: Subtract delayed signal instead of add

Th. zero: Through Zero. Delay Dry signal to 1/2 way between high and low delay deviations of delayed signal.

Tempo: Sweep speed

St.df: Stereo Difference between LFO's

LFO Type:

Rnd: Random. Add random noise to LFO to help reduce unmusical effects of linear interpolation between samples.

New Effect based on Ring, Steve Harris LADSPA plugin. Auto Freq recognizes the frequency of the note played (Only monophonic sorry ..) in order that you can use Ring as a simple monophonic synthesizer.

New Effect based on Harmonic, Steve Harris LADSPA plugin.

Multi Band Distortion, use the Cross1 and Cross2 to select the ranges of each band.

20Hz -> Cross1 = Low
Cross1 -> Cross2 = Mid
Cross2 -> 20KHz = High

You can select a waveshape type for each band.

"Arpie" is a pet name for Arpeggiator. This effect is a modified version of Echo. The delay line is played back at higher rates to create arpeggios at octave intervals. You can select a preset sequence, then adjust the number of steps to repeat from the sequence to gain a wide variety of patterns. Arpie can be adjusted to achieve thick and beautiful soundscapes as well as very edgy electronic trance and techno sounds. Of course, with slower tempos such that several phrases are repeated one can get the tempo-doubling side effect to lay out paint-pealing leads without hardly moving your fingers.

Wet/Dry: Amount of arpeggiated accompaniment in the mix

Arpe's: Mix pitch shifted delay line with normal delay

Pan:

Tempo: Measured in beats per minute

Subdivision: Ratio to subdivide measure. A setting of 1 is equal to 1 measure at Tempo. For example, 1/4 means the length of the delay line is equal to a quarter note at the selected tempo. The sequence will jump 4 steps per measure.

LRdl.: Same as for Echo

L/R.Cr: Same as for Echo

The primary purpose for this effect is to offer a differently natured Noise Gate than the "Noise Gate" module. The turn-on characteristic is modeled after that of a BJT transistor, thus it behaves more similarly to an analog noise gate. It is likely that Expander uses less CPU resources than Noise Gate, but neither are very resource intensive modules.

The basic behavior is the louder you play, the louder it gets until it gets pinned at the maximum "Level". As a noise gate, the volume is exponentially increased in the lower volume levels until Threshold, at which point the gate is entirely open.

To use the effect as an expander, it is best to set the threshold at higher levels, and shape at moderate levels (less than 10). Short attack/release times in this mode will cause it to work much like one would expect from a dynamics expander. In this mode it is essentially a compressor in reverse.

To take the idea of expander a little further, one can set long attack times and short release times to achieve a string swell effect akin to adjusting the volume pot on the guitar...only the effect does it automatically for you.

Finally, the filters may be used for a bit of frequency shaping before a distortion effect, or used in front of the Compressor to help brighten the tone.

A. Time: Attack Time, time to open the gate

R. Time: Release Time, time it takes the gate to return to off state

Shape: Gate transition shape. Large number means gate turns on suddenly at threshold. Small number means gate gradually turns on Noise suppression is best at higher numbers, but the gradual expansion above the threshold can be a more musical way for gate to turn on.

Thrhold: Threshold. The gate is reducing signal volume below this level. Moderate threshold setting with smaller number for shape is useful for string swell effects.

Level: Boost the output. 1 ... 127 represents a fractional increase in volume from 0dB to 20dB

HPF: High Pass Filter

LPF: Low Pass Filter

Effect based on Stereo Shuffling paper by Michael Gerzon. That convert L/R signals to M/S, Mid and Side. You can equalize one of this bands with a parametric four band, the Rev selector select the M/S band to equalize, ON = S, OFF = M. That produce interesting spatial sound, also can be used to remove certain frequencies on the M/S channels.

A different type of view of the Analog Phaser effect revealed a basic structure supporting high-order filters with adjustable number of filter stages ...most of the work was already done, so thanks again to Paul Nasca for this piece of code from the original ZynAddSubFX Phaser.

Starting with a phaser (as opposed to wah-wah or EQ) was inspired by a phaser circuit modification suggested by Mark Hammer, a frequent forum poster at diystompboxes.com/smfforum. This modification allowed two phase stages to be converted to low-pass stages with a switch in an analog phaser. This is the type of effect marketed as a "phase-wah". Of course, converting all of these phase stages to low pass stages makes this circuit look an awful lot like what is found inside the old Korg Delta DL-50 synthesizer.

Converting a normal OTA phaser into a synth filter would be a relatively ugly hack in the circuit bending world, but a simple exercise in software bending, and with a more elegant final product. With a little bit of experimentation, I had a filter sounding much like something found in an analog synthesizer. Thinking back to an analog envelope filter I once built, I determined high pass filter stages would also be of great utility. The final result: A very flexible filter module able to accomplish many flavors of low pass, band pass and high pass filter shapes. Possibilities range anywhere from simple wah-wah sounds to definite mushy filters from synthland.

You can operate up to 12 high-pass and 12 low pass stages simultaneously. At present, the user is protected from resonance near instability. If you are a synth filter lover and want to be able to make your filter unstable, then it is a very minor adjustment to the source code if you wish to internally amp up the feedback range to allow this. Finally, the SynthFilter comes with wet/dry mix so you can accomplish interesting phase-wah sounds.

Wet/Dry: Like many other filtering effects, it is not as simple as Wet/Dry. Destructive interference happens between the wet and dry signals to varying degrees as this slider is adjusted, so new & interesting notch and comb filtering effects evolve as you move this from wet to dry.

Distort: This adds some nonlinear response to the filter. In addition to adding audible distortion, this parameter has the effect of making the filter less resonant. Set to zero for a "clean" filter sound.

Tempo: The filter can be modulated by an LFO. This sets the rate.

LFO Type: The shape of LFO you wish to use.

Subtr.: Subtract. Make output of filter negative, thus changing which parts of the signal will cancel when using Wet/Dry mix.

St. df: Stereo Difference. Delay between LFO's for left and right channels.

Width: Width of the LFO, also think of this parameter as LFO Amplitude.

Fb: Feedback amount. Positive or negative values make the feedback positive or negative. This parameter makes the filter more resonant.

LPF Stg: Number of first order low pass filter stages.

HPF Stg: Number of first order high pass filter stages. HPF and LPF are series processed, so combining the two creates a band pass filter.

Depth: Filter "start frequency". Set Width and E. Sens to zero then use this to manually sweep the filter with a MIDI message. More simply, use this parameter with a MIDI expression pedal to make Synthfilter behave as an interesting type of wahwah.

E. Sens: Envelope Sensitivity. This adjusts how much the filter sweep responds to the dynamics of the input signal. Numbers greater than zero cause the filter to sweep upward as your playing gets louder. Numbers less than zero cause the filter to sweep downward as your playing gets louder. Set to zero for no effect. Use "Depth" to adjust the filter start frequency. For example, if filter is to sweep downward, set Depth to a large number and E. Sens to a negative number.

