Musical Role: COLOUR the whole sound

FUNCTION	PURPOSE
FILTER BANK [TD]	The centers of several bands of frequencies are emphasised, creating resonance effects. Example uses: lightly blur and warm a sound keep even or odd harmonics, discarding the others bring out the subharmonics of a sound colour with the harmonic (overtone) series colour with equal divisions of a frequency band The degree to which the filter focuses tightly on the bands determines how much one hears pitches: i.e., high Q and narrow bands. With lower Q and wider bands, more of the original sound 'comes through'. 'Q' is the focus of the filter. Reference Manual   Back
FIXED [TD]	Use a fixed roll-off slope to boost or cut frequencies. Example uses: hi-pass filtering: cut below a specified frequency lo-pass filtering: cut above a specified frequency notch filtering: cut around a specified frequency Boosting instead of cutting achieves the opposite effect: boost below emphasises the lower frequencies, boost above emphasises the higher frequencies, and boost around creates an area of resonance. The roll-off slope is known as Q. Perhaps professional engineers know why. Reference Manual   Back
FILTER LOHI [TD]	Create LO- or HI-PASS filters by specifying the outer limits of the frequency band to filter. Example uses: lo-pass: stop-limit is above pass-limit. This removes the upper frequencies, leaving the lower ones. It produces a softer, muffled tone. It can also be used to remove unwanted hiss in a sound, though a more elaborate CLEAN function is often needed. hi-pass: stop-limit is below pass-limit. This removes lower frequencies, leaving the higher ones. It produces a thin, perhaps metallic, tone. It can also be used to remove unwanted bass resonance. The musical use, as suggested here, is simply to colour the tone of the sound, making it either brighter or muffled. Reference Manual   Back
FILTER VARIABLE [TD]	Select modes for Lo- Hi- Band-Pass or Notch filters with only a single specifying frequency to specify. Example uses: easily compare the effect of different filter types create strong, pitched tones (resonance) by adjusting the acuity parameter Because only a single frequency is specified, this function has only a limited effect on the sound, but it is useful for exploring resonance effects placed at specific frequencies. Reference Manual   Back
FOCUS FOLD [TD]	Move (fold) the partials of the spectrum into the bounds of a specified frequency band. This has a marked effect on the colour, the timbre of the sound because the partials are so greatly moved from their original pitch levels. Example uses: folding frequencies up into a tight band can create eerie high pitched effects, sometimes like an ethereal wind or high and slightly ringing folding frequencies down creates deep, complex tones In the spectral domain, the frequency and amplitude of each partial is known as a result of the FFT analysis. This function is a good example of how this information can be used: in this case, shifting their pitch levels such that all the frequencies of the sound end up within the specified band. It is a very interesting function to explore. Reference Manual   Back
RING MODULATION [TD]	This is a great technique first used with analogue equipment. It both adds and subtracts a specified frequency, thereby creating two 'sidebands', one above and one below the original sound. Example uses: make a sound 'hollow' by giving a fairly high value for the modulating frequency, e.g, 1000 create beat patterns by specifying a very low frequency value, e.g., 5 enrich the timbre of a sound with a mid-range frequency value, e.g., 100 This process adds timbral edge and variety, and also makes other functions more effective because of the additional artefacts in the sound. Reference Manual   Back

