Fehrenberg Audio DynCurve [WiN]

• Product: DynCurve
• Developer: Fehrenberg Audio
• Version: 1.0.6
• Format: VST3
• Requirements: Windows 10 or later
• Source: fehrenberg-audio.com/dyncurve.html

DynCurve is a dynamics processor built entirely around a freely editable transfer curve — input level on the X-axis, output level on the Y-axis — with no fixed threshold, ratio, or knee parameters. Compression, expansion, upward expansion, and gating coexist on one continuous curve, shaped by placing and adjusting control points. Independent curves for Stereo, Mid, and Side processing run simultaneously, each with separate envelope detection and sidechain conditioning. A dual release architecture, blendable RMS detection, and an asymmetric envelope mode extend the dynamic model beyond conventional peak-follower compressor behavior. The envelope follower’s output is displayed in real time behind the curve editor, making the relationship between incoming signal and gain change directly observable. DynCurve answers the search for threshold-free, visual dynamics control across stereo, M/S, and multichannel formats.

Key Takeaway

DynCurve replaces the threshold/ratio/knee model with direct transfer function editing — a workflow that reads faster for engineers who think in gain relationships across level ranges, and slower for engineers accustomed to parameter-per-function compressor interfaces. The gain change meter shows both attenuation and amplification simultaneously, which matters when the curve applies upward expansion in one level range and compression in another. Producers working in stereo-only sessions with no M/S or multichannel requirements get the full curve editing system but leave the multichannel soundstage architecture unused. AAX inclusion makes it viable in Pro Tools mixing chains where most curve-based dynamics tools are absent entirely.

Transfer Curve Architecture

A conventional compressor applies gain reduction when a signal exceeds a threshold, with a ratio that determines how much reduction is applied above it. Threshold, ratio, and knee together describe a specific shape of transfer function — but that shape is only one of an infinite set. DynCurve exposes the transfer function directly. Any shape is achievable: a curve that applies gentle compression between −30 and −18 dBFS, upward expansion below −40 dBFS, and a hard ceiling above −6 dBFS occupies three conventional dynamics behaviors on a single control surface.

Control points anchor the curve at specific input-output level pairs. Segment behavior between points switches between spline interpolation — which produces smooth, curved transitions — and linear, which produces straight-line connections between points. The Shift key snaps any point to the 1:1 diagonal, restoring unity gain at that input level without numerical entry. Right-clicking anywhere in the editor opens the Curve Ruler, which projects the cursor position onto the transfer function and displays the corresponding input level, output level, and gain difference at that point. This resolves the primary precision concern with freehand curve editing — without it, placing the curve at a specific gain reduction amount at a specific level requires repeated adjustment and metering confirmation.

The limitation of the model is its departure from muscle memory. Engineers who reach for threshold and ratio as primary parameters require a conceptual reframe before the curve editor produces results faster than a conventional compressor interface. Presets cover standard compression shapes — the curve representations of knee, ratio, and threshold exist within the editor — but the building blocks are unfamiliar. The first session with DynCurve takes longer than the second for this reason.

Envelope Follower Visibility

Dynamic processing operates on a control signal derived from the audio, not the audio itself. Attack and release times determine how that control signal follows the waveform — a fast attack captures transient peaks before they pass; a slow attack lets them through while the gain reduction establishes around the sustain. The control signal shape determines what the transfer curve actually processes. Two engineers applying the same curve to the same audio with different attack times process structurally different signals.

DynCurve displays the envelope follower’s output in real time behind the transfer curve editor. Every attack, release, and RMS adjustment immediately shifts the visible signal waveform within the same display that contains the curve. The curve’s placement relative to that signal becomes directly observable — whether the curve’s compression region intersects the signal’s sustained body, its transient peaks, or its noise floor is visible without switching to a separate meter. Adjustments to sidechain low cut, stereo link, and RMS mix all update the same display.

The attack stage offers two characters in addition to time control. Snappy mode accelerates the envelope follower’s initial response at the onset of a rising signal — a faster effective transient capture at the curve’s operating range. Relaxed mode follows a smