|Home · Standalone · rtcmix~ · iRTcmix · Tutorials · Reference|
an open-source digital signal processing and sound synthesis language
|about · links · contact|
VOCODESYNTH -- oscillator-bank channel vocoder
quick syntax:VOCODESYNTH(outsk, insk, dur, AMP, nfilts, CFREQLO/FTABLETRANSP, CFREQMULT/FTABLEFORMAT, transp, filtbw, windowsize, smoothness, threshold, attack}, release, hisigmix, hifreq, inputchan, PAN, WAVETABLE[, SCALINGTABLE, CFREQTABLE])
CAPITALIZED parameters are pfield-enabled for table or dynamic control (see the maketable or makeconnection scorefile commands). Parameters after the [bracket] are optional and default to 0 unless otherwise noted.
p0 = output start time (seconds) p1 = input start time (seconds) p2 = duration (seconds) p3 = amplitude multiplier (relative multiplier of output signal) p4 = number of filters p5 = if p4 > 0, lowest filter center frequency (Hz/oct.pc) if p4 == 0, transposition of function table (oct.pc) p6 = if p4 > 0, center frequency spacing multiplier (> 1) if p4 == 0, function table format (0: octave.pc, 1: linear octave; Hz if > 15.0) p7 = amount to transpose carrier oscillators (Hz/oct.pc) p8 = filter bandwidth proportion of center frequency (> 0) p9 = power gauge window length (seconds, try 0.01) p10 = smoothness -- how much to smooth the power gauge output (0-1, try 0.5) p11 = threshold -- below which no synthesis for a band occurs (0-1, try 0.0) p12 = attack time -- how long it takes the oscillator for a band to turn on fully once the modulator power for that band rises above the threshold (seconds, try 0.001) p13 = release time -- how long it takes the oscillator for a band to turn off fully once the modulator power for that band falls below the threshold (seconds, try 0.01]) p14 = amount of high-passed modulator signal to mix with output (amplitude multiplier, try 0.0) p15 = cutoff frequency for high pass filter applied to modulator. This pfield ignored if p10 is zero. (Hz, try 5000) p16 = input channel p17 = pan (0-1 stereo; 0.5 is middle) p18 = pfield reference for wavetable to use p19 = pfield reference for table giving the carrier scaling curve, as [freq, amp] pairs p20 = pfield reference for table giving list of center frequencies (if p4 is zero) p3 (amplitude) and p17 (pan) can receive dynamic updates from a table or real-time control source. p5 (cfreqlo/ftabletransp), p6 (cfreqmult/ftableformat), p18 (wavetable), p19 (scalingtable, if used) and p20 (center freq table, if used) should be references to pfield table-handles. Author: John Gibson, 8/7/03.
VOCODESYNTH performs a filter-bank analysis of the input channel (the modulator), and uses the time-varying energy measured in the filter bands to trigger wavetable notes with corresponding frequencies (the carrier).
Usage NotesThis is a bit different from the other VOCODE-family instruments like VOCODE2 and VOCODE3 in that it does a resynthesis of the signal. This instrument is similar in some respects to PVOC, but it is a channel vocoder using a bank of band-pass filters instead of an FFT analysis (like with phase vocoders).
If parameter p4 ("nfilts") is > 0, then p5 ("CFREQLO/FTABLETRANSP") is interpreted as the lowest center frequency of the filters. This can be specified in Hz or oct.pc, < 15.0 is the point at which it will be interepreted as oct.pc. p6 ("CFREQMULT/FILTMULT") will be interpreted as a multiplier of p5 -- it will space the center frequencies of all the filters (up to "nftilts", p4) as a multiple of each succesive center frequency. For example, p6 = 2.0 will make a stack of octaves; p6 = 1.5 will make a stack of perfect (Pythagorian) fifths. To get stacks of an equal tempered interval (in oct.pc), you can use the cpspch convertor:
p6 = cpspch(interval) / cpspch(0.0)However, if p4 == 0, then p20 ("CFREQTABLE") has to be present. It is interpreted as a reference to a table containing the oct.pc center frequencies of all the filters used. Parametet p6 ("CFREQMULT/FTABLEFORMAT") then serves as a flag for how the table referenced in p20 will be interpreted. If p6 == 0, then the table will be interpreted as containing oct.pc pitch values. If p6 == 1, then the table will be interpreted as linear octave values. NOTE: This only applies if the values in the table ar < 15.0; otherwise they will be interpreted as direct frequency values regardless of the setting of p6. The number of filter bands is determined by length of function table.
An example of this second method for center frequency specification:
Parameters p14 ("hisigmix") and p15 ("hifreq") allow some of the original signal to appear in the output. This is helpful in capturing sharp noise transients (such as consonants in speech) that aren't captured well by standard FFT analysis. p15 is ignored if p10 is set to 0.
p18 ("WAVETABLE") is a reference to a waveform table that will be used by the internal wavetable oscillators.
p19 ("SCALINGTABLE") is a table used for scaling the carrier notes, interpreted as [frequency, amplitude] pairs.
VOCODESYNTH produces both stereo or mono output.
Sample Scoresone example:
rtsetparams(44100, 2) load("VOCODESYNTH") rtinput("mysound.aif") inskip = 0.0 dur = DUR() amp = 4.0 env = maketable("line", 1000, 0,0, .1,1, dur-.1,1, dur,0) numbands = 25 lowcf = 300 interval = 0.025 cartransp = 0.00 bw = 0.009 winlen = 0.001 smooth = 0.98 thresh = 0.0001 atktime = 0.001 reltime = 0.01 hipassmod = 0.0 hipasscf = 2000 carwavetable = maketable("wave", 10, 20000, "sine") scale1 = 0.5 scale2 = 1.0 scalecurve = maketable("curve", "nonorm", 100, 0,scale1,1, 1,scale2) spacemult = cpspch(interval) / cpspch(0.0) VOCODESYNTH(0, inskip, dur, amp * env, numbands, lowcf, spacemult, cartransp, bw, winlen, smooth, thresh, atktime, reltime, hipassmod, hipasscf, inchan=0, pan=1, carwavetable, scalecurve) cartransp += 0.001 spacemult += 0.002 VOCODESYNTH(0, inskip, dur, amp * env, numbands, lowcf, spacemult, cartransp, bw, winlen, smooth, thresh, atktime, reltime, hipassmod, hipasscf, inchan=0, pan=0, carwavetable, scalecurve)
See Alsomaketable, CONVOLVE1, LPCIN, PVOC, SPECTACLE, SPECTACLE2, SPECTEQ, SPECTEQ2, SPECTACLE, VOCODE2, VOCODE3