CDP DOCUMENTATION – MAIN INDEX

Back to TOP of page

DISTORT – distortion based on wavesets (pseudo-wavecycles)

The DISTORT functions are variants of the grain concept, but the segments are pseudo-wavecycles called wavesets, found in zero-crossings. The wavesets are processed by repetition, omission, replacing waveforms, adding "harmonics", and so on. Cycles can be processed singly or in groups. Many processes do produce distortion with small groups of cycles, because of rapid and uneven changes of waveform, but larger groups produce smoother and often unpredictable results.

Back to TOP of page

ENVEL – alterations to the amplitude envelope

The 'envelope' is the overall loudness contour of a sound. You can extract an envelope (either as a binary or breakpoint text file) and edit or transform it in over 30 different ways. Envelope functions can alter the dynamic shape, creating fades and swells, warp it in many ways, or totally re-draw it, creating new envelopes. Most importantly, you can impose or replace the envelope of one sound with that of another.

Most envelope functions use a parameter called window-size. This is the time-period used to measure the amplitude successively and it determines how accurately the envelope is defined: a small window-size gives a precisely defined envelope; with a larger one, amplitude values are averaged over a longer time.

Back to TOP of page

ENVNU – specialised operations on envelopes

A small group which extends the scope of the ENVEL functions. PEAKCHOP 2 produces an envelope based around the peaks in the soundfile; when this replaces that of the original source, the result is a gated sound with silence-separated events, particularly relevant to the RETIME programs.

Back to TOP of page

EXTEND – processes which extend and/or segment a sound in time

The EXTEND group is concerned with repetitions of various sorts, which prolong the sound. The iterations tend not to repeat the whole file, but only segments of it. Typical of this approach are:

• LOOP, which not only loops a segment but can move the looping start-point through the file to create a form of time-stretching.
• ZIGZAG, which reads segments forwards and backwards, using an ever-changing start-point, and
• DRUNK, which takes a random walk through the file, extracting and splicing segments selected by a certain randomness of movement and segment size.

SEQUENCE and SEQUENCE 2 are quite versatile sample-playing functions, which can play a sequence of sounds at given times, with optional transposition, loudness and duration. (Soundshaper has a MIDI to SEQUENCE 2 conversion utility which enables the program to render standard MIDI files.)

Newer standalone functions attached to the group include CERACU and SHIFTER, which can execute complex polyrhythms.

Back to TOP of page

FILTER – techniques to filter or focus on frequency bands

Filtering changes the tone-colour of a sound, attenuating some of the harmonics and emphasizing others. The filter's power is well known through the classic "subtractive synthesis" model in which a synthesised waveform, rich in harmonics, is shaped by a time-varying filter. The classic filter types are:

• Low-pass: cut off top frequencies
• High-pass: cut off bottom frequencies
• Band-pass: a mixture of Low and High which preserves a frequency band; and hence Notch, in which a band is rejected
• All-pass: all frequencies are let through but with changed phase relationships

CDP's filter functions include all the classic types plus particularly effective Filter Banks (c.f. a Graphic EQ), which can be used to "tune" the sound by strongly emphasising certain frequencies over others. There are also spectrally based filters (see HILITE).

Back to TOP of page

FOCUS – focus or sustain spectral amplitude data

The FOCUS functions complement those in HILITE and BLUR by emphasising certain aspects of the spectral envelope, or operating across a number of time-windows. The time-based functions freeze, step-freeze or hold the sound over a period, while SUPERACCU (replacing the older ACCU) sustains energy in each spectral band (a little like a piano pedal). EXAG exaggerates the spectral envelope and FOCUS similarly emphasises the spectral peaks by filtering around them. FOLD radically alters the birghtness of a sound by octave-transposing frequencies into a more concentrated range.

Back to TOP of page

FORMANTS – extract and impose the spectral envelope (formants)

Formant files extract the time-varying spectral envelope. This can then be imposed on a different sound or combined with a pitch file to create a new sound. The group includes a spectral version of the classic Vocoder, in which a bank of filters controlled by the characteristics of one sound is imposed onto a second sound.

Back to TOP of page

GRAIN – granulate and manipulate grains

The main granular program is MODIFY BRASSAGE and its variants MODIFY SAUSAGE, WRAPPAGE and NEWTEX. The last two of these are included in the GRAIN group: WRAPPAGE is a multi-channel version of BRASSAGE, while NEWTEX generates a time-varying texture of transposable segments cut from one or more sources.

