![]() |
|
![]() |
This section is about the selection of processes, as well as Infile types and file conversion.
CONTENTS
PROCESS MENUS
Keyboard Shortcuts
Menu Items
INFILE TYPES
FILE CONVERSION
CONVERSION SETTINGS
PROCESS TABLES
PROCESS LIST
PARAMETER FILES
PROCESS SEGMENT
CUT and PASTE
Non Cut+PastePROCESS MENUS
Soundshaper Process Menus give access to all available CDP functions.
There are six menus for CDP processes and two for supporting functions:
- EDIT/MIX: editing and mixing soundfiles (.wav).
These functions cover standard studio operations, with many innovative approaches as well.- SOUNDFILES: processes which transform soundfiles (.wav, also .evl /.env).
These functions alter sounds by speed/pitch, direction, segmentation and re-assembly, distortion etc.- SPECTRAL: processes which transform sounds in the frequency-domain (.ana).
These functions alter not only frequencies but also the spectral envelope and its evolution over time.- PITCH: extraction and processing of pitch data files (.frq / .pch.)
These functions alter and combine pitch data extracted from frequency-analysis files.- SYNTH: synthesis functions (.wav output).
These create new sound sources by generating and combining waveforms.- XTRA: functions that are supplementary to the main groupings., including legacy items.
SUPPORT FUNCTIONS:
- DATA: functions that generate or manipulate text data files (.txt / .brk).
- INFO: functions that give information about files or convert units.
CDP processes are described in the CDP sections of this manual, via pop-up descriptions on the Main and Parameter pages (F1), on the Ensemble Software website and in the CDP documentation. You can also get a usage statement for any CDP process or function by entering its name (in CDP format) at the command-line.
POP-UP MENUS
From version 5.08, new pop-up "context" menus have replaced most submenus. A new MENU button in the taskbar recalls the last-used menu (or SHIFT-CTRL+Z). The new menus have a brief description of the process, for comparison with other processes in the menu. They also have memorable keyboard shortcuts for immediate access at any time (see below).
![]()
KEYBOARD SHORTCUTS
An extensive system of keyboard shortcuts gives almost instant access to many processes and the pop-up menus.
- Soundfile menus: Shift+Ctrl+[intial letter], for example Shift+Ctrl+C for the CHANNELS menu.
- Spectral and Synth menus: Ctrl+Alt+[intial letter], for example Ctrl+Alt+M for the MORPH menu.
See the Menus themselves and the supplied file Shortcuts.txt for details. (It is not possible to assign your own shortcuts at the present time.)
MENU ITEMS
The menus are based on CDP program groups and names, but do not follow them exactly. However, the CDP name is given in the Information Panel at the bottom of the Main Page. Clicking on this links to the CDP documentation:
![]()
The menu item is accompanied by a colour-coded icon, indicating the number of inputs (1, 2, or > 2):
![]()
If there can be additional inputs, an ADD INPUT panel appears immediately following selection. (For details see SECONDARY INPUTS.)
Otherwise:
Selecting the process leads to a PARAMETER PAGE, unless the process takes no parameters.
INFILE TYPES
The colour of the menu-item icon indicates the native CDP input file-type. In Soundshaper, AUTO-CONVERSION (see below) handles the conversion between most file types, so that you simply have to choose the process. Where auto-conversion is not possible, you may have to add a cell that makes a specific conversion to the infile type required.
In Soundshaper, the different file types are colour coded as follows:
Input/Output files: Type: .wav /.aif Sound .ana Freq. Analysis (Spectral) .frq Pitch .trn Transposition .for Formant .evl Envelope Text Datafiles:
.brk Breakpoint .env Envelope Breakpoint .pch Pitch Breakpoint .txt Text .mix Mix .mmx Multi-channel Mix Most primary inputs are either Soundfiles (.wav), Frequency Analysis (spectral) files (.ana) or Pitch Files (.frq).
FILE CONVERSION
CDP employs several different file types. Natively, you must supply an infile of the correct type.
In Soundshaper, however, most of the necessary file conversion is done automatically: you simply select the process. For example, if a cell containing a soundfile (.wav) is the input to a spectral process, it is converted behind the scenes to a frequency analysis file (.ana) by the current PVOC settings.
The following patch is an example of Auto-Conversion the colours show that a variety of file types have been used in succession, but no cells have been needed for the conversions:
![]()
Auto-conversion applies to primary infiles and also to secondary ones selected from the grid. In the above example, MAKE requires two inputs: the first follows the previous process, PchExagg; for the second, you would click on the cell GetFmts.