Musical Role: COMBINE features of different sounds

FUNCTION	PURPOSE
COMBINE INTERLEAVE [SD]	Alternate short sections of 2 or more sounds. Example uses: rough-grained, flapping effect (many windows) fine-grained, sandy effect (few windows) new timbral effects by mixing several sounds in this way. There is always a certain 'motoric', mechanical quality to the outputs of this function, more pronounced when the inputs contrast. Working with more than 3 inputs is worth exploring! Key Parameter: windows – the number of analysis windows from each sound to interleave. Reference Manual   Back
COMBINE MAX [SD]	Subtle and somewhat unpredictable mix because the inputs are compared on a window-to-window basis, with only the loudest window being retained. Example uses: smooth, seamless mix of similar sounds with steady amplitude contours. features from contrasting sounds come and go The more the input sounds vary in timbral quality and change in their loudness contours, the more unpredictable and interesting the results. I have always been inspired by Rob Waring's mix of a female singing voice, a whistle and a deep drone. Note that a large number of inputs are allowed. Reference Manual   Back
ENVELOPE REPLACE [TD]	Here we take the loudness contour from one sound and put it onto another sound. There is no change of timbre, but the loudness of the second sound now rises and falls according to the pattern derived from the first. Example uses: introduce pulsations into a steady sound prepare a sound for a MORPH by giving it some patterning from the sound it will turn into This function only really makes sense when a steady sound is enlivened with the patterns of a changing sound. Smoothing out a lively sound only reduces its interest. Reference Manual   Back
FORMANTS VOCODE [SD]	This technique is known as cross-synthesis: the tonal characteristics cross from one sound to another. Example uses: one sound moves 'to the tune of another' – not just the normal amplitude contour of the samples as in the time-domain, but influenced by the colour highs and lows impose vocal speech on other types of sound associate the 'wrong' timbre with the performance characteristics of an instrument The key element is the spectral envelope. This means the changing pattern of the amplitude of the partials contained in the sound. Thus different tone colours are emphasised at different times. It is this changing tone pattern in sound 2 which is imposed on sound 1. Thus, if sound 1 is water and sound 2 is a voice, the timbral rise and fall of the voice is heard in the water, i.e., the water 'speaks'. It is clear, then, that this function will work best when sound 2 has very perceptible changes in the loudness curve of the tone colour: prominent strong frequency components. Reference Manual   Back
MORPH MORPH [SD]	Create a transition between the spectra of 2 sounds. This means interpolating, making gradual numerical transitions, between the frequency (spectral) content of one sound to that of another. The aim is to hear the timbre, the tone qualities, of the 2nd sound coming into the 1st and gradually replacing it. The two sounds need to be reasonably close in pitch and tone for this to work effectively. Example uses: make a seamless transition from one sound to another bring in the timbre of sound 2 before raising its amplitude bring in the (unaltered) sound of sound 2 before making the transition to its timbre with the help of other processes, as suggested below, feel sound 1 warp out of shape as it changes into sound 2 The software allows linear (steady) transition or exponential curves (getting faster or slower). The psychoacoustic aspect of audio transitions of this nature is still relatively unexplored. While visual images perceptibly distort very easily in the moment of transition, audio tends to blend more readily, i.e., it just sounds like a smooth change to the 2nd sound. The Reference Manual therefore discusses ways to prepare a perceptible half-way state that gets across that a morph is taking place: an alteration, a warping of shape, of form. Reference Manual   Back

Musical Role: Basic EDITING, for convenience, or use a soundfile editor / audio sequencer

FUNCTION	PURPOSE
SFEDIT CUT [TD]	Cut out and keep part of a soundfile within specified times. Example uses: remove unwanted portions before & after a recording: keep central portion FROM - TO specified times remove attack portion by omitting it from the portion to be kept: makes source of sound ambiguous cut out and keep consonants or vowels as separate soundfiles keep most 'interesting' part of a sound for further use/processing This is a vital tool. It can be used aurally with the FROM-TO option in Soundshaper (tick 'Play between markers'). Alternatively, one can perform this operation visually, using a graphic sound editor to block out the part to save, and then Save Selection. Reference Manual   Back
GAIN [TD]	Alter the loudness of a soundfile. Example uses: increase the loudness of a recording which is too quiet reduce the loudness of a sound which tends to distort ('overmodulate') during some types of processing increase amplitude after a FILTER operation makes it too quiet GAIN is done by multiplying the amplitude values of the sound by a gain_factor. For example, 2 doubles the amplitude, and 0.5 halves it. The -32766 to +32767 scale is used. SNDINFO MAXSAMP displays the maximum sample found in a soundfile and advises on the highest gain_factor that can be used. Note that too great an increase can cause 'digital noise': excessive differences of amplitude level between adjacent samples. Reference Manual   Back
ENVELOPE DOVETAIL [TD]	Another basic editing tool, it smooths the beginning and end of a soundfile. Example uses: eliminate a click at the beginning and/ or a sudden stop at the end soften a very sharp attack – attacks can be fine-tuned in endless ways, especially as there is the choice between a linear (steady slope) or exponential (slow to fast or v.vs. slope) shape. Always starting and ending at amplitude 0 (silence) and allowing indefinitely long slopes, this function provides a quick way to get a click-/bump-free start and end to the sound. Longer slopes and exploring linear and exponential shapes also provide ways to go beyond simple editing to real tonal changes. This can be especially helpful when preparing a sound for use in a TEXTURE function, because every iteration in TEXTURE starts at the beginning of the sound, unlike the granular functions, which move steadily through the sound. Reference Manual   Back

