Device Details
Device Overview
| Name/Version: | Hysteresis V2 1.0 |  |
| Author: | OspreyInstruments | |
| Description: | Magnetic saturation (tape/transformer) is subject to a time-variant effect called hysteresis, in which the output voltage may follow different paths along a loop when rising, and falling. They will switch paths only when they have reached a saturation threshold, and change direction. There are complex electromagnetic formulas that require solving in order to "faithfully" model this effect, however, I took a more direct/naive approach and modelled it using offset waveshaper functions. Of course this is less authentic than a designated tape emulation plugin (if you need one, I would point you to ChowTapeModel), however, it does effectively replicate many of the characteristics, and is a useful soundshaping tool. V2 features a completely redesigned hysteresis algorithm. I have migrated from a very basic loop leg switching criterion (exceeding amplitude threshold of -0.99 or +0.99) to one in which the following conditions must be met: 1) Absolute Difference Signal* below defined threshold (default 0.01). This ensures that the signals from the ascending and descending legs have sufficiently converged. 2) Directional change in input signal. The signal is moving from a peak value back towards 0. This selects the correct loop leg: ascending or descending. 3) Absolute Difference Signal* has changed direction from descending to ascending. This is the ideal time to trigger the switch assuming all other conditions are met, as this is the closest together the two loop legs will be, thus minimizing the jump. *Absolute Difference Signal is produced subtracting the output of one loop leg from the other. The absolute value ensures that this is a positive function. Thus, it will peak when the two signals are furthest apart (input amplitude 0, but subject to biasing) with a minimum trough when the signals are closest together (near -1 and +1). The "closeness" of the functions is largely dictated by the window size (wider windows have narrower convergence and will be harder to switch) and the input gain to drive the waveshaper functions (higher "voltages" will map to higher output voltages and square off the curves to create a flatter top and bottom). Other new features include the addition of the second "asymptotic limiter" soft clipping algorithm, which has more rounded saturation curve (although I am also driving it harder). The harmonic response does not fall off as abruptly as tanh() so it ends up being quite a bit brighter. You can now access the switching threshold and switching ramp time (crossfade) directly, as well as a de-emphasis attenuvertion parameter that enables you to choose how much deemphasis you would like (if any), including negative values that will reenforce the EQ moves made in the emphasis stage, rather than inverting them. This way, you can easily dial in the frequency response of the clippers, and selectively determine if there should be any net equalization changes. I will keep Hysteresis V1 up, although I think this version sounds better, so I may switch the algorithms in Spectacle and VibeStrip over to this. |
Device Details
| Tags | effect |
| Live Version Used: | 10.1.18 |
| Max Version Used: | 8.1.5 |
| Date Added: | May 07 2023 05:57:14 |
| Date Last Updated: | Not updated yet |
| Device Type: | audio_device |
| Download URL (report link) | https://ko-fi.com/s/1e1ee62d18 |
| License (more info): | None |
Login to comment on this device.
[ browse device library ]