There is also an automatic output conversion: if Options | Auto-FFT is checked, a spectral output file (.ana) is auto-converted to a soundfile (.wav), so that it can be played. Both .ana and .wav files are retained and the most suitable will be used for the next process.
The chart below illustrates the various possible auto-conversions:
![]()
- Soundfiles .wav can be converted from .ana.
This is the only "backwards" auto-conversion available.- Spectral files .ana can be converted from .wav.
- Binary pitch .frq can be converted from .ana or .wav.
- Breakpoint pitch .pch or .brk can be converted from .ana or .wav.
- Formant .for can be converted from .ana or .wav.
(Single-formant files also use the suffix .for.)- Binary envelope .evl can be converted from .wav.
- Breakpoint envelope .env or .brk can be converted from .wav.
To get back to sound from pitch (.frq, .pch/.brk) and formant (.for) files, we have to combine them using the function MAKE (see patch example above). There is also a function MAKE2 which combines pitch, formant and envelope data.
Transposition files (binary .trn or breakpoint .brk) are not auto-converted as they are also the result of combination: either of two pitch files or two transposition files (binary or breakpoint).
Textfiles (.txt), apart from the breakpoint types (.pch, .env, .brk), are also not included as they cannot be primary or secondary infiles to a process.
Hint: if auto-conversion is not possible, it can be useful to copy a cell to a spare row, do the conversion and then select the converted cell as input.
CONVERSION SETTINGS
Auto-conversion uses preset conversion values for PVOC (.wav to .ana), pitch extraction ( to .frq or .pch) and envelope extraction (to .evl or .env). All have default values which are preset on startup.
You can change the settings and also make specific conversions (using one extra cell) from these menus:
- PVOC: Spectral | PVOC | Settings / Analyse (Ctrl+F9) /Synth (F9)
There are also toolbar buttons for Analysisand Synthesis
.
An explicit PVOC Synth is sometimes necessary to complete a patch with a soundfile, for example to save the final output to a permanent file.
- Pitch: Pitch | GetPitch | Set Defaults / GetPitch using defaults or parameters.
- Envelope: Soundfiles | Envelope |Extract | Set Defaults / Extract envelope using defaults or parameters.
PVOC SETTINGS
The most widely used of these is PVOC (Phase Vocoder) for converting to and from the spectral domain.
![]()
The two parameters required for analysis are:
- No. of points. This determines the accuracy of the FFT conversion: more points give better frequency resolution, but worse time resolution. The default is 1024 (N=1024 of the PVOC program). This is approximately twice the number of channels (like bands in a Graphic Equaliser) used to analyse the sound. (There is one additional channel at 0Hz.)
- Filter overlap factor (1-4). Default is 3.
Other PVOC parameters are not set by the user but are determined from the two that you set.
The PVOC page shows these values (Channels, Chan. Bandwidth, Decimation Factor, Time between Windows) and adjusts them as you alter points and overlap. The number of channels and the number of windows are often used as parameters in spectral processes, so it is worthwhile reading about these factors in more detail.Note especially:
- Greater frequency resolution produces lesser time-resolution and vice versa.
- CDP processes involving more than one analysis file (such as VOCODE) generally require the same analysis settings.
- For most purposes the defaults work very well.
PROCESS TABLES
An alternative selection method to the menus is to click on the process name in one of the PROCESS TABLES (formerly called Process Grids), where functions are laid out in a table:
![]()
The tables are accessed from the top of the corresponding menu: e.g. Edit/Mix has the Edit/Mix Table as its top item. There are also the keyboard shortcuts Alt+1 to Alt+5.
SHORT FUNCTION NAMES
The Process Tables, like the Patchgrid, use Soundshaper's Short Names for CDP processes. Most of these mirror the CDP names (but usually omit the program "group" name); some come from the original CDP name (e.g. SCRUB). Short Names are also shown when you pass the mouse over a Process Menu item, at the top of the Parameter page, in the output cell on the Patchgrid, and in the Process List (see below).
PROCESS LIST
The Process List parallels the File List (Pool), in providing a drop-down list of favourite processes.
Click on the PROCESSES button, as necessary, to reveal the PROCESSES PANEL:
![]()
To launch a process from the list: select the item in the drop-down box and double-click on it, or click the OK button.
A Process list is simply a text list of processes, one process per line. These may be:
- General: for example, STACK, using Soundshaper's Short Names and recalling the process as you would do from a Process Menu.
- Parameter Set: use your own title and parameter values for a process, such as the "CloseStack" example in the screenshot.