Musical Role: LENGTHEN or SHORTEN a sound, roughening its surface texture

FUNCTION	PURPOSE
DISTORT REPEAT [TD]	A form of wavecycle distortion: it repeats irregular lengths of soundfile, or groups of irregular lengths. Example uses: create a repetitive, churning effect extend the length of a sound The more repeats and larger the groups creates more of a churning effect while lengthening the soundfile. It can become too long and overstay its welcome, so moderation is in order. Reference Manual   Back
EXTEND LOOP [TD]	Move steadily through a sound, creating regular separate or overlapping segments. Example uses: separate the segments, creating a hic-cup effect (step longer than the segements) introduce regular pulsations into the sound (step length close to normal rhythms, such as semiquaver, triplet or quaver.) draw out a text by creating a stuttering effect move slowly through a sound with dense overlaps (very short step Surprisingly flexible, this tool is useful whenever some form of regularity in the output sound is desired. Reference Manual   Back
EXTEND SCRAMBLE [TD]	Here we have a random function controlling the placement of the segments. The process can be made to repeat, rescrambling the previous result. Example uses: mild displacements some loss of regularity and recognisability of the source significant amount of jumbling up of the original reduce original to a churning mass of sound Useful to roughen up or thoroughly jumble a sound, dramatic purpose and psychological impression is a musical consideration, especially when using words. Reference Manual   Back
GRAINMILL [TD]	Comprehensive granular manipulation facilities. Example uses: extend with timestretch (grain repeats) while adjusting grainsize (segment length) and density (grain overlap – or gaps) widen the pitch band in which the grains occur, possibly with time-varying upper and lower contour shapes. provide variations in the loudness of each grain create dense washes of sound GrainMill is different in that a large number of parameters are available at once, that ranges can be specified, that they can change over time, and that there is a built-in graphic editor to create time-varying contours. Within ranges the program selects a random position somewhere inbetween the limits for each grain of sound. Reference Manual   Back
MODIFY BRASSAGE [TD]	Simple granulation of a soundfile. Example uses: produce surface ripples on a sound introduce an echoey, reverberant quality roughen and reduce recognisability gradually slow down a sound while roughening its surface This is a straightforward tool for a minimal amount of granular processing. The various Modes allow you to focus on specific types of effect, such as Mode 2, which only does timestretch. Mode 6, however, provides several options, and Mode 7, everything. GrainMill is another version of Mode 7. Reference Manual   Back
STRETCH TIME [SD]	Literally stretch out the frequencies of a sound. Example uses: expose the beat patterns created by near frequencies make timbral transitions occur slowly slow down a sharp attack – which often reveals that multiple percussive impacts are in fact occurring draw out the syllables in spoken text, which emphasises the sonic role of the consonants enhance reverberation qualities Fascinating when applied to high-pitched screeching sounds, a timestretch yields equally interesting results when applied to dense harmonies, sharp attacks, sounds that contain strongly tone colours, and the spoken word. Reference Manual   Back