The other Grain functions manipulate the grains in 'grainy' sounds, i.e. sounds with silences separating the segments/grains. BRASSAGE Mode 5 can create such sounds, as can ENVNU PEAKCHOP combined with ENVEL REPLACE. Grains can be treated in many ways, such as being duplicated, omitted, repitched, repositioned, shuffled and reversed in order.

Back to TOP of page

HILITE – emphasise spectral amplitude, including spectral filtering

The spectral filtering functions are arguably the most powerful of the HILITE group: FILTER implements a number of classic filter types with the precision only possible in the frequency domain and GREQ (graphic EQ) is a versatile (though visually non-graphic) filter bank; TRACE retains the N most prominent bands in each time-window and VOWELS successfully imposes the formant characteristics of vowel-sounds onto any sound.

Back to TOP of page

HOUSEKEEP – basic soundfile housekeeping utilities

Some HOUSEKEEP functions are in fact sound utilities, handling basic channel functions (CHANS and CHANPHASE), gating functions (GATE and EXTRACT), or tidying-up facilities (TOPANTAIL2, EXTRACT 3 and 4, DEGLITCH and ENDCLICKS). RESPEC modifies the soundfile header directly.

The rest of the group consists of file-handling utilities – BAKUP concatenates files for archival; BUNDLE creates a text list of files and SORT sorts such a list; BATCHEXPAND expands a batchfile to take more inputs; COPY makes multiple copies of a soundfile.

Back to TOP of page

MODIFY –Wide variety of effects, including brassage, transposition, loudness, panning, delay and ring-modulation

The MODIFY group has most of the fundamental time-based functions such as SPEED (transposition), LOUDNESS, SPACE (panning) and REVECHO (delay and echo) as well as more radical traditional techniques such as Reverse, Ring-modulation and Shredding (in MODIFY RADICAL).

BRASSAGE implements granular reconstruction (also known as "granular synthesis"), typically creating thickly textured sounds out of overlaid segments (grains), and CDP's program GRAINMILL provides a standalone graphical interface for this. The technique is closely related to the separate group of GRAIN functions.

Back to TOP of page

MORPH – Create smooth transitions between sounds, using amplitude and frequency analysis data

Morphing interpolates between the spectra of two frequency-analysis files, creating a gradual transition between them. The Release 7 functions NEWMORPH and NEWMORPH 2 are particularly effective, including tuning options and morphing between the frequencies of spectral peaks. Another powerful function is GLIDE, which makes a frequency transition between two single spectra over any time; a spectral envelope from a suitable sound can then be imposed on the result.

Back to TOP of page

MULTICHANNEL – multi-channel diffusion and processing

The functions in the MULTI-CHANNEL group are all standalone programs which have been brought together purely for reference purposes. Several are multi-channel equivalents of earlier mono/stereo functions.

A second sub-group manipulate channels and sounds in multi-channel space, involving spatial rotation and panning, re-ordering channels, distibuting sounds across a spatial PANORAMA, or segmenting and dispersing the fragments spatially (FRACTURE). Mention should also be made here of the MULTI-CHANNEL TOOLKIT, which similarly includes functions for channel re-ordering and rotation.

The third sub-group is concerned with multi-channel mixing: the multi-channel mixing program NEWMIX MULTICHAN has its own multi-channel mixfile format and a support program MULTIMIX to help with creating such files.

TANGENT and TRANSIT

Multi-mix files are also the output of the TANGENT and TRANSIT groups of functions, which move a sound or a sequence of sounds along a tangent to (Tangent) or across (Transit) a circle of loudspeakers.

Back to TOP of page

MULTI-CHANNEL TOOLKIT – multi-channel file handling functions

The Multi-Channel Toolkit is a group of versatile tools for handling multi-channel files, including a special emphasis on ambisonics and the WAVE_EX file format. The tools include functions to extract channels, change the channel order, mix and interleave multi-channel files, play them, obtain information about them, plus COPYSFX (CDP's "Swiss Army Knife") which can copy/convert between a very wide range of soundfile formats.

Back to TOP of page

ONEFORM – operations on single formants

The ONEFORM group complements the FORMANTS functions, which extract and employ time-varying spectral envelope (formant) files. GET extracts a single spectral envelope from a formant file at a specific time. PUT and COMBINE produce sounds of static amplitude (which may still vary in frequency content), the first by imposing the single formant on a spectral file, the second by combining it with a pitchfile (c.f. COMBINE MAKE).