A. Time: Attack Time. Sets how quickly the envelope detector responds to input dynamics.

R. Time: Release Time. Sets amount of time for envelope detector to release the build-up from loud dynamics.

*Attack and Release are measured in milliseconds.

*Attack and Release times have no effect if E. Sens is set to zero.

Offset: Separation between High Pass and Low Pass filters. If both HP and LP filters have stages 1 or more, this adjusts the bandwidth of the resulting bandpass filter. Set to zero, a bandpass filter with a narrow peak will result, or set large, the bandpass filter will be wider.

Four Bands volume modulated with two LFOs. You can select for each LFO the type, Tempo(speed) and LR difference (St.df). Cross points are to defining the band frequency range.

20-> Cross1 = Low
Cross1->Cross2 = Mid Low
Cross2->Cross3 = High Low
Cross3->26 KHz = High

The Combi choice select the LFO for each band

1 = LFO1
2 = LFO2
o = Bypass
x = Disable Band.

Convolotron A module optimized for physical speaker/cabinet modeling.

A lame EE joke: "If you do that once, I will convolve your face with a Dirac Delta. If you do it again, I will convolve your head with an impulse train."

Some jargon associated with this effect:

IR Impulse Response. Filters are often classified according to their impulse response. An impulse response is the output of the filter (or physical system) resulting from application of a short-duration, one-time, "impulse" of energy. An audio signal can be constructed mathematically as a series of such impulses separated by an infinitesimal period of time, and scaled by a certain factor representing the real-time amplitude of the signal. If one impulse is applied to a filter, the filter begins to react with its characteristic impulse response. Then if another impulse is applied a little later, the same impulse response is invoked and is added to the first. If we think of the audio signal as a rapid succession of impulses, then the result of adding up all the impulse responses will reconstruct what the physical system will do in response to the input signal. This process is called convolution.

Convolution. A process of successively adding the impulse responses for any signal that has been or will be applied to the system. In the case of realtime audio processing the impulse responses are causal in nature, so we can ignore the future and apply the effects only for what has already occurred.

Convolotron. A rakarrack module that takes an audio file as an IR, and convolves it with the input signal. This signal may be an IR recorded from an amplifier, a microphone, a cheap computer speaker system, horns, bells, gongs, reverb...kicking a door...

Some caveats: Convolotron is very CPU hungry. Rakarrack does straight-forward time-domain convolution. Another Linux software application to compare is JConv, which is not as CPU hungry because there are some frequency domain math tricks which allow the user to sacrifice some quality to reduce processing load. As a result, Rakarrack requires some "horsepower", but is better suited to amplifier/cabinet modeling assuming your CPU can handle it.

Parameters

Wet/Dry: Self explanatory. For amp responses, most of the time you will want this all wet.

Pan: The effect is processed in mono (L & R are mixed on the input). Pan routes the output left or right.

Level: The actual IR file will determine the gain of the effect. This helps you adjust to a good level.

Damp: Darkens the sound of the IR.

Fb: Feedback. Pretty self explanatory.

Length: Measured in milliseconds. This tells Convolotron where to truncate the IR. Set this to the maximum value your CPU can afford without Xruns. A longer Length will result in a more true model of the system being convolved.

Safe Mode: Automatically limits Length internally based on sample rate and info from your /proc/cpuinfo to help prevent freezing your computer.

Preset: Several amp IR's are available by default for your convenience.

User: You may download or generate your own IR's. You can even do silly things like loading ring tones or segments of a song (if you like to experiment). Must be wav format, and keep in mind that convolotron only uses the first "Length" milliseconds of the file...so if you have a file with 100ms silence at the beginning, you will get nothing but silence on the output.

Looper

Autism: A pervasive developmental disorder characterized by severe deficits in social interaction and communication, by an extremely limited range of activities and interests, and often by the presence of repetitive, stereotyped behaviors.
(The American Heritage® Dictionary of the English Language, Fourth Edition copyright ©2000 by Houghton Mifflin Company)

It's a joke...

Wet/Dry: Balance the mix of the recorded loop with the incoming signal to the output.

Level 1: Volume level of track 1. '64' is unity. Higher is louder than the original signal, lower is more quiet.

Level 2: Volume level of track 2.

Tempo: Set the tempo of the song. This control responds to TapTempo. However, time stretching is not used in the Looper, so changing Tempo during recording, or on a recorded loop will have no effect on the recorded tempo.

Time Sig: Time Signature. Use this to set the beat of the song. The loop will be quantized to the nearest measure.

MS: Metronome Sound. You can select N (Normal mode) to hear a "mark" on every first beat of the measure, then a lower "tick" for the remaining beats, or H (High tick) to hear the "mark" sound on every beat, or L (Low tick) to hear the lower tick sound on every beat.

Reverse: Play loop in reverse.

Auto Play: If this is checked, the looper will begin to play automatically after recording the first time. It also will cause the Record button to start play automatically. Disable this option to synchronize Looper with Jack Transport

Play/Stop: What more needs to be said?

Pause: Pauses whatever is happening.

Record: Unless AutoPlay is selected, Record only arms recording. Play/Stop will set the tape rolling. First time, it begins recording and will continue to record until you push record again. At this point you have defined the length of the loop. If you press record again, everything you play will be overdubbed onto the active track(s). To erase the track, you need to have the track active and select "clear". This operates on the active track only if R1 or R2 (corresponding to the track) is selected. This allows you to record on track 2 while hearing track 1.

R1, R2: Enable recording on track 1 and track 2, respectively. If neither are checked, then the record button will not do anything.

Track 1, 2: Enable Play, Stop, Record on the selected track. Again, both the track and the R button need to be selected to record on the specified track.

Clear: Erase the selected track and set length to zero.

Lnk: Link track 1 and track 2 with the same length. Activate this if you want track 1 and 2 to play at the same time. Deactivate if you want something like a verse-chorus-verse-chorus song structure where track 1 is the chorus part, and track 2 is the verse part.

M: Enable Metronome. Note Looper has its own metronome so it is able to synchronize with your play/pause commands. Pressing play resets the metronome to beat 1, no matter where it is in the measure. You will probably confuse yourself if you try to use the master metronome and Looper metronome simultaneously.

Example sequence of operation:
Set Wet/Dry to 0, Level 1 to 64, Auto Play Active, Track 1 active, R1 active. Press record and begin to play into the loop. When finished, press record again, and your loop will immediately start up from the beginning. Play along, or press record again to add more parts to the loop. Harmonizer with Octave down for bass simulator can be of assistance. MuTroMojo and Synthfilter with reverse envelope sweep can be of assistance for bongo drum sounds.

MutroMojo

The name is derived from two well known effects using the same filter topology: Mutron, Mojotron. For those not well accustomed to classic analog pedal effects, these were envelope controlled filters built upon the state-variable-filter. This filter structure has three "states" which can be extracted: High pass, Low pass, Band pass. This is easy to implement digitally, and is much more efficient for the processor than to compute the three filter bands as separate filters.