Musical Role: REARRANGE a sound

FUNCTION	PURPOSE
BLUR DRUNK [SD]	This process jumps about randomly within a specified range of analysis windows (less than 64). You tell it where in the sound to begin, and it then proceeds towards the end until the output duration you chose is completed. If a big jump takes it past the end of the sound, it is reflected back to the beginning. Example uses: mix up the attack portion of the sound start at some useful place after the attack move about in a restricted area – the original sound (such as words) comes through better, but randomly granulated (small range) jump about wildly – there may be very little sense of progression through the sound (large range) Without time-varying parameters, this function is relatively restricted (EXTEND DRUNK is more flexible). If the sound is fairly steady-state, there will be little perceptible effect: it works better with rapidly changing sounds, such as spoken text. Note that you can also lengthen a sound with BLUR DRUNK. Reference Manual   Back
BLUR SHUFFLE [SD]	This shuffle works on analysis windows, which are only ca. 0.02 seconds long. The mechanism is the same as DISTORT SHUFFLE: the domain specifies how many windows to use before moving on to the next set and gives them a letter name, such as abc, and the image defines the order into which they are to be shuffled, with or without repeats, such as abbaccb. Example uses: longer, but with little change in tone or texture (use a groupsize of 1) slightly more texturing, but still relatively clear (use a groupsize of 2+ with a more complex image) echoey bubbling (use a groupsize of 10 or more) DISTORT SHUFFLE operates on irregular lengths of wavecycle between zero crossings, which are both irregular and generally longer. BLUR SHUFFLE operates on windows, which are of a regular length and very tiny. This means that you have to work harder to produce the shuffle effect: such as by a longer domain, a more mixed up image, and especially, a larger value for groupsize (i.e., repeats of the image). The groups of windows are formed before shuffling, so a longer section of sound is affected with each shuffle operation. Reference Manual   Back
DISTORT REVERSE [TD]	Although the output of this function moves steadily through the original soundfile, it turns around groups of (pseudo-)wavecycle backwards. Example uses: roughen up the texture of a voice (small values) – what is being said remains recognisable large values (the number of cycles in a group can be as high as 5000) mostly have the effect of reversing the sound, with very little distortion mid-range values are interesting, e.g., 100. Now small portions of the sound are reversed, with sharp, sudden changes. This is a new and flexible way of playing with reversing a soundfile. Reference Manual   Back
DISTORT SHUFFLE [TD]	Allows you to pattern a number of wavecycle elements. The DOMAIN defines the elements (e.g., abc) and the IMAGE defines the pattern (e.g., abcbbccaaa). Example uses: introduce cyclic effects into the distortion create repetitions, symmetries etc. extend the length of the sound This is a powerful way to tease the ear with pattern in the midst of a roughly churning sound. There is a version in the spectral domain called BLUR SHUFFLE that operates on analysis windows rather than wavecycles. Reference Manual   Back
EXTEND SCRAMBLE [TD]	Here we have a random function controlling the placement of the segments. The process can be made to repeat, rescrambling the previous result. Example uses: mild displacements some loss of regularity and recognisability of the source significant amount of jumbling up of the original reduce original to a churning mass of sound Useful to roughen up or thoroughly jumble a sound, dramatic purpose and psychological impression is a musical consideration, especially when using words. Reference Manual   Back
REVERSE [TD]	Play a soundfile backwards (reverse the order of the samples). Example uses: long fades become gradual crescendi spoken text becomes an abstract series of sounds reduces recognisability of the original brings out different sonic qualities join backwards & forwards verions for long fade in/out Although a very simple process, reversing a sound can help prepare for more effective use of other types of transformation. Reference Manual   Back
SFEDIT JOIN [TD]	Splice two or more sounds end to end, with a definable amplitude slope at the joins. Example uses: join together original + reversed sound, or the opposite: reversed + original. Depending on the amplitude envelope shape, these will dip and rise in amplitude, or rise and dip – slow pulsations. join together 2 or more contrasting sounds and then timestretch with granulation, thus turning the contrasts into gradual transitions link several sounds with a silent soundfile inbetween each one in order to create a single soundtrack of separated sounds As suggested above, joining soundfiles can often be a way to prepare a sound for further processing. Reference Manual   Back
SUBMIX MIX [TD]	Place several soundfiles on a time line, occuring separately or overlapping. Example uses: assemble a passage of music out of several sounds enrich a sound by mixing together several variants at the same or nearly the same start time create new timbres by mixing disparate sounds balance the levels of several sounds and place in horizontal space create large-scale echo or stretto effects by overlapping the same sound at very small time intervals Mixing is the fundamental assembly process in audio music. More often than not, it will be achieved with full audio sequencer software. It can be done quite simply with the CDP software, if this is convenient. Reference Manual   Back