Back to TOP of page

PITCH (&Harmony/Freq) – operations based on partials

The PITCH spectral group consists of functions which directly alter frequency content and includes two transposition functions (TRANSP and OCTMOVE). CHORD(F) does not produce a conventional chord, but rather a timbre enhanced by superimposing transposed spectra. The most powerful functions in the group are probably PICK and TUNE, which tune a spectrum onto the specified partial sets; TUNEVARY can do this in a time-varying manner.

Back to TOP of page

PITCHINFO – information about partials or extracted pitch traces

PCHINFO functions are for binary pitch files (.frq), extracted from spectral files by REPITCH GETPITCH. They either report to the terminal or produce textfile reports. Basic file header information is provided by SNDINFO PROPS. INFO gives the maximum, minum and mean pitch in the pitch file and the total pitch range. ZEROS reports whether the file contains unpitched windows. You can also hear the pitch the file has found, or view it (or transposition data .trn) as a pseudo-soundfile.

Back to TOP of page

PSOW – manipulate pitch-synchronous grains (FOFs)

PSOW is a set of experimental grain processes for vocal sounds. It attempts to find and then manipulate FOFs, which are formant waveforms (grains), used to synthesise the singing voice. (The term comes from IRCAM's CHANT program.) PSOW processes require a pitch-trace taken from a spectral analysis file. The programs allow you to alter formants independently of pitch, often with unexpected results.

Back to TOP of page

PVOC – FFT analysis and resynthesis

The public-domain PHASE VOCODER (PVOC) is used by CDP to convert between the waveform representation of sound in the time-domain (.wav / .aif files) and time-varying spectral analysis files (.ana) in the frequency domain – and back again.

While in the frequency domain, the different "bins", or frequency bands, are manipulated by Spectral and Pitch processes to create effects that could not be achieved in the time-domain. CDP has been especially innovative in this regard.

Back to TOP of page

REPITCH – Extract and alter pitch-related frequency analysis data

Apart from REPITCH TRANSPOSE(F), which transposes pitch without altering time, the REPITCH programs are concerned with extracting pitch from frequency analysis files and processing the resultant pitch files, which can be in binary or breakpoint-text format.

To re-synthesise sound, COMBINE MAKE combines the pitch file with a formant file (.for), which represents the time-varying spectral envelope. Typically, the formant file might come from another sound. When processing pitch, it's important to realise that the dynamic profile will be that of the formant file and that changes to pitch will only be convincing to the extent that the frequencies of the altered pitches are also present (i.e. strong or weak) in the formant file.

Back to TOP of page

REVERB – functions to reverberate soundfiles

The REVERB group's two main programs are REVERB and ROOMVERB, comprehensive reverberation functions incorporating such features as reverberation time, dry/wet mix, absorption and early reflections. The programs use built-in sets for small, medium and large rooms, or they can use a datafile created by the separate ROOMRESP program. There is also optional filtering and pre-delay (giving the effect of a bigger room). ROOMRESP's data can also be used by TAPDELAY, a tapped delay line.

Also in the group is FASTCONV, an FFT-based convolution program, which can simulate a wide range of reverberation types using a sampled impulse response of a building or other responsive space. (Many suitable impulse-response files are available on the Internet.) More experimentally, the impulse-response input can be any soundfile. Convolution can also implement an FIR linear-phase filter.

Back to TOP of page

RHYTHM – rearrange and retime events in a soundfile

The RHYTHM group consists largely of one program, RETIME, which has 14 modes dealing with the retiming of events in a soundfile. There are three divisions: functions that create or retime silences, those that retime amplitude peaks, and those that retime silence-separated events.

Back to TOP of page

SFEDIT – basic soundfile editing

One of the largest function groups, SFEDIT has the fundamental cutting and splicing functions found in visual editors, but CDP also extends these operations into relatively uncharted areas.

For example, MASKS covers portions of the sound with silence; applying different masks to copies of the sound effectively partitions it in different ways. Similarly, ISOLATE, and PARTITION partition sound into separate output files, each retaining the original segment times; like the MASK outputs, these files can be further processed independently prior to a possible remix. RANDCHUNKS and RANDCUTS also output multiple files, cut at random within a specified range. TWIXT and SPHINX switch between a number of sounds seen as being streamed simultaneously. PACKET outputs short enveloped sound grains that can be used with TEXTURE, BRASSAGE or PANORAMA.