The items in this case represent parameter files (.prm), which are text files saved on the Parameter Page, or here on the Main one.To add a process to the list: first run it and click on the leftmost button Add Process to List (shortcut CTRL+F8).
- You are asked whether you are adding a previously saved process (.prm file).
- If yes, a file selector opens in a sub-folder for that process, within your User-Folder (i.e. a place where you might collect parameter files together).
- If no, you are asked whether you are saving parameter values with the process.
- If that is Yes, then a file selector invites you to name a Parameter File (.prm) in the relevant Process sub-folder, then saves the file and adds its name (and path) to the Process List. If there are datafiles, these are also saved in the Process sub-folder.
- Otherwise, the general process name only (e.g. STACK) is added to the list.
Note that a Parameter File simply saves the set of parameter values for the given process, so can be used with any suitable infile.
PROCESS LIST FILES
The default Processes filename and location is: [User-Folder]\PROCESSES\Processes.txt. If this file exists, it is loaded on startup. Any number of processes can be in a list, and there is no limit to the number of Process list files you can have.
There are 5 buttons in the PROCESSES PANEL. From left to right these are:
- ADD PROCESS button: add the last-run process to the list, as described above.
- REMOVE PROCESS button: remove the currently-shown process from the list.
- LOAD PROCESSES button: load a Process list.
- SAVE PROCESSES button: save the current Process list to file.
- CLEAR button: clear the current list (does not erase the Processes file).
PARAMETER FILES
In addition to selecting parameter sets via the Process List, you can load a single preset process by dragging and dropping a Parameter File (.prm) onto the Main Page.
Parameter files may be saved on the Parameter Page (File | Save Snapshot to Parameter File) or via the Process Panel (as described above). As soon as the file is dropped, Soundshaper opens the appropriate Parameter Page with its parameter settings; you can then click RENDER to run the process or edit the settings further to your new requirements.
If you have favourite settings for a process, it is worth saving them to a Parameter File on the Parameter Page or adding them to the Process List. Parameter files are also created in the TEMP. OUTFILES folder whenever you run a process, so a simple copying/renaming of this file allows you to catch any parameter settings you have found useful. Because you have to click RENDER to run the settings, you are not tied to the values you entered when you saved the Parameter File. This is potentially a very versatile method of collecting single-process presets, as you can name them and save them however you like.
PROCESSING PART OF A SOUNDFILE
Being able to set loop-points and markers opens up possibilities for processing part of a soundfile (or a spectral file if Auto-FFT is checked). In fact CDP only processes whole files, so processing just a part is a matter of making a behind-the-scenes CUT and PASTE at selected points. It is not appropriate for all processes - see more below.
![]()
The PROCESS panel offers three choices, to process:
- the whole file: the default and normal CDP operation
- between Loop Points, which are stored with the data for the cell
- between Markers, using the currently displayed markers, which are just times, not associated with any particular cell or soundfile.
CUT WITHOUT PASTE
To CUT a file at loops or markers without an associated PASTE:
- Select PROCESS: At Loop Points or At Markers, as appropriate.
- Click the CUT button, or Edit Menu | Cut (Shift+CTRL+K).
- Process the selection. Only the selection is retained and processed.
Alternative edits (in the EDIT context menu at the right, or the full Edit/Mix menu) are ZERO CUT (cut at zero crossings) or EXCISE (discard the selected portion and close the gap). INSERT SILENCE may also be relevant in some contexts. To access full parameters for these edit functions, select PROCESS Whole file.
CUT AND PASTE
To cut, process and paste the result back into the original file:
- Select PROCESS: At Loop Points or At Markers, as appropriate.
- Select an appropriate CDP process. Most processes in the Soundfile and Spectral menus qualify.
- Soundshaper automatically CUTS the file at the given points; this appears on the patchgrid as SEGCUT.
- The parameter page appears (if there are parameters) for the CUT portion. Any timings are in relation to the CUT filelength.
On returning to the Main Page:- the display previews the PASTED file, i.e. the original with the selected portion processed.
- A Preview paste checkbox appears, checked; this can be unchecked to play just the processed segment out of context, and re-checked to switch back to the preview. (Re-select the cell after unchecking to achieve this.)
- A second or subsequent process may also be made at this stage, before PASTEing. There is no need to uncheck Preview (or de-select "At Loops" or "At Markers"); just add one or more further processes. Remember you are still processing the CUT segment.
- Click the PASTE button:
to confirm the paste edit.