MuTroMojo allows you to adjust the three bands at whatever levels you want. This flexibility allows anything from various wahwah effects to synth filter effects, and possibly even sounds of the moogerfooger (though Synthfilter structure is better suited to "moog-like" sounds). Like any good Rakarrack effect, it has an LFO synchronized to Tempo. This means Tap Tempo will set the rate of the LFO effect.

Additionally, you have the envelope follower applied with the E. Sens parameter, and smoothed by Smooth. Here's the breakdown:

Wet/Dry: It's obvious that this mixes wet and dry signals, but this interacts like a phaser or flanger. By mixing some Dry into higher order filters (increase Stages) you can create notches in the frequency response. With individualized control over HP, BP, and LP filter bands, you can create some interesting comb filter responses. This is more than simply "amount of effect".

LP, BP, HP: Low Pass, Band Pass, High Pass, respectively. Greater than zero mixes the output in phase, less than zero mixes the output out of phase. This can be alternated to create different types of destructive interference between filter bands. For the more technically minded, you can dust off your filter cookbook and tune in to a wide variety of filter responses. This can also be used as a sort of parametric EQ or tone stack.

Stg: Stages. Each filter stage is a second order filter. "Stg" sets the number of filters in series. The number of stages will improve the amount of rejection above/below the cutoff. A tip: You can use this module as a compromise for up sampling quality when using up sampling. For example set Wah and Range to maximum levels with LP at 32 and BP, HP at 0. With this configuration you have a 12dB/Octave low pass filter set at about 6kHz. For clean electric guitar, you won't lose much of the original signal because most guitar pickups form such a filter at 5kHz. Then the upsampling can be set to Zero Order Hold, and this filter will smooth it out. To improve quality, increase the MuTroMojo Stages. In this mode it acts as an "Interpolating Filter", which is able to reconstruct the missing information between samples. The higher quality upsampling options do the same thing, but they use much higher order Sinc interpolation filters and try to maximize bandwidth. Unfortunately these higher order filters use more CPU, and they also have bad transient responses. If you can sacrifice the bandwidth, you can use lower order filters for nearly the same result. This is where the MuTroMojo can give you more choice with whether you want to trade bandwidth or CPU usage while balancing transient characteristics. If none of that made sense, then maybe you can ask on the rakarrack IRC channel for a better explanation...or look up sample rate conversion and interpolation for discussion of the trade-offs and benefits. Otherwise, just experiment with settings that sound good to you, and forget about the technical details of why.

Width: Deviation of the LFO. This describes how far it sweeps.

Tempo: LFO Rate expressed in beats per minute. This is synchronized when Tap Tempo is active.

Res.: Resonance. Increasing this makes a higher and more narrow peak at the filter resonant frequency.

Range: Upper bound on the filter sweep range.

Wah: Sweeps the filter resonant frequency from the minimum bound (hard coded in each preset) up to the upper bound (set by Range). This is the parameter you would want to map to a MIDI expression pedal for a wah wah effect.

E. Sens: Amount that filter sweep responds to input signal dynamics. Positive number means the filter center frequency sweeps upwared, negative means it sweeps downward. Zero means the filter does not respond to dynamics of the input.

Smooth: Smooth the envelope detector shape. This has no effect unless you have E. Sens set to something other than zero. If E. Sens is set to a useful level, then the Smooth parameter adjusts how quickly the filter center frequency responds to changes in signal dynamics.

Echoverse

This effect is intended to be an "in your face" wall of sound.
This has been adapted from the normal "Echo" effect, but with some modifications that help improve the sound and behavior. With Wet/Dry mix, you will notice it is possible to get 100% wet (echoes only). The reverse is a little more "clean". More importantly, this effect synchronizes to master Tempo, and is responsive to TapTempo control.

SubDivision allows you to divide delay times to 2,3,4... beats per measure. Also added is "Extra Stereo" effect that makes the effect come alive in 3D space (assuming you're amplifying or recording in stereo).

Wet/Dry: Direct/Reverberant mix. It's like changing the distance you are located from the sound source.

Reverse: Same as Echo

Pan: Move the effect to the left or right

Tempo: Beats per minute. This is synchronized to TapTempo when active.

LRdl.: Difference between left channel and right channel delay times.

Fb.: Feedback. Other descriptive words are Repeat, Regen or Tail.

SubDivision: Delay time can by synchronized to the length of a musical measure subdivision. For example, the image to the left indicates delay time is set to 1/2 notes of a song played at a Tempo of 90 beats per minute.

Damp: High frequency damping.

E.S.: Extra Stereo. This is the amount of the effect applied.

Angle: This is the angle from which the sound appears to be coming.

Coil Crafter - Pickup Emulation/Converter

This unit contains two filters tuned to typical pickup frequency responses. Origin is used to match the pickup you're using. This filter "undoes", or reverses, the frequency response of your source pickup. "Destiny" is what kind of pickups you want it to sound like is in your guitar. This applies the frequency response of the pickups selected.

Two caveats:

A) This does not make your Gibson Les Paul sound like a strat. It makes it sound like you put single coil pickups in your Les Paul. The Tone control helps to take off some of the lows so you can somewhat emulate different guitar body types.

B) You have to adjust parameters for the "Origin" pickup in order to cancel the effects of the pickups you are actually using. Without good electronics lab test equipment, this is largely guesswork tuned into range by your ears. If it sounds good, then who cares if it's technically right.

Gain: Volume control

Tone: Simple low cut

Origin: Pickups you are using. This will "undo" your pickup's response.

Freq1: Resonant frequency of origin pickup.

Q1: Resonance of origin pickup.

Destiny: This is the type of pickup you wish to emulate.

Freq2: Resonant frequency of desired pickup response.

Q2: Resonance of desired pickup.

Pos: Uses harmonic exciter to emulate switching to neck position pickup.

Low Shelf Tone Booster

This is a simple Low Shelf Filter that can be used as Tone Booster, is good to put after Distortion effects in order to adjust the distortion tone or cut undesired freqs. Also with a really hard gain you can obtain a nice distortion in combination with the "invisible" Compressor/Limiter that rakarrack has on the end of the chain, if you try that be sure to low the main Output volume.

Vocoder

For those not already familiar with vocoders, this is the effect responsible for many robot sounds on several popular songs.

The basic structure is this:

Your voice from a microphone needs to be connected to the "Aux" input of Rakarrack. Second, you will have some sound source coming through Rakarrack's normal processing chain. This is likely a guitar or synthesizer. When you speak into the mic, you will hear the sound of your instrument with a robot-like voice. The best types of signals are constant and rich in frequency content. For example, distorted guitar chords are a good "carrier" for the Vocoder. Various synthesizer strings sounds or brass emulations are also good sources for the voice.