Musical Role: REDUCE – data reduction/thinning

FUNCTION	PURPOSE
BLUR BLUR [SD]	The frequency and amplitude data for defined lengths of soundfile are averaged, thus smoothing and blurring the sonic content. Example uses: create a smooth 'ambient', abstract flow take the edges off a sound slur speech Frequency and amplitude data in the spectral domain are organised into 'windows', each of which is only e.g., 0.02 seconds long. The number of windows is specified and the data in each group of this size is averaged. Reference Manual   Back
BLUR SUPPRESS [SD]	Remove the loudest components of the sound. Example uses: make a sound higher, lighter and somewhat metallic, especially voices (e.g., remove ca. 10-60 components) significantly reduce a very strong sound (e.g., remove ca. 100 components – much more than 100 and there may be very little left!) The strongest partials are often the fundamental frequency and first few overtones. Taking these away leaves the higher overtones and inharmonic partials. This leaves the higher and more complex partials, which can sound thin, metallic and cramped – such as a voice heard at a distance coming out of cheap headphones. Reference Manual   Back
HILITE TRACE [SD]	Take away frequencies except those in a specified number of analysis bands. This can be used: to clean up a noisy sound – keep e.g., ½ the bands to make a sound ambiguous – keep something like 50 bands to reduce a sound to a mere burble – keep e.g., 5 bands The number of bands to keep will alter with the nature of the input sound and the desired acoustic result. Reference Manual   Back

Musical Role: Add SPACE around a sound

FUNCTION	PURPOSE
MODIFY REVECHO [TD]	Give a sound a reverberant quality. Example uses: slightly enhanced, a little more 'alive': modest values for delay, mix (e.g. 0.5) and feedback (e.g., between 0.1 and 0.5 – negative values can create distortion) [Mode 1, Standard Delay] strongly reverberant: higher values for delay (e.g., 25ms-100ms), mix and feedbackweighted more towards 1 (more of the delayed signal) [Mode 1, Standard Delay] make the sound flap, like a sheet in the wind: explore the modulation depth and frequency parameters available in Mode 2 (variable delay) create the ambience of large spaces, such as a football stadium: Mode 3 This is a fairly straightforward approach to reverberation, using fixed (constant) delay times. Feedback too close to 1 tends to be too echoey. Some unusual effects can be achieved in Mode 2 by pushing the values of the various parameters towards their upper limits (or negative values for lfofreq, while Mode 3 makes it fairly easy to create the effect of large, open spaces. You can play with echoes in an interesting way by making the time between echoes (size) quite long, and/or by adjusting the number of echoes (count) – quite different parts of the sound can be made to overlap. The overall length of the sound will be extended if higher values are given for these parameters. Reference Manual   Back
MODIFY SPACE [TD]	Place the sound in the Left-Right stereo field, either at a fixed point or moving from place to place. This process is called panning (verb), or a pan (noun). Example uses: move from full left to full right or full right to full left, or anywhere in between move back and forth across the stereo field several times, and to different locations each time locate a number of sounds differently in the horizontal space so that, when mixed, they are nicely spread and do not cover each other – this is usually done in the mix itself. other functions enable you to narrow or reverse (mirror) the stereo image The pan operation requires a mono input and produces a stereo output. The mirror and narrow operations require stereo inputs. The location and movement of sounds in the aural space is an essential part of creating living sonic landscapes. As multi-channel facilities expand, the possibilities for doing so will become more and more interesting. Reference Manual   Back

Musical Role: – Alter internal TIMING of a sound

FUNCTION	PURPOSE
FOCUS FREEZE [TD]	Freeze the spectrum (i.e., continue in a steady state) at specified time points, either after the time points or before them. The key is to select times in the sound where there is interesting material to freeze. Example uses: freeze after a given time until next unfreeze point (e.g., a2.0 3.0: freeze at 2 sec., unfreeze at 3 sec.) take a later window for freezing and start using it earlier (e.g., 2.0 b3.0: freeze the window found at 3 sec. and start holding it steady at 2 sec.) in general, play around with starting and stopping the sound use this process not to create a whole new soundfile, but just to find some interesting spectral material to CUT out and use elsewhere You can freeze either amplitudes or frequencies only, or both together. The latter is very frozen, but the first two hold the one steady while the other parameter continues to move as normal. There really is a lot of scope to explore the possibilities of this rather strange function. Reference Manual   Back
FOCUS HOLD [SD]	This is like FOCUS FREEZE when it freezes both amplitudes and frequencies. You freeze between specified start and end times, of which there can be several. Example uses: Start (hold) and stop (release) at various points during the sound, e.g., at regular or irregular time intervals. When the sound is released after a hold, its normal movement resumes. Thus it is like the game of statues in which you freeze in position until released to move again. The sense of release is very strong and can be used to create tension and release effects. There are no 'after' or 'before' facilities in FOCUS HOLD, so it is simpler than FOCUS FREEZE and an easier way to do the starting and stopping of the freeze operation. Reference Manual   Back
FOCUS STEP [SD]	Freezes spectral content for specified (regular) lengths of time. Example uses: standard 'sample-hold' effect find interesting spectra to cut & use elsewhere introduce a jangling effect have a smooth, fine-resolution sandy effect The key factors is that the holds are regularly spaced. Thus the length of the holds determines the nature of the effect. Reference Manual   Back

