15.0


With artifacts: Increase the "Quantization smoothing" parameter.
Creation of parallel voices

Set desired chord.

If necessary, change the "Humanize" parameter and press "Create voices" to generate the voices
again.
Pitch-sliced-objects and VIP objects
The basic frequency analysis reduces the pitch slices to pieces which ideally correspond to one note in
the music's melody or one syllable in speech. For speech the pitches correspond to the level of the
vowels.
After the analysis, all of the slices are shown selected (blue).
The subdivision into slices can be corrected manually. The slices are cut apart with the scissors tool, and
then merged together with the rubber band gripper tool ("Ctrl" held down).
The blue line represents the mid-range pitch for each object and results approximately from the average
value of the pitches in the slice.
In selected pitch-sliced objects the functions found in the "Editing sliced-object" group can be applied,
i.e. the correction of the pitch sequence and the mid-range pitch.
The option to reset the corrections for selected slices is also available.
Selecting sliced objects
Sliced objects can be selected and deselected. "Ctrl + A" selects all slices; clicking on the "empty space"
deselects them.
Fundamental frequency analysis correction
General options
The fundamental frequency analysis and the related fundamental frequency-dependent
pitch-slice-analysis may fail or cause errors for the following reasons:

Reverbed material,

Runtime stereophonics,

False analysis,

Voiceless vocals (sibilants...),

Noisy sounds.
In the last two cases, analysis cannot return a result as the fundamental frequency is not available.
MAGIX Elastic Audio
provides the user with the opportunity to achieve good results using many manual options, whereby the
fundamental frequency analysis does not or cannot deliver any results.
Error in the fundamental frequency analysis
Symptoms:
Sound:
The pitch varies strongly for a short period and does not correspond to the set value or characteristic.
When using the monophonic voice
algorithm, additional strongly distorted noises arise. The algorithm, which depends on the fundamental
frequency recognition itself, fails.
Graphics in the editor
: The curve exhibits a short strong jump and behaves in a very jerky manner.
Solutions:

With the help of the eraser you can try superimposing the original curve at these points.

If you set high values for the curve smoothing parameter, then short "hiccups" can be reduced in
severity or prevented completely.
Correction of slice borders
Page 229