This particular implementation uses the analog Vocoder model. These units were implemented with several parallel filter banks. One filter bank was used on the voice, and the other on the instrument. The bank on the voice was used to "find out" what frequency bands are being used by the voice, then the power from each of these bands was used to scale the corresponding bands on the instrument so it has more or less power in these bands. Briefly put, the power in the signal of each frequency band from the mic was used to control the volume on each corresponding frequency band for the incoming instrument. Think of two 32 band equalizers. The first 32 band EQ has a mic plugged into it. As you speak into the mic, you can see the VU meter bars bounce on each corresponding band (if this is the kind of EQ that displays VU bars for every band). Imagine something being linked to the VU meter bars that can push the sliders up and down on the Equalizer used to process the instrument. This is what a vocoder does.

Rakarrack uses a 32-band filter bank to implement the effect. These are all second-order band pass filters spaced evenly on a logarithmic scale from 300Hz to 3600Hz, thus covering most of the important formant frequencies.

Additionally there are some basic hard-coded utilities to help in processing the voice channel:
Gate: Anything below -75dB is gated. You can "cheat" the gate with the Input level adjustment. If too much noise is getting through to the vocoder when you're not talking, you can decrease the input level until the noise floor disappears. Of course, if you have a noisy preamp, mic, lots of background noise...etc, this gives you poor signal/noise ratio, then there is little we can program into the application to help improve the dynamics :-(.

Compressor: A simple compressor is applied to the vocals with a ratio of approximately 2:1. It is implemented using add/subtract/multiply/divide instructions to avoid CPU costly log and pow functions. As a result, the compression curve looks more like saturation on a vacuum tube or JFET, and the compression ratio increases steadily with the input signal level. Consequently, it is a soft-knee compressor and lends itself well to vocal processing as a easy soft-knee compressor. Because of the way this compressor behaves, increasing the Input level is tantamount to decreasing the threshold and increasing ratio simultaneously. It is designed this way so it works effectively without the user needing to know anything about it. Incidentally, this same compressor is used to create the Sustainer effect.

The take-home message with regard to the "invisible" utilities is that the Input level parameter gives you a lot of power over the dynamics of the microphone input, and can help in getting a good consistent vocal sound onto the instrument being processed.

Here is the list of parameters:

Wet/Dry: We have been over this many times.

Pan: Move the processed signal to the left or right

Input: Microphone input level. Explained in detail above.

Muf.: "Muffle". Maybe more like "blur" or "smear". This increases the averaging time on the power coming into each frequency band. Adjusting to a large level has an effect of making it sound like the voice is in a very large reverberant room. Lower levels achieve a more articulate sound.

Q: Resonance of the filters. Increasing this number makes the filter bands more narrow. Extremely high values make the filters so resonant it sounds like metallic reverberation. This also "smears" the vocals. Unless you are going for extreme sounds, the best range is 65 to 90.

Ring: If you are going for extreme sounds, this is the ticket. This causes the voice from each mic band to be multiplied with each instrument band. The adjustment is basically modulation depth.

Level: Final output level.

Sustainer

A very simple "no frills" soft knee compressor good for making notes sustain. All you have to think about is "sustain" and output volume.

This module was inspired by, and is built upon, the vocal compressor used in the Vocoder. This compressor has a more "bright" sound than the normal compressor. Read up on Vocoder for more information about this compressor.

The Attack time is equal to the Release time, and this time is set to 50ms.

Gain: Gain recovery after the compressor. This is a simple volume control.

Sustain: This lowers the threshold and increases the ratio as the number increases. Adjust anywhere from mild compression to a more extreme "breathing" compression sound.

Sequence

8-step sequencer with multiple modes of operation. The sequencer controls effects that are hard-coded internally which can be activated by the Mode selector. To date the sequencer includes various modes for modulating a band-pass filter, amplitude, and there is also a pitch bending mode based on the code used in Harmonizer and Shifter.

Some of the uses include a sequenced tremolo (try Tremor mode) or something akin to the Sample/Hold filter (Stepper mode), or a sequenced wah wah (lineal).

Preset: Several built-in preset configurations exist to showcase the possibilities of the effect.

Wet/Dry: Nothing special here.

Sliders 1 - 8: Level of effect at each step in the sequence. If a filter, sliders set filter center frequency. If Tremor, sets volume of signal at the corresponding step in the sequence.

Tempo: Rate in Beats Per Minute at which the sequencer changes to the next step.

Q: Filter resonance. This may or may not do anything with modes that don't use a filter. For certain modes, it may be used to control a "hidden" parameter, so it is worthwhile to try it to see if it does something.

St. df: Stereo difference. This delays the right channel sequence from the left by the number of steps selected (1-8). This makes it sound as though the sound source is moving all around the room.

Range For most modes, it provides a choice ranges over which the sequence sliders control the parameter. For example, on a filter, it controls the maximum and minimum center frequencies.

Mode Select different modes of operation. The modes include:

Lineal: Filter & optional amplitude. Moves smoothly from one step to the next.

Up Down: Filter & optional amplitude. Returns to zero between steps, and steps up to the level adjusted in the sequence sliders.

Stepper: Filter & optional amplitude. Steps abruptly from one level to the next. This is the most like a Sample/Hold modulated filter.

Shifter: Frequency shifter. Bends smoothly from one frequency to the next

Tremor: Amplitude. This is a sequenced Tremolo that can be switched between two different stepping modes by the "Amp" switch.

Arpeggiator: Frequency shifter. Pass directly from one frequency to the next one, the semitone is adjusted by the sequence sliders, semitone = "slider value / 10".

Chorus: Frequency Shifter, this mode use small amount of pitch shifter in order to generate a chorus effect.With St.df=0 normal operation, with St.df=1 added Extra Stereo, with St.df=2, Q can control the pan.

Amp: Invoke amplitude modulation when available in the mode. In Frequency shifter modes Amp is used to switch down the frequency.

Finally, here is a reward for those who read the Help:

What is a sequencer effect without a random mode? There is an "Easter Egg" hidden in this effect that invokes a random sequence. Pull all the step sliders (1-8) to zero. If the sum of the values on these sliders is less than 4, then the Sequencer will step randomly. To return to a sequenced operation, bring one of the sliders to something greater than 4. At this point, the Sequence queue contains the randomly generated levels from the last complete cycle through the queue. In random mode, these are refreshed on each cycle so the sequence does not repeat, but once you leave random mode the current random sequence remains in the queue. You can "erase" each of these one-by-one by adjusting the corresponding slider.

Shifter Pitch Shifter

Pitch shifting effect built on the Harmonizer DSP engine. This flows with the general concept of a synthesizer pitch wheel or a whammy pedal.

This effect has the ability to shift pitch up or down an octave.
In Trigger mode, the effect will cause the pitch to bend up (or down) when the signal has exceeded the threshold. The rate for upward bending is set by Attack. Time to return to normal pitch after the signal has fallen below the threshold is set by Decay. A second mode is available to make use of the Whammy slider for a pedal controlled digital whammy effect.

Wet/Dry: Mix some of the unaffected signal to the output.

Int: Musical interval setting the maximum pitch deviation. In Trigger Mode with interval value set to "0", the real interval is set to "1" and can be controlled by the Whammy slider.

Gain: Overall effect volume.

Pan: Pan effect to left or right channel.

