USING "Sustain a specific FOF within a sound" to produce Song.

This process, on the FOF menu, can be used to extend pitched vowels in a source, and make them sing.

Success depends on the choice of material, and the parameters applied,
and output may have to be modified by small edits and/or (low pass, or band pass) filtering to produce a satisfactory result.

The pitch of the original sound must first be extracted using extract pitch from analysis file: to textfile.
(A successful pitch-track is essential).
The resulting file is the first parameter PITCH BREAKPOINT FILE.
The second parameter FREEZE is the time at which the sound is to be frozen and extended.
You can user the Sound View window to see the sound, and mark an appropriate time.
Choose a moment within a clearly pitched vowel sound.
If the output you get is at an unexpected pitch, or has an unsatisfactory quality,
try adjusting the freeze time by very small ammounts.
The OUPUT DURATION OF WHOLE SOUND will determine by how much the FOF is extended.
e.g. if the original sound is 2 seconds long, and the specified output duration is 8 seconds, the FOF will be extended to 6 seconds duration.
NB: IF you do not set the output duration it defaults to the duration of the source,
and the process will refuse to run.
Similarly, if you set the output duration to a value less than the source duration,
the process will refuse to run.
Normally you will want to grab just one grain.
Grabbing more than 1 grain usually tranposes the pitch downwards
(2 grains down an octave, 3 grains down 1 octave and a 5th, and so on down the subharmonic series).
Large number of grains produce audible loop-repetition effects.
If you now run the process (with vibrato depth and transposition zero)
you should produce an output with a static frozen FOF.
From this you can ascertain if the pitch is what you expected.
If the output has high frequency emphasis (sounds artificially "nasal"),
try setting SMOOTH FOFS and running again.
(Using this sometimes produces a discontinuity or 'bump' at the start of the frozen portion.
If this happens you can edit it out later).

To produce a more realistically \"sung\" result
- Set the vibrato frequency to somewhere between 6 and 15 cycles.
- Set the vibrato depth between .25 and .65
Test the output by running the process again.

To confine this vibrato to the extended FOF only (and to possibly add some randomised frequency and depth variation)
use the VibLocal button.
- The depth and freequency must be preset, in the parameter boxes, to non-zero values, before VibLocal is used.
- The degree of randomisation of either parameter (if required) can be set in the small window that pops up.
  Change the values using the Arrow keys.
- Breakpoint files will be generated and placed in the Vibrato Frequency and Depth parameter boxes.
You can now rerun the process and hear the output sound.
You can cause the pitch of the FOF to change by supplying a semitone transposition breakpoint file
for the GRAIN PITCH TRANSPOSITION parameter.

If you have Tabula Vigilans there will be a small piano-keyboard type key to the right of this parameter,
You can use this to create such a breakpoint file.
- First acknowledge that MIDI capture will proceed (click on the window which appears).
- Next hit the reference pitch from which the transpositions will take place.
  Normally this will be the pitch of the extended FOF itself.
- Now play a pitch line.
  It is best to use staccato articulation, as the events can be rejoined as required, afterwards.
- Hit the bottom note on your MIDI keyboard, to stop the MIDI capture.
- You will be asked if you wish to pre-sustain the entered pitches.
  The captured data, if staccato, will consist of a series of silence-separated events with quite steep on/off envelopes.
  If you intend to smooth the onsets of these events, you need to pre-sustain them by the likely smoothing (see below).
- You will next be asked if you wish to extend the last entered event to meet the end-portion of the sound.
  If you selected to freeze the sound at a moment not at the very end of the source,
  there will still be a segment of unmodified source retained after the frozen extension.
  If you opt to do the extend, your last MIDI event (provided it occurs before the end of the frozen segment)
  will be sustained to join onto this end-segment of the source.
- You will also be asked if you wish to apply the loudness envelope from the MIDI entry.
  Preserving the loudness not only retains the loudness envelope,
  but is also used to separate the extended FOF into distinct events separated by silence
  (only if you separated your MIDI input events by silence).
  If you opt to do this, a breakpoint file will be automatically generated and placed in the LOUDNESS CONTOUR parameter box.
If you now run the process again, you will hear a sound where the sustained FOF is transposed by your input data.
Typically, the steep envelopes will make it sound rather artificial.
Finally, if you generated silence-separated events, you can articulate these, form the Artic button.
(The Artic button will not function until you have generated such silence-separated events).

The Artic interface offers several options.
- You may "Stitch" the end of the sustained FOF to the remnant of the end of the source.
  Often there will be a loudness discontinuity here, manifesting itself as a bump or click in the sound.
  The Stitch facility allows you to specify an envelope to apply to the output, at this point,
  to remove this discontinuity.
- In the panel below, each silence separated event in your FOF-extension is assigend an event number
  and a set of options to apply to each event.
- Before you can articualte any event you must turn the articulation On
  by ticking the relevant Articulation On box, or the All Events: Artic On box.
- Once the articulation for an event is On you can enter values for Decay to minimum of
  and Minimum at time fraction in the boxes below the event number.
  These set the minimum level to which the event will decay, and the time-proportion between events at which the minimum is reached.
  (Also see the alternative, Rise Time and Decay Time below).
- Events may also be Sustain-ed, i.e. they are extended to meet the next event, but not slurred into them.
- Events may also be Slur-red, i.e. they are extended to merge into the next event, with no articulation between them.
- The Gain of each event can be adjusted.
- You can enter explicit values for the Rise Time and Decay Time of that event.
- You can also set global rise and decay times in the boxes above (GLOBALS) using the Arrow keys.
  You can apply these to all the On events by clicking on set rise or set decay.
  You can also remember, and recall these global values (from one session to another) using the Save and Recall buttons.
  
  Note that if you set a rise-time of (say) 0.3 seconds, but, when you created the source from MIDI input,
  you set the pre-sustain to less than 0.3 seconds, the previous pitch will extend into the rise-attack
  and the previous pitch will be heard to gliss towards the new pitch.
When you run the Articulation, it writes a new LOUDNESS CONTOUR envelope to the parameter box.

You must Run the Process again to hear the result.