TimeStretch Filter

TimeStretch

TimeStretch (clip, float "tempo", float "rate", float "pitch", int "sequence", int "seekwindow", int "overlap", bool "quickseek", int "aa")

TimeStretch allows changing the sound tempo, pitch and playback rate parameters independently from each other, i.e.:

Sound tempo can be increased or decreased while maintaining the original pitch.
Sound pitch can be increased or decreased while maintaining the original tempo.
Change playback rate that affects both tempo and pitch at the same time.
Choose any combination of tempo/pitch/rate.

Parameters:

The speed parameters are percentages, and defaults to 100. If tempo is 200 it will play twice as fast, if it is 50, it will play at half the speed. Adjusting rate is equivalent to using AssumeSampleRate and ResampleAudio.

The time-stretch algorithm has a few parameters that can be tuned to optimize sound quality for certain applications. The current default parameters have been chosen by iterative if-then analysis (read: "trial and error") to obtain the best subjective sound quality in pop/rock music processing, but in applications processing different kind of sound the default parameter set may result in a sub-optimal result.

The time-stretch algorithm default parameter values are

Sequence     82
SeekWindow   28
Overlap      12

These parameters affect the time-stretch algorithm as follows:

Sequence: This is the length of a single processing sequence in milliseconds which determines how the original sound is chopped in the time-stretch algorithm. Larger values mean fewer sequences are used in processing. In principle a larger value sounds better when slowing down the tempo, but worse when increasing the tempo and vice versa.
SeekWindow: The seeking window length in milliseconds is for the algorithm that searches for the best possible overlap location. This determines from how wide a sample "window" the algorithm can use to find an optimal mixing location when the sound sequences are to be linked back together.

The bigger this window setting is, the higher the possibility of finding a better mixing position becomes, but at the same time large values may cause a "drifting" sound artifact because neighboring sequences may be chosen at more uneven intervals. If there's a disturbing artifact that sounds as if a constant frequency was drifting around, try reducing this setting.
Overlap: The overlap length in milliseconds. When the sound sequences are mixed back together to form a continuous sound stream again, this parameter defines how much of the ends of the consecutive sequences will be overlapped.

This shouldn't be that critical parameter. If you reduce the Sequence setting by a large amount, you might wish to try a smaller value on this.
QuickSeek: The time-stretch routine has a 'quick' mode that substantially speeds up the algorithm but may degrade the sound quality.
aa: Controls the number of tap the Anti-alias filter uses for the rate changer. Set to 0 to disable the filter. The value must be a multiple of 4.

The table below summarizes how the parameters can be adjusted for different applications:

Parameter name	Default value magnitude	Larger value affects...	Smaller value affects...	Music	Speech	Effect in CPU burden
Sequence	Default value is relatively large, chosen for slowing down music tempo	Larger value is usually better for slowing down tempo. Growing the value decelerates the "echoing" artifact when slowing down the tempo.	Smaller value might be better for speeding up tempo. Reducing the value accelerates the "echoing" artifact when slowing down the tempo	Default value usually good	A smaller value than default might be better	Increasing the parameter value reduces computation burden
SeekWindow	Default value is relatively large, chosen for slowing down music tempo	Larger value eases finding a good mixing position, but may cause a "drifting" artifact	Smaller reduce possibility to find a good mixing position, but reduce the "drifting" artifact.	Default value usually good, unless a "drifting" artifact is disturbing.	Default value usually good	Increasing the parameter value increases computation burden
Overlap	Default value is relatively large, chosen to suit with above parameters.		If you reduce the "sequence ms" setting, you might wish to try a smaller value.			Increasing the parameter value increases computation burden

Notes:

This is NOT a sample exact plugin. If you use it, slight inaccuracies might occur. Since we are dealing with float values rounding errors might occur, especially on large samples. In general however inaccuracies should not exceed a few 10's of milliseconds for movielength samples.

Currently the SoundTouch library only supports 1 and 2 channels. When used with more than 2 channels, each channel is processed individually in 1 channel mode. This will destroy the phase relationship between the channels. See this thread for details :- TimeStretch in AVISynth 2.5.5 Alpha - Strange stereo effects ?

SoundTouch is used in float sample mode.

Examples:

TimeStretch(pitch = 200)

This will raise the pitch one octave, while preserving the length of the original sample.

TimeStretch(pitch = 100.0*pow(2.0, 1.0/12.0))

This will raise the pitch one semi-tone, while preserving the length of the original sample.

TimeStretch(tempo = (100.0*25.0*1001.0)/24000.0)

This will change the tempo from Film speed to PAL speed without changing the pitch.

Credits:

This function uses:

http://www.iki.fi/oparviai/soundtouch
http://www.surina.net/soundtouch

Changelog:

v2.55	Initial Release
v2.57	Expose soundtouch parameters

$Date: 2008/12/24 22:55:01 $