Attack: Sets how quickly the pitch bends upward when input has exceeded the threshold.

Decay: Sets how quickly the pitch returns to "zero" after signal is below threshold.

Thrshold: Level at which the signal will trigger pitch bending.

Down: Pitch bends up by default. Check this box to make pitch bend downward.

Whammy: If Whammy mode is selected, this acts as a smooth pitch bend from "zero" (in tune) to a maximum set by Interval. This is intended primarily for external MIDI controllers, whether it may be a MIDI foot pedal, pitch wheel, Ardour automation, or a MIDI sequencer software.

Mode: Sets mode of operation.

StompBox

A stand-alone stompbox emulator.

Between the various equalization tools and distortion modules, Rakarrack offers great flexibility to get a wide range of distortion sounds, but this method requires much experimentation, while some things are simply not possible. What if you are a person who wants to dial in a good sound quickly, and don't enjoy tweaking seemingly endless possibilities?

StompBox aims to simplify the process by giving the user a familiar "StompBox" interface with only "Level", "Gain", and a 3-band EQ, which generally covers all the features presented by a typical stompbox.

Additionally, most of these "Stomp Box" modes are physically informed models of actual stompboxes, with the most notable features derived from the circuit's schematic. Please enjoy these models as a unique creation of their own, and not as something expected to be a part-for-part replacement. It is my belief (as the developer) that these models have some convincing characteristics, but I am aware of many subtle things that are not taken into account. Further, there are many variants of a specific type of pedal on the market; for example, an Overdrive pedal is basically the same circuit for almost all models marketed as "overdrive". The differences are minor variations that the specific designer determined produced a better sound. Even though the names of the modes indicate one model or the other, the internal signal processing in Rakarrack includes variations of its own: some to counteract unfortunate digital artifacts and some applied as common "mods" that tweakers apply to their physical stompboxes.

Perhaps a good analogy would be that of an artistic painting versus a photograph. The Rakarrack stompbox models are more like an artistic painting, where the artist (programmer) looks at the object (schematic), "paint" (write code) the prominent features, then add details that make it become an identifiable image of the original object. In contrast, a photograph captures things exactly how they appear, with artifacts such as resolution and lighting/color distortions. Many commercial products attempt more of a "photographic" approach to stompbox modeling, where each part and component is modeled in detail. This of course, makes the CPU usage increase dramatically while IMHO little value is added in terms of making a *GOOD* sound (however they do make it much more true to the original). I think there would also be risk of legal issues if we attempted to model commercial stompboxes exactly and (especially) present them to our users in that form. For the most part you will be able to tell what stompboxes were used as a standard for a specific type of sound, as the names are relatively tell-tale. Hopefully we have kept things generic enough to avoid aggravating the industry. If any person of such authority takes issue, let it be known we are happy to make changes & remove things you feel are infringing on your patent or trademark rights. However, being free software, this is probably greater benefit in the form of free marketing for these industries.

Last, not all of the stompboxes modeled actually have a 3-band EQ. For these cases, one or two of the EQ bands is configured to emulate the EQ control on the pedal while the others have a configuration that matches the "theme" of the pedal, and they have no effect at "0" setting. Thus, any pedal model will have EQ controls that behave similarly to the original.

Now with all the explanation and disclaimers set aside, let's give some insight to what this thing is able to do. The first important parameter to understand is the "Mode" selector. This is where you get to select the flavor of stompbox distortion you wish to model:

Amp: This is a model twice derived from a tube preamp. The general circuit flow follows that of a typical overdrive pedal (Tubescreamer type), however the waveshaper uses dynamically modulated symmetry with a soft clipping function somewhat emulating the behavior of a vacuum tube. The reason it is considered twice derived is because the overdrive pedal originally used the frequency contouring of a tube preamp along with a soft diode clipper to (poorly) emulate the characteristics of an overdriven tube. Rakarrack "Amp" model therefore is a model of a model, but the clipping routine is much more dynamic and "alive" than the average diode clipper. The pre-emphasis filter was modified somewhat to pick up more of the lower end to make it better suited to fat blues lead tones. The 3-band EQ is constructed based on criteria derived partially from a typical tube amp tone stack blended with some of the character of the overdrive pedal tone control. If Low and Mid bands are left flat (zero), then the high band behaves like the OD pedal Tone control.

Grunge: This will explain itself to those who are familiar with the source model. The original has only high and low EQ bands. The Rakarrack model implements the mid band to allow for some mid scooping, or a mid boost if desired. Set mid at "zero" to get the most true behavior of the original.

Rat

Fat Cat: Rat and Fat Cat are minor variations on the same thing. In truth, the difference between the two is subtle. The EQ behaves like the original if low and mid bands are left flat (zero). The high band behaves as the original tone control. The mid band EQ is designed to work on frequencies associated with the nasal sound. The low band is very broad and can be adjusted to make the pedal sound very "beefy".

Dist+: The name says it all for any who know the original. The original has no tone control, so all three bands are fabricated. The EQ on this is of the same type as for the Rat model. To get the sound of the original pedal, leave mid and low flat, then push high to its maximum.

Death: A good mode for a raw-edged chainsaw guitar sound and death metal enthusiasts. One may easily guess the original model if it is stated that the internal circuit is identical to "Grunge", only some contouring modifications and 3 bands of EQ. This model, unlike the original, has a gain control.

Mid Elves Own: Say it out loud and it will will sound similar to how the original stompbox name sounds. This model excludes the sweepable mid-band EQ. Instead, the mid-band picks a good place and leaves it fixed there. It is an intentional design decision not to include the extra slider for mid sweep. This is to maintain the simplicity of a generic interface for all models. The Rakarrack Parametric EQ as well as MuTroMojo modules are capable of providing a sweepable mid scoop or boost if desired.

Fuzz: The first kind of distortion to come in a stompbox. It makes everything sound as you hear things while experiencing mild electrocution from the chassis of a poorly grounded amplifier. This does not capture any particular circuit, but rather summarizes an era (although elements of the various different fuzz boxes were considered in the design). The controls on this unit do not follow the pattern of the stompbox models as fuzz originally was a 1 or 2 knob wonder. Here is the guide:

Level: This does what it says

Gain: This is mostly consistent with the name. This makes it more distorted but at the same time the lows get cut from the input increasingly as gain is increased. This is similar to how the classic Fuzz Face behaves.

Low: This has nothing to do with the low frequency EQ. This controls the bias in the virtual "circuit". As you adjust up or down, the waveform gets clipped increasingly asymmetrically in one direction or the other. It has no effect at "0" setting, although the clipping characteristic is naturally asymmetric. This setting can make the fuzz sound rather "nasty", and is a good parameter to map to the ACI to change dynamically.

Mid: A mid-band EQ, but not like a normal mid-band EQ. This one operates on the input before the distortion, and is tuned to act like a stuck wah wah pedal. Settings less than zero cut mids from the input mix, making the fuzz sound a little more "woolly".

