2018 NAMM Show: Frap Tools has announced the Fumana, an all-analog dual filter bank with various controls to access dynamically each individual band.
Each filter bank is composed of an array of 16 independent 48dB/oct bandpass filters. Having two identical filter arrays, allows you to use one of those for spectral analysis (of a modulator signal) and the other for spectral re-synthesis (of the main signal).
Each array is sub-grouped in odd and even bands (numbering these increasingly from 1 to 16 from left to right). It is possible to apply different signal to odd and even bands, both for main and/or modulator arrays. This, together with the 16 envelope followers connected to each band of the modulator filterbank, and the 16 VCA connected to each band of the main filterbank, allows the FUMANA to perform a single 16 bands or a dual 8 bands analog spectral transfer.
While the upper section provide access to each main filter’s direct output, the lower section has 4 voltage controllable parameters to manage two independent functions for further spectral editing: tilt and parametric scan.
The tilt, as the name suggest, tilts the spectrum, adding and subtracting amplitude to the 16 bands, emphasizing lower or high bands, and attenuating high or lower bands: this is managed only with the green potentiometer.
The parametric scan, instead, works with three parameters: peak/notch, band selector and width. Like in a parametric equalizer, where is possible to set peak/notch gain, frequency and slope, the FUMANA uses three parameters to set which is the center band to perform this manipulation, then if that band should be emphasized or attenuated, and how this action is extended to nearby bands.
In addition to the 16 bandpass outputs, there are other 4 outputs provided: the all bands output, the all envelope followers output, the odd bands and the even bands output. These two have the options to be inverted in phase with two switches, in order to obtain even more combination when summed together, maybe/also with the “all bands output”, using, for example the 333 module.
Since a spectral transfer tool may be used as a “vocoder-like-effect”, the FUMANA provides an input for an external noise which may be used for unvoiced (fricative/sibilants) sounds as words containing or starting with s, f, z, ch and other fricative sounds ([s] [z] [?] [t?] [d?] [ts] [?] [f] [v] [?] [?] [?] etc etc).
Features:
- Two 16 bands filterbanks: main filterbank of 16 fixed bandpass filter 48 dB/oct, and a second one only for analysis and creation of envelopes to be used as modulation signals
- parametric band scanning: gain/attenuation control (peak/notch approach), scanning center, scanning width
- capable of true 16 bands analog spectral transfer: if the modulation signal is a voice, some call also call it vocoding
- odd & even input/output: the filterbank can be split in two, using for example two different modulators to modulate a source, vice versa, or using odd bands and even bands independently. This translates also in a fully independent dual 8 band vocoding.
- flat frequency response: from band 2 to 15 we guarantee a tolerance within ±1.5dB (globally ±4dB) and each band’s amplitude is individually trimmable (by technician only). This means that it doesn’t add unwanted colors, doesnt take off that much low or high frequencies… you know what i mean…
- unvoiced section: when used as vocoder, it is possible to add a noise source to this input, which will be added, to the main outputs according to the spectral content. This helps recreating a few timbres that can’t be normally transferred.
Pricing and Availability
The Fumana is expected to be available in February 2018, priced at US $999/999 Euro. See the Frap Tools site for more info.
Remember that somewhere in the 1970’s the Elektor magazine published a series of articles with pcbs allowing everone to build a similar device, which could also be used as a Vocoder. In comparison with existing multiband filter modules today, such as Buchla, this one mainly distinguishes itself by the double array of 16 bands.
Cool for those who like to perform an alternative way of dynamic filtering / vocoding.