Musical Role: – TUNE a sound to a harmony

FUNCTION	PURPOSE
FILTER VARIBANK [TD]	The 'vari' part of the name means that the frequencies specified in the bank of filters can change over time. The frequencies specified should lie where the loudest partials of the sound occur, or the filter bank will 'miss' the main part of the sound. Example uses: tune a sound to a recognisable harmony, or several harmonies that change over time control the balance between original and harmonised qualities by adjusting Q (filter focus) and roll-off (amount the amplitude drops away) The required datafile is very straightforward. It consists of 1 or more lines giving the time and then the pitch (as frequency or MIDI Pitch Value) and amplitude of each note in the harmony. When there are several lines with some time inbetween, the harmony changes gradually from the first harmony at the first time to the second harmony at the second time. As with all CDP time value automation files, the harmony will stay the same if two lines at different times are the same. Abrupt changes are made by placing two different lines at very close, adjacent times. Reference Manual   Back
PITCH TUNE [SD]	Operating directly on the spectrum (frequency content) of a sound, this is the most effective way to tune a sound to the pitches of a chord. Example uses: impart a harmonic flavour to a sound create a harmonic dimension cleanly, without too much of the resonance effect that can occur in FILTER BANK. Effective if the original sound is not too tightly focused in the first place. While FILTER VARIBANK can alter chords over time, tuning in PITCH TUNE remains the same throughout. (N.B. the newer TUNEVARY can change frequency over time.) Reference Manual   Back

