1 Shepard tones
2 inharmonic glide
3 self-glissando
infile – input analysis file made with PVOC
outfile – output analysis file
-f extract formant envelope linear frequency-wise, using 1 point for every N equally-spaced frequency-channels
-p extract formant envelope linear pitchwise, using N equally-spaced pitch-bands per octave
-i quicksearch for formants (less accurate)
glisrate rate of glissing in semitones per second (negative value for downward glissando)
hzstep partials spacing in inharmonic glide. Range: FROM channel-frq-width (usually 43Hz) TO Nyquist/2 (i.e., ¼ the sample rate)
-ttopfrq top of spectrum: must be > 2*channel width of the analysis. Default: Nyquist (i.e., ½ the sample rate)
This program extracts the time-changing spectral contour (the spectral trajectory) of the sound. It therefore retains any spectral articulation, such as patterns of speech, but replaces the signal by an endlessly rising (positive values for glissrate) or falling (negative values for glissrate) shepard tone or inharmonic glissando. It uses linear interpolation in determining the spectral contour.
In Mode 1 (Shepard tones)
the key parameter appears to be glissrate. A high positive or negative value (e.g., + or -50) introduces quite a bit of noise into the sound, whereas a low value such as 0.3 results in a twangy vibrating sound.In Mode 2 (inharmonic glissando)
the two most salient parameters are glissrate and spacing. Higher postive or lower negative values for glissrate, e.g., +24 or -24, +100 or -100, gives a very rapidly cycling glissando (a pretty mushy twang). Low values for spacing (e.g., 45) gives a fairly deep twang, whereas high values, e.g. 1101, results in a rather hollow sound.The number of channels (frequency-wise formant extraction) or bands per octave (pitchwise) affects the resolution of the analysis without changing the resulting sound effect very much, and adjusting the top frequency can offer a certain degree of filtering control.
In Mode 3 (self-glissando)
the original tonal quality of the sound is largely preserved, especially with low values for glissrate. With higher values of glissrate, a higher, somewhat hollow glissando is added to the sound, otherwise very much like the original input.
End of STRANGE GLISIn effect, the original sound is replaced by the glisandoing tone, but the articulations of the original are retained, i.e., the rhythm and attack transients of the spectral envelope are notably the same. However, the tone of the sound is greatly altered, acquiring in Mode 2 as noted above a mushy twang with low values for spacing and a rather hollow vibrating sound with higher values for spacing. One set of trials on a low drum sound transformed it into something not unlike a didgerdoo.
Because formants are involved, input sounds with significant resonance features work well with this function, e.g., vocal and some drum sounds, especially those lower in the frequency range.
1 Normal inversion
2 Output sound retains the amplitude envelope of the source sound
infile – input analysis file made with PVOC
outfile – output analysis file
STRANGE INVERT is derived from an idea developed by Allen Strange. It inverts the spectral envelope, relative to the overall spectral envelope. This means that the energy asociated with the highest frequency bands is transferred to the lowest ones and vice versa. As the lowest partials in many sounds have the greatest amplitude and the highest ones the least, the result is typically a much brighter timbre. Mode 2 retains the amplitude envelope of the sound, so will sound closer to the original, with just the timbral change.
STRANGE INVERT does NOT invert the frequencies themselves. For this see SPECNU SLICE, Mode 5.
This function was formerly SPEC INV, which required that its infile be generated by SPEC BARE (zeros channels which don't contain harmonics). The SPEC BARE function is now built into STRANGE INVERT, meaning that the result may have more of a 'harmonic' quality than it might have. This will be related to the degree that the original sound was distinctly pitched. The main reason for including the BARE function, however, was simply to make the spectrum 'easier to manipulate'; it doesn't go so far as to harmonise the sound as such (see the PITCH functions for this).