JOIN is a standard splicing program, but there are also simple sequencing functions JOINSEQ and JOINDYN (c.f. SEQUENCE / 2 in the EXTEND group.)

Back to TOP of page

SNDINFO – basic soundfile information

SNDINFO functions either report to the terminal or produce textfile reports. The functions cover such aspects as soundfile properties, length, maximum level, lowest level, loudest channel and channel differences, zero-crossings, and start-times. (RMSLEVEL, in the Multi-Channel Toolkit, reports the overall power level of each channel.) PEAKFIND lists the times of loudness peaks found and is an important data file for other programs such as RETIME. PRTSND converts soundfile data (time-varying amplitude) to text. Some functions require two or more infiles for the purpose of comparing soundfile levels, lengths, and summing lengths.

Other reporting functions specific to particular Function Groups are found in that group. These include cycle count for DISTORT, grain count for GRAIN, find pan for MODIFY SPACE and stretch factor for STRETCH TIME.

UNITS is a general-purpose musical calculator, offering a number of useful unit conversions for pitch and frequency, transposition ad interval, speed and time, and loudness as gain and dB. TIMESMP converts time in a soundfile to sample-count, for sample-accuracy editing.

Back to TOP of page

SPEC – gain and editing utilities in the spectral domain

A collection of spectral utilities. CUT and GAIN are spectral counterparts of time-domain functions. BARE zeroes channels which are not harmonics, according to a pitchfile used for comparison. The reference manual lists some possible uses.

GRAB retains a single analysis time-window at the specified time. This is required by MORPH GLIDE (only). MAGNIFY expands a single window over a specified time, not unlike the FOCUS functions HOLD and FREEZE.

CLEAN has been superseded by SPECNU CLEAN; both require a sample of noise for comparison, so that the noise frequencies may be spectrally removed.

Back to TOP of page

SPECINFO – information about spectral data

SPECINFO functions either report to the terminal or produce textfile reports. Basic information includes the number of time-windows, and conversions between channel number and frequency. File header information is provided by SNDINFO PROPS. Various functions report on the time-varying spectral peaks in the sound or extract the amplitudes of partials; NEWMORPH 2 (documented in the MORPH function group) lists the frequencies of the most prominent spectral peaks. You can also view the changing level as a pseudo-soundfile or print (part of) the analysis data to a textfile. (This produces huge files and is best used for tiny portions.)

Back to TOP of page

SPECNU – various functions, mainly to clean up frequency analysis files

Three of the original functions in this group (CLEAN, REMOVE and SUBTRACT) are concerned with cleaning up files, though the latter two may have more general uses.

SLICE partitions the spectrum into frequency bands based on permutations of channel numbers and outputs these as separate files. This is an important resource for separate processing of the files, prior to a possible remix. For example, a mono source can be turned into pseudo-stereo by giving the slices different pan positions. SLICE Mode 5 inverts the spectral frequencies around a given frequency value, which destroys the formant relationships in speech or singing, while retaining the rhythm.

Newer functions in the group include RAND, which randomises the order of spectral time-windows. (The Release 7 program SPECGRIDS is identical to SLICE, Mode 1.)

Back to TOP of page

STRANGE – some unusual and unpredictable effects

The "strangeness" of this Function Group is perhaps subjective. GLIS attempts to create glissandi inside the changing spectral envelope, with a limited degree of success (SUPERACCU does it better). WAVER oscillates between a harmonic and inharmonic state. Its detuning can be subtle or extreme; if minimal, it can be used to give a chorus effect. INVERT inverts the spectral envelope in each time-window, transferring the energy in the lowest frequency bands to the highest ones and vice versa; the result is typically a much brighter timbre.

More useful is SHIFT, which performs a linear frequency shift of all or part of the spectrum, turning harmonic sounds into inharmonic ones. An upwards shift compresses the musical intervals between the partials and a downwards shift stretches them apart, not unlike STRETCH SPECTRUM.

Back to TOP of page

STRETCH – stretch time or frequency data

A mini-group, consisting of two processing functions:

• STRETCH SPECTRUM warps the spectrum by stretching or compressing its frequency components, so that, for example, a harmonic sound may become inharmonic. This function is closely related to STRANGE SHIFT
• STRETCH TIME stretches or compresses the duration of the sound without changing the pitch. The degree of stretching or compression is itself time-variable.

Attached to the group is the utility program STRETCHA, which calculates the stretch factor needed to convert one duration to another, especially in relation to a given number of beats and tempo.