Musical Role: Move UP/DOWN or GLISSANDO

FUNCTION	PURPOSE
DISTORT PITCHWARP [TD]	Bends the pitch through parts of an octave. Example uses: about ¼ octave (0.3) puts a tremor into vocal sounds higher values create a roller-coaster effect The fact that pseudo-wavecycles are present (irregular lengths) introduces distortion. On top of this the changing pitch levels, which can be minuscule or major, raise and speed up or lower and slow down the original sound material. Unless a time-varying automation file deliberately makes pitch changes instant, this will happen with glissandi between the pitch levels. Reference Manual   Back
FOCUS ACCUMULATE [D]	Create sustaining effects internal to the sound. Example uses: slow glissandi inside the sound, while its overall pitch remains the same: the decay has to be long enough to accommodate the size of the gliss a rich, somewhat mobile reverberant effect: low decay with minimal octave range for the gliss (e.g., 0.1) reverberant fluttering: decay values approach 1, and the octave width increases There is a balancing act here between the length of the decay and the scope of the gliss. A good place to start with this function is a very slow decay, near the bottom of the range, e.g., 0.005 along with a modest negative gliss, e.g., -0.4. Even less gliss results in a reverberant quality without too much sliding about, while more gliss increases the background movement. Higher decay values increase the reverb/sustaining until the whole sound becomes like an aural quicksand. Reference Manual   Back
MODIFY SPEED [TD]	The simplest way to alter the pitch level of a sound is to transpose in semitones (Mode 2) – parts of semitones can be used: a 100th part of a semitone is a 'cent'. Example uses: raise or lower the pitch level of a sound make voices high, fast & squeaky, or low and growley create deep, rich, long tones Because this is a time-domain transposition, raising the pitch level makes the sound shorter as well as higher; lowering it makes it longer as well as lower. While this makes voices sound rather artificial, it is very useful to get fuller, richer versions of a sound. Time-varying transposition creates glissandi between reasonably separated time-points, while instant changes of level can be achieved by making two time-points virtually adjacent, such as 2.00 and 2.01 (they have to be at least slightly different). Also note Mode 5, which enables you to ACCELerate or decelerate a sound over time. The spectral version, REPITCH TRANSPOSE makes it possible to transpose without altering the duration. Reference Manual   Back
PITCH TRANSP [SD]	This transposition function preserves the spacing proportions between the partials, which maintains the harmonic character of the sound. It operates either in fixed octaves or in semitones. Example uses: raise or lower a clearly pitched sound, intending to preserve its tone colour use the up + down option with fairly large values to split the sound, similar to a ring modulation use the up + down option with very small values to colour the sound with beat patterns Clearly pitched sounds are the most suitable inputs for this function. FOCUS ACCUMULATE + STRETCH TIME + PITCH TRANSP is a powerful combination: create internal movement, (CUT out the most interesting portion), time-stretch it to reveal the transitions, and move it up or down or up & down, keeping its harmonic character. Some very beautiful and haunting sounds may result. Reference Manual   Back
PITCH TRANSPOSE [SD]	Move whole sound up or down without altering its duration. Example uses: transpose a clearly pitched sound to several different pitch levels and MIX to create a chord – all the notes of the chord will be equally long, which cannot be attained in the time-domain change the pitch level of any sound without having too much impact on its timbre transpose a sound only a fraction of a semitone up and down and MIX these two variants with the original: a subtly enhanced, slightly reverberant result is created – if the transposition is even ½ a semitone, fairly strong beat patterns will begin to emerge This is an important alternative way to transpose a sound, avoiding the speed changes that occur in the time-domain, with all the timbral change created as a side effect. Reference Manual   Back
STRANGE SHIFT [SD]	This is a non-linear shift of frequencies. Example uses: push a sound into the acoustic 'stratosphere', making it extremely high and thin create long glissandi into or out of the acoustic 'stratosphere' or down into the depths It is the non-linear nature of this shift that alters the timbre of the sound: the spacing between frequencies is increasingly compressed (up) or expanded (down) instead of remaining the same. Reference Manual   Back
STRANGE WAVER [SD]	This comes across mainly as a frequency (pitch) vibrato, although the Reference Manual refers to an oscillation between harmonic and inharmonic states. This is because the stretch factor tends to shift frequencies upwards and squeeze them closer together. Example uses: The upward squeezing effect creates an inharmonic effect, which sounds somewhere inbetween hollow and tinny. The higher the value for stretch, the more this occurs. distorting the original sound. Thus you are playing with distortion of a sort and a ringing colouration of the sound. The of the waver (in Herz, or cycles per second) can be very slow, very fast, or changing over time. The faster it changes, the more a flapping effect is created. Low values for frequency and stretch produce modest changes, but higher values whizz it round considerably. The perception of the vibrato as a pitch vibrato is most noticeable on clearly pitched sounds, but not so much with more complex sounds, including vocal sounds. With the latter, the inharmonic qualities are more prominent. It is easy to go over the top with this function. Unless a wild flutter effect is needed, controlled movement brings out the changes it makes more effectively. Reference Manual   Back

---

Orientation

There are now many ways to alter sounds with computer software. This toolbox contains a selection of 19 tools, grouped by musical task.

These tasks work in two different domains, and it is important to understand the difference:

1. the 'time-domain' – operates on 'normal' soundfiles. These have data stored as a series of 'samples', like a string of beads, each of which has a time when it occurs and an amplitude, i.e., how loud it is. These sonic beads can be rearranged, duplicated or omitted (operations on the time data
) or be made louder or softer (operations on the amplitude data – the overall shape of the amplitude contour is called the amplitude 'envelope').
2. the 'spectral-domain' – operates on 'analysis' files. These have data created by a Fast Fourier Transform (FFT) process, stored in a complex, overlapping way as frequency (perceptible waves) and amplitude, i.e., how loud each perceptible wave is. To use the spectral-domain functions, a normal soundfile has to be converted ('ANALYSE') to an analysis file. After running a spectral process, the analysis file has to be converted back to a normal soundfile ('SYNTHESISE'). It is possible to audition an analysis file directly, using Richard Dobson's PVPLAY program, implemented in several ways in Soundshaper.

The table above lists 10 groups of musical tasks, lists the program functions which perform them in one way or another, and gives a necessarily general description of what to expect as a result. Follow links to the full Reference Manual for a complete explanation of all the parameters and how the function works.