High This is a tone knob modeled according to the Big Muff tone control. Positive numbers cut lows while boosting highs. Negative numbers cut highs while boosting lows.

Reverbtron

A convolution Reverb and delay processor.

Reverbtron is based upon the engine of Convolotron (thus the name), but takes into consideration that time-domain convolution has expensive taste for CPU cycles. Convolving a reverb impulse response in real time is impossible without the use of special hardware dedicated to intensive DSP functions unless something in the impulse response is sacrificed for lower processing requirements.

In the case of Reverbtron, the full shape of the response it preserved from the head to the tail, but true replication of the frequency response is sacrificed. In short, the "magic" actually happens in the IR conversion utility used to convert the .wav IR file to a .rvb file for Rakarrack. Various elements of the waveform are considered, then the .rvb file is constructed as a list of pairs: Time, Reflection Amplitude. This file is actually a plain text file, and you may edit this file or generate your own custom files. In fact you could use plain utilities such as bash scripting, perl, octave, scilab, etc. Internally Reverbtron convolves this directly with the signal using the time index to determine at what delay amount to perform the multiplication. When Length is reduce, Reverbtron simply subsamples the file at even intervals. As one can see, the quality of the resultant reverb relies 100% on the construction of the .rvb file. This has the side effect for enabling future improvements to Reverbtron for users who install Rakarrack from their distribution repository. Then they only need to download .rvb files as the Rakarrack team improves the IR file processing.

A tip to reduce CPU usage to an amount that even some slower processors can easily manage is open the Rakarrack Preferences and set the internal sampling rate to a low rate, down to 8kHz. It may seem a low quality sound will result from using a samplerate of 8kHz, but with some magical mathematics in the downsampling and upsampling process it is possible to replicate the input signal perfectly in the upsampling routine, only band-limited to 1/2 the sample rate. In other words, the loss in quality is no worse than applying a steep low-pass filter at slightly below 1/2 the sample rate, assuming you are using one of the sinc interpolation up/down sampling selections. There will be non-musical artifacts resulting from zero-order hold or linear interpolation, although this may be acceptable to you if your CPU is too slow to handle the higher quality selections. Almost all selections of resampling will be faster than processing at a higher samplerate.

Here are the parameters:

Wet/Dry: Amount of dry signal to mix with the convolved signal. Notice that the IR file defines the maximum wetness. The Fade parameter fades the head end of the IR where the amount of dry is defined. Usually this works well by mixing all wet and setting fade until it appears there is silence at the beginning for a duration equal to the incoming instrument's note attack time.

Pan: Fade the output more to the left or right.

Level: Helps to compensate for louder and softer IR volume levels.

Damp: Reduces high frequency content in the signal. This can help to tame poorly behaved IR files, or simply to impose a different character on the IR.

Fb: Feedback. Feed output to input. Be careful as this has the same sensitivity and character as microphone feedback. Setting initial delay or fade to a modest level helps to create more of a regenerative sound instead of feedback. Damp can be used to tame the feedback as well.

Length: This is measured in the number of points to process. A length of 800 means that literally 800 reflections will be processed, but it also means 800 multiplication and addition instructions will happen for every frame. This is an indication of the amount of subsampling performed on the IR file. Contrary to intuition, a high number does not always make the reverb sound better. There can be multiple "sweet spots" along the range, even with numbers less than 300.

Stretch: Want your IR to reverberate for a longer period of time? Have a long one and want to shorten it? Stretch changes the time base. A negative number shortens the IR, a positive number makes it longer. You can stretch some IRs long enough to create good discrete echoes; and very complex ones at that.

I. Del: Initial Delay. This parameter delays the entire reverb response. This is a way of thumbing your nose at physics. You can create reverb effects that don't happen in the real physical world.

Fade: Reduce the level of the initial response of the IR. This can be used to increase the wet level on an IR that contains much of the direct response.

Diffusion: Diffuse the percussive echoes. Synthesizes an artificial Head Related Transfer Function (HRTF) and performs a second convolution on the output. Increasing the slider value increases the number of reflections in the HRTF, and will better diffuse the discrete echo sounds. For instruments with fast or percussive attacks this can be a very necessary control to help make the reverb more smooth and natural sounding.

ES: Extra Stereo. Emulate a spatial effect in stereo. This works approximately like this: Signal->LPF->Right->LPF->Delay->Left->FeedbackToSignal....and it just keeps feeding back on itself in a loop until it decays. The delay time is set internally by a guess at the room size based on properties of the IR file. The method assumes in a real room your ears will be most sensitive to reflections from the walls on the left and right side. Each time a sound reflects off a wall, generally it loses more high frequency content than low frequency content, and a reflection will be delayed and filtered by the time it passes your left ear, bounces off the right wall and returns to your right ear. Anybody with technical understanding will know this is a very general approximation, but the result is pleasant, and certainly creates an illusion of space when listening in stereo (particularly with headphones).

LPF: Low Pass Filter. Sets the cutoff point where high frequencies are damped in the ES option.

Safe: Limits Length regardless of the setting.

User: Browse for your own .rbv file. These files are generated by invoking the rakverb command on a .wav format IR file. "rakverb -i foo.wav" will generate a file called "foo.rbv" in the same directory. This is the file you would want to load into Reverbtron.

Preset: Rakarrack comes with several IR files already processed and ready to go so you don't need to search the web looking for impulse responses.

Echotron

Multi-tap delay with virtually unlimited taps (limited to 127, but if more than that is desired, use Reverbtron).  In the most simple case, assign the timing and spacing of delay taps in a simple text file, then load as a custom "user" file.  Tempo is normalized to 1 measure = 1 Second at 60 beats per minute.  Simply think of a "1.0" in the time column of the text file as "one measure".  For example, if you want 8th note echoes at Tempo, assign delays in multiples of 0.125.  Potential musical subdivisions are limited only by your creativity and ability to do fractional mathematics as pertains to musical rhythm and timing.  If you don't like math, but you like to experiment, then simply try a bunch of numbers less than 6 and see what happens.  Maybe someday we will make a GUI editor to generate the text files

A more advanced feature of Echotron is the assignment of filters in each delay line tap. In the text file, you can configure a state-variable filter with center frequency, resonance, and individually mix the 3 bands, as with MuTroMojo. Up to 32 filters may be configured in the text file. If you wish to bypass a filter in any line of the file, set Stages to 0 (zero). The first 32 occurrences of filters with stages greater than 0 will be processed. Any filter parameters defined after the first 32 are ignored.

You can assign left/right panning in the text file to create interesting rotational patterns.

There is a maximum delay time of 6 seconds, but there is no minimum delay time; for example, a delay time of zero is perfectly acceptable if using the filters to create a phaser or wah wah (be careful with the feedback parameter).  Other small delay times are possible for creating chorus and flanger effects.  You may also set the delay times the same for multiple taps if you wish to create a comb filter with the state variable filter, or even an 8-band equalizer is possible.

From the GUI, you can enable/disable the filters, enable/disable filter modulation, enable/disable delay line modulation, and all is synchronized to Tempo.  You may also limit the number of Taps processed from the file.