The result appears in the next cell as SEGPASTE.- PASTE is also available in the Edit and Markers menus as Paste Segment (shortcut CTRL + =).
- Note that the new cell SEGPASTE does NOT inherit any Loop Points used for the Paste (though these can usually be set using GET LOOPS and SET LOOPS).
Please also note:
- The CUT/PASTE cycle is not complete until the PASTE button is clicked and the Preview checkbox disappears.
- If necessary, the CUT/PASTE operation can be cancelled in the Markers Menu: Cancel Cut/Paste Segment (shortcut CTRL + -). This is useful if at some point you forget to switch back to Process Whole File and trigger a fresh SEGCUT. (In this case, cancel CUT/PASTE, then UNDO the extra Segcut cell.)
- If you cancel the CDP process, you must also undo the SEGCUT cell (unless your process is not the first one following SEGCUT).
- Some processes may cause the PASTED insertion to lengthen the file or create a silent portion. Further editing using different loop points and EXCISE may be required.
- Changes in the number of channels caused by the process are catered for behind the scenes.
- CDP is very fussy about a PASTE into the end of a soundfile. To minimise the problem, Soundshaper adds a small amount of silence to paste into and also does a butt-end join.
- SEGCUT/PASTE can use up many processing cells. For this reason, Soundshaper now supports up to 99 cells per row, though the recommended default is 16, with only 12 columns visible at once, as before.
CUT plus USER-CHOSEN PASTE
Although the CUT button on the right of the page is not linked to a particular paste file (to paste into), if you change your mind after processing you can still paste into a chosen file:
- First select Loops or Markers to process, so that Soundshaper can use the timepoint where you made the CUT as an insertion point for the PASTE.
- If you're not pasting into the same file you CUT from, click on the cell for the paste file you intend to choose and set its Loop-start/Left Marker to the same insertion point: if using Markers, this timepoint will already be set; if using Loops, use GET LOOPS to grab the time and SET LOOPS to set it in the new file.
- Click the PASTE button and the ADD INPUT panel appears, allowing you to select your paste file from the patchgrid. If you have not chosen Loops or Markers, Soundshaper suggests you use the Marker time.
PASTE versus INSERT
Soundshaper's PASTE does not use full parameters. If you simply want an ordinary INSERT with parameters, choose Insert from the Edit menus (or Shift+Ctrl+V).
Note, however, that in CDP INSERT, the 2nd sound is inserted into the 1st. If, for example, you want to INSERT cell A_2 into A_0, there is no point in selecting A_0 as the 2nd file; this would insert A_0 into A_2. The correct method in the patchgrid would be to copy A_0 to B_0, then select A_2 as the 2nd file (B_0 --> INSERT A_2 --> B1).
The PASTE procedure in Soundshaper uses INSERT, but reverses the normal order of the files, putting the specially extracted Paste file as Infile1 and the current cell (to be pasted) as Infile2.
PROCESSES NOT USING CUT and PASTE
Certain CDP processes and functions are not suitable for Cut/paste and the facility is not offered. In general, these are editing functions, certain multi-channel processes, synthesis and 'timed' processes:
- INFO: Most Information functions can use Loop Points or Markers and report on the selected segment. They produce a SEGCUT, but this does not appear on the patchgrid.
- 'TIMED' PROCESSES: Many CDP functions process the whole file with reference to one or (more often) two time points, e.g. START and END. In some of these, the output is the whole processed file; in the others, it is just the processed part. (For list see Appendix: Timed Processes.)
If Option 3 (Loops/Markers to Params) has been set, then selecting Process At Loop Points or At Markers sends the appropriate time-point(s) to the Parameter Page. (You can over-ride these values, of course.) Several editing functions simply use the given times without visiting the Parameter page at all. If you want full parameters for these, select Process Whole File.
- TIMES FILE functions: these require a list of times and Markers can be used for this. For details, see Markers Menu and Appendix: Times Files.
- OTHER EDIING functions not using times or times files: including Trim, TopNTail2, Join group, Presil,Endsil; all channel functions, including multichannel pan.
- SYNTH functions and equivalent, such as EnvCreate and RecSF.
- MULTI-OUTPUT processes (e.g. Randchunks) use SEGCUT only, not PASTE.
Most processes that manipulate multi-channels also offer SEGCUT only, including Polyrhythms, Shifter, TexMchan and Graintex.
- PITCH Processes: the whole of the Pitch suite, except PchFix, which has 'time' parameters.
- Non-patchgrid output: All functions producing non-patchgrid output, such as text files.
RETURN to top of page |