Back to TOP of page

SUBMIX – mixing and related processes

The centrepiece of the SUBMIX group is MIX, which mixes the sounds listed in a text mixfile, the lines consisting of: soundfile name, start-time, no_of_channels, ch1_level, ch1_pan [ch2_level, ch2_pan].

There are many support functions which manipulate the data in mixfiles, such as PAN, ATTENUATE, SHUFFLE, SYNC, SYNCATTACK, SPACEWARP and TIMEWARP.

Other mixing functions are also available, which do not use mixfiles. These include:

• MERGE – a quick mix of 2 soundfiles
• MERGEMANY – a quick mix of several soundfiles
• BALANCE – mix 2 soundfiles, using a balance function
• CROSSFADE – quick crossfade between 2 soundfiles
• FADERS – mix several mono or stereo files using a time-changing balance function

The two INBETWEEN functions produce multiple outputs, involving a transition in which the balance shifts from predominantly one sound to the other. Finally, an important workhorse – INTERLEAVE – combines mono signals to create a stereo or multi-channel file. (A similar function, INTERLX, is found in the Multi-Channel Toolkit.)

In general, CDP Mix functions are best for mixing small sounds to create larger units. Although larger mixes are perfectly possible, a visual editor is more intuitive and tends to have better controls.

Back to TOP of page

SYNTH – functions to synthesise sounds

Only a small number of synthesis functions are provided within CDP, which is predominantly a sound-processing system. The older functions create basic waveforms, noise, silence, a chord for test purposes, a tone of variable pitch within a frequency band (SPECTRA), and a click-track.

The NEWSYNTH program is in a different league. Modes 1 and 2 (akin to CSOUND's ADSYN) generate a waveform with a time-varying spectrum and any number of harmonic or inharmonic partials. Mode 2 allows each wave-cycle envelope to be shaped, to a degree. In Mode 3, a multi-channel synth, pulsed selections of the spectrum are sent to a multi-channel output. Each output event is a random selection of partials from the overall spectrum and subject to octave-transposition over a specified pitch range.

Mention should also be made of undocumented functions developed for sonification and distributed with Release 7. (These will be documented in due course.) Sound can be created from time-series text data, or the data can be assumed to be a pitch-trace. The SPECTRUM programs can generate a time-varying spectrum from "frq-amp" pairs, and there are supporting spectral analysis functions which can generate such text data from input sound.

Back to TOP of page

SYSUTILS – system utilities

A group of standalone support programs, including functions to play and record soundfiles (PAPLAY, RECSF) and play spectral analysis files (PVPLAY), list available audio devices (LISTAUDEVS) and (sound)files in a folder (DIRSF), and create a shortcut to a file (ALIAS, PC only). There is also a MAC utility (CDPCONV) to convert spectral files from PPC to Intel format. COPYSFX, which converts between soundfile formats, is also included in this group (as well as in the Multi-Channel Toolkit).

COLUMNS is a major data-massaging program which is implemented in Soundloom's Table Editor and Soundshaper's Data Editor. Around 100 functions alter or generate data textfiles, for possible use with CDP programs. The COLUMNS functions are not concerned with the appropriateness of the resulting data for any particular purpose: the user must ensure that ranges and values are suitable for the intended parameter.

Back to TOP of page

TEXTURE – create textures of sounds

The TEXTURE suite is based on algorithmic repetition: the processes repeat and transpose the input sound or sounds in various ways to create a texture of events ('notes'). The note events are whole sounds (which may or may not be played to the end), but the user can easily edit one or more sounds to provide segmented events for the processes; ISOLATE, PACKET and PARTITION are three functions which can produce suitable small segments.

Note events can be treated as simple repetitions, or repeated in groups, or with a timed rhythm, or as transposed ornaments or fully-defined motifs (timed and transposed). Key parameter values such as pitch, duration, gain (level), spatialisation and sound number are generally chosen randomly from within a specified time-variable range (which can also be a fixed value).

In each of the eight main functions, repetitions can be pitched at random within the defined pitch range, or restricted to a user-defined pitch-set or harmonic-field (which octave-transposes the pitches); the set/field can be time-varying. More than one input sound can optionally be used, and the range of sounds chosen can be time-varied.

It is not easy to grasp the differences between the Texture programs, and the suite is best explored at first using a simple note sample or similar short sound. CDP supplies two sets of tutorial examples for this purpose.

Back to TOP of page