This effect is the swiss army tool for stereo ping-pong and rotational delays, flangers, phasers (thanks to state variable filter) and even stereo spatialization techniques.  Hopefully we have done a good job demonstrating the most noteworthy possibilities in the default files installed with Rakarrack.

Here is a description of the parameters:

Wet/Dry: Mix unprocessed signal with processed output.

Pan: Pan processed output to left or right channel.

Tempo: Beats per minute. Synchronizes with master TapTempo.

Damp: High frequency damping in the feedback loop.

Fb: Amount of Feedback (regeneration)

L/R Cr.: Amount of blending left & right channels. Less than zero means subtract left from right & right from left.  Greater than zero means adding left to right, right to left.  At +/-32 left & right are mixed 50/50.  At +/-64, left/right are completely swapped.

Width: Width of the LFO.  This adjusts the LFO amplitude.

Depth: Filter center frequency.  "0" means it is centered on the frequency designated in the text file.  >0 Shifts the filter up in frequency, <0 shifts them down.  This is a good parameter to assign to a MIDI expression pedal.

St. df: Sets stereo time difference between LFO right and left channels.

LFO Type: Select the modulation shape.

AF: Activate Filters.  If the box is checked, filters defined in the text file will be applied to the delay taps.

MF: Modulate Filters.  If the box is checked, modulation will be applied to the filters' cut-off frequencies.

MD: Modulate Delays.  If the box is checked, the delay line will be modulated (like a chorus or flanger).

#: Sets the number of taps to process sequentially from the text file.  For example, if you have a text file that defines 20 echoes, you can limit it to use the first 2 or 3 or....whatever you want up to 127.

File: Select from several files supplied by the Rakarrack team.  This will include a broad spectrum of uses so you can get started using Echotron without ever touching a text file.

User: So you want to do something you can't do with one of the default files.  You edit your own text file with the desired delay times, levels, and filter pattern, browse to this file, select, and voila! Your Echotron has morphed into a completely unique effect.

Text Files

Below is an example text file with explanation of each field and any caveats you may need to know:

Filter:  This field multiplies the tempo for the filter LFO modulation. For example, if you are playing a song at 80 beats per minute, but you wish to make the LFO half as fast as the echoes, you enter a 0.5 in this field to adjust the filter LFO speed to 40 beats per minute.

Delay:  This field multiplies tempo for the delay line modulation. Same concept as for filter, but this is the LFO applied to the delay line. Note there is not a field for delay time subdivision. This is because you define this by the times you put into the Time column, so such a field would be redundant.

Pan:  Ranges from -1.0 to 1.0.  less than 0.0 is pan left, greater is pan right.  0.0 puts the delay equally to left and right.  Anything magnitude of +/-1.0 or greater will be treated as extreme left or right.

Time: This is the real time at 60 beats per minute. An easier way to think of this as a musician is increments of measures. "1.0" means 1 measure at the tempo selected in the effect GUI. The above indicates a quarter note in the first line, and a 1/2 note in the second line.Ranges from -6.0 to 6.0

Level: how loud you want the echo to return.  1.0 returns it exactly as loud as it came it.  More makes it come back louder, less makes it softer.  You can use both positive and negative numbers.Ranges from -2.0 to 2.0

LP: Mix State variable Low Pass Filter amount.

BP: Band pass filter amount.

HP: High Pass Filter amount.

*For LP, BP, and HP when filters are activated, these also adjust the level.  If you don't want the filter to have an effect at a certain time, then set LP, BP, and HP all to 1.0.Ranges -2.0 to 2.0

Freq: Filter center frequency.Ranges 20.0 to 20000.0

Q: Filter resonance.Ranges 0.0 to 300.0

Stages: Number of filter stages.  This is virtually unlimited, but you can crash your CPU with the processing requirement if you aren't careful.Ranges 1 to 16

Caveats:

Always separate the fields with a <TAB>, and not spaces.  Echotron looks specifically for TAB separation.

You can enter as many lines as you want, but only the first 127 will be used by the program.

There are only 8 filters available.  After the 8th line, filter parameters are ignored by the program.

It is not wise to set Q values greater than 300, but there is no arbitrary limit.  At some high range the filter is likely to go unstable.

It is not wise to set Level greater in magnitude to +/-1.0.  There is some normalization internal to Echotron, but it is best to think of 1.0 as the volume knob's maximum range.

It is best to copy one of the default files distributed with Rakarrack and edit it as desired.  Always save with an extension .dly so you don't confuse it with other kinds of text files.

Stereo Harm

This is a Stereo Harmonizer, two voices in stereo, the SEL and MIDI functions are same as Harmonizer, please read the Harmonizer help,  in this effect Chrm L and Chrm R where added, that is Chroma, you can use Interval L/R set to "0" and modify this Chromas to reach a "Open" stereo chorus effect. This Chorma parameters can be used also in normal Mode.

In SEL and MIDI modes Chroma parameter doesent has effect, because the frequency is selected internally by the program.

CompBand

Is Four band Compressor, using the Compressor effect, the Wet/Dry parameter is included  for  mastering purposes,  probably if you use with a single instrument you will want to hear the Wet.

Four bands are availables  L(Low), ML (Mid Low), MH (Mid High), H (High). You can control the ratio and threshold for each one of this bands.

The Cross Sliders is to determine the frequency range of  each band in the form:

0-> Cross1 = Low

Cross1->Cross2 = Mid Low

Cross2->Cross3 = High Low

Cross3->26 KHz = High

Opticaltrem

Short explanation of Tremolo, for those who are "newbies" to guitar effects:  A tremolo is like an automatic volume knob.  You set the rate and depth, and it's like having somebody change the volume knob up and down at the set rate and amount set by depth.  No effects processor would be complete without a tremolo.  In prior versions of Rakarrack, there is the AutoPan function in the stereo Pan effect.  In Auto mode, the Pan is a tremolo, but is not a "classic" sounding tremolo.

This module emulates the tremolo found in classic guitar amplifiers (which also can be found in several stompbox tremolo units).  The reason for the name "Opticaltrem" is because of the photo-electric device that was used to change the electrical resistance of the signal path in the amplifier circuit.  The device in that configuration makes a tremolo effect when exposed to a pulsing light source.  Incidentally, this is the same type of component used in "Optical Compressors".  If  you have ever heard that term, now you know what it means.

Here is a more technical explanation:

The volume knob is nothing more than a variable resistor.  The volume knob varies the amount of signal current going into the next stage of the circuit.  In many old guitar amplifiers this effect is emulated by a variable resistance that can be changed by exposing it to light.  Usually the variable resistance element consisted of a light-proof container (opaque box) containing a Cds (Cadmium Sulfide) cell and a lamp, or LED for more modern units.  When Cds is exposed to light, its electrical resistance decreases, and thus its terminals behave like the resistance between the wiper and outer lug on a potentiometer (volume control).