The spectral envelope is shaped by the amplitude level of the various frequencies, changing through time. Inverting this shape should reduce the level of some frequencies, while making others more prominent. So what should be most noticeable would be a timbral change in the sound, perhaps deeper and richer in character. But it will be important to explore the possibilities by submitting different types of source to it, ranging from strong, focussed pitches, to sounds of a more amorphous nature.
End of STRANGE INVERT
Return to List of STRANGE functions at
top of this file
Return
to Main Index for the CDP System.
Return
to Spectral Index
1 Shift the whole spectrum
2 Shift the spectrum above frq_divide
3 Shift the spectrum below frq_divide
4 Shift the spectrum only inside the range frqlo frqhi
5 Shift the spectrum only outside the range frqlo frqhi
infile – input analysis file made with PVOC
outfile – output analysis file
frqshift linear shift in Hz of spectral frequencies (same for all); for downward shifts, use -frqshift, i.e., a negative value (Range: -22050 to 22050)
frq_divide frequency at which the shifting starts or stops
frqlo frqhi define a range inside or outside of which shifting takes place
-l logarithmic interpolation between varying frequency values (Default: linear).
(The -l option is useful only if any of the above input parameters are time-varying, AND frqshift values must be all positive or all negative for this to work.)
frqshift, frq_divide, frqlo and frqhi may vary over time
The fact that this shift is linear means that it will produce inharmonic results: it will alter any harmonic relationships between partials. The linear (i.e., additive) change distorts harmonic relationships and musical intervals, which are based on multiplication. An upwards shift reduces the intervals between partials, while a downwards shift stretches them.
This function can be used to create major alterations in the timbral character of a sound. A shift upwards will emphasise high frequencies in a way which makes a dense, inharmonic sound. A shift downwards will produce a deeper, more gong-like sound.
SEE ALSO: SPECFNU Mode 13, which frequency-shifts partials under formants
End of STRANGE SHIFT
Return to List of strange functions at
top of this file
Return
to Main Index for the CDP System.
Return
to Spectral Index
1 Standard spectral stretching for inharmonic state
2 Specify spectral stretching for inharmonic state
infile – input analysis file made with PVOC
outfile – output analysis file
vibfrq the frequency of oscillation, i.e., number of oscillations per second. (Range: 1/infile_duration to 172.265628)
stretch the maximum spectral stretch in the inharmonic state (Range: 1 to 2205)
botfrq the frequency above which spectral stretching happens
expon defines the type of stretch (must be > 0)
vibfrq and stretch may vary over time
STRANGE WAVER introduces an oscillation towards and away from 'inharmonicness' in the spectrum of a sound. The program moves towards inharmonicness by means of the stretch factor. The rate of oscillation, vibfrq, can be fixed or time-varying, using a time vibfrq breakpoint file. The effect can be very much like a frequency vibrato – it depends on the relationship between vibfrq and stretch.
The minimum value for vibfrq is related to the duration of the infile: 1/infile_duration. Thus, durations less than 1 result in a minimum vibfrq which is > 1, and durations greater than 1 result in a minimum vibfrq which is less than 1. Vibfrq will always be greater than 0. The 0.32 minimum value recommended in SoundShaper is for an input file of about 3 sec duration. Longer input files will result in a lower a lower acceptable minimum value. For example, a 10 second file will have a minimum value of 0.1 sec. Whether you would want the effect to operate at this rate is another matter.
Vibfrq values greater than 24 start to produce a 'flutter' effect.
Higher stretch values will raise the perceived pitch of the output.
The stretch factor affects the spectral character of the sound, but does not alter the length – it is not a time-stretch. Take care with the stretch parameter: this function can produce 'wispy' artefacts which higher values will tend to increase.
The vibfrq and stretch factors interact in a very interesting way, which is well worth exploring.
This progam is a useful way to create vibrato and flutter effects, especially via the vibfrq parameter. However, a very low vibfrq combined with a rather high stretch produces something quite different.
End of STRANGE WAVER
Return to List of strange functions at
top of this file
Return
to Main Index for the CDP System.
Return
to Spectral Index