If the response of the pulsing lamp was perfectly linear with the voltage from the LFO, and if the Cds cell response was perfectly linear with the light, then the effect would sound exactly the same as Pan.   As it is with most physical electrical devices, most things have interesting nonlinear properties.  Even more, there is a time dependency in the system.  It takes time for a lamp to heat up and turn on.  When current is removed, there is still some heat in the filament, and light fades out in a time-dependent way.  Then, the Cds cell itself has some "memory".  The resistance does not change instantly when light is applied and removed, nor does it work the same for charging and discharging.

Perhaps that explains why there is a need for a special "Opticaltrem" module in Rakarrack to emulate this behavior.  We hope this will bring to mind the sound of the tremolo in an old tube amp.

Depth: Amount of effect.  At full depth it is almost turning off entirely between pulses.

Tempo: Speed of pulsating sound.

Rnd: Adds some random noise to the LFO to emulate the imperfections of an analog LFO.

LFO Type: The shape you want it to use.

St.df: Stereo difference between left and right LFO's.

Pan: The effect can be panned to the left or to the right.

Vibe

Chorus/Phaser effect emulating a rotating speaker.

This effect was added at the request of a Rakarrack user wishing to have a software UniVibe. This effect is based on the original UniVibe circuit and mathematically models components most likely to contribute to the overall effect. Of course, the originals varied widely from one unit to the next due to some components being manufactured under less advanced process control. On these components we made our best guess and tuned by ear. This is the same situation for those attempting the same kind of thing with modern analog clones.

The sound of the original type of unit can be easily found by searching the internet for "Vibe effect demo". Its popularity among guitar players makes for the proliferation of clones and audio clip demonstrations, and ultimately an attempt by the Rakarrack team to emulate this. We hope you will find our version satisfactory and useful.

Things not available in the original units are the inclusion of stereo processing paths and feedback, not to mention the variety of LFO's standard to all Rakarrack modulation effects. Consequently, this effect can be taken to sonic destinations the original was never able to achieve. The feedback scheme applied in software, for example, would require a messy modification using several extra electronics components to properly apply to a real analog unit. In software this is more elegant and produces a lovely phasing effect.

Wet/Dry: The original 'Vibe had a switch allowing you to select "Chorus" or "Vibrato". Internal to the circuit, this switch only selected between different amounts of wet/dry in the final mix. The "Chorus" setting is a bit of a misnomer because it is actually more of a Phaser. A Wet/Dry of 0 (50/50) corresponds to the "Chorus" setting. A wet/dry setting of -64 (all wet) corresponds to the "Vibrato" setting of the original analog unit. Later clones of the circuit added the wet/dry mix as a pot so you could mix anything in between. Rakarrack applies this philosophy.

Width: Width of modulation (LFO) sweep.

Depth: How deep the modulation can go on the lowest end of the sweep. A small number will introduce more "thump" in the response.

Tempo: Speed of the LFO sweep

Rnd: Adds some randomized "noise" to the LFO to help make it sound less mechanical.

LFO: Type Modulation Shape

St. df: Amount of delay between left and right channel LFO's. At 0 or 127 the LFO's are 180 degrees out of phase. At 32 or 96 means there is a quadrature relationship between the two. If you don't know what that means, then you only need to know this setting to makes left and right channels sound different by adding stereo width to the effect. Anything near 64 will have minimal stereo spreading.

Fb: Feedback. 0 is no feedback. The original Vibe does not use feedback. A setting of -64 is extreme negative feedback and causes a more intense phaser sound. +64 makes the whole effect sound more "full".

L/R.Cr: Left/Right channel crossing. Mix left channel into right channel and right channel into left channel. Less than zero mixes the channels out of phase from each other, greater than zero mixes left and right like a normal mixer. This parameter can have an interesting outcome depending on the setting on St.df. In some cases the interference between left and right can be used to make a tremolo effect, particularly when St.df is set to 32 (quadrature).

Pan: Pan the effect to the left or right. This can be used to change the color of the effect between left and right channels if Wet/Dry mix is set to a certain amount dry.

Ininity

This is an implementation of and audio counterpart to the classic barberpole illusion: the filters appear to be continuously ascending (or descending) akin to the manner in which the colored strips on the barberpole appear to have no beginning or end. There are other illusions that come out of this effect model as well when you increase the rate.

One such illusion is the apparent beat frequency. Set the "Rev" switch to reverse left and right channel directions and mix all wet. Turn up the rate until you hear a beat frequency. If you are wearing headphones, remove one from the left or right ear: the beat frequency is gone. With stereo speakers, mute one speaker and the beat frequency goes away. It is most interesting with headphones because the left and right channel signals are completely isolated. The perception of a beat frequency is entirely in your head!

The infinity is also capable of frequency shifting, as in Single Sideband Modulation, not pitch shifting. This does crazy robotic things to vocals, or in a more subtle form may be used to slightly detune your instrument. You may also use it to detune a delay then use jack to create a feedback loop so you have an echo which is perpetually ascending or descending.

Apart from illusions and crazy sound effects, there is a wide range of subtle swirls or stereo panning and doppler shift effects possible.

Think of this as an 8-band EQ where each band perpetually moves up the spectrum to the top of the range, then starts over at the bottom. This will give an understanding for adjusting the level on the bands #1-8.

Wet/Dry: Mix un-filtered signal with the filtered signal.

Res: Filter resonance, also known as an adjustment of the Q factor. More positive is more resonance, negative is less resonance. It is suggested to set Res to 0 or less when using Infinity for frequency shifting or doppler effects.

1-8: Level of filter bands. Infinity uses 8-bandpass filters spaced evenly on a logarithmic scale from Start to End. The mix level of each filter band is adjusted by these controls. If all are set to the same value, they have practically no effect except when resonance is set quite high. It is suggested to use alternating +/- settings. A varying volume (like tremolo) effect can be obtained by setting them at different levels and low Res with some 0 in between.

Rev: Reverse the direction between left and right channels. If Right channel filters are sweeping upward, left will be falling when this is activated.

Stages: This adds up to 12 phaser stages to each filter band. A strong phaser effect can be obtained by setting this to something greater than 4 and alternating the bands #1-8 by 64, -64, 64, -64... There is an interesting switch internally that interacts with the Pan parameter when Stages is greater than 8. This makes pan into a variable rate parameter, which allows for rotating doppler effects.

AutoPan: As long as stages is, this acts only as an auto-panning effect that alternates amplitude on left and right channels. When stages then it acts as a variable rate control allowing doppler frequency bending to create the illusion of rotation.

St. df: Stereo offset between left and right filters.

Start: Where the filters start. Set lower than end if you want filters to sweep upward. Set greater than End if you want filters to sweep downward.

End: Where the filters end.

Tempo: The rate the filters sweep up or down.

Subdiv: Tempo frequency subdivision. Numbers greater than zero make the rate slower. Numbers less than 0 make the rate faster, even up into audible range frequencies for the crazy pitch shifting and ring modulation type sounds. Notice this has no effect on the AutoPan frequency. AutoPan Tempo is hard-coded to cycle every full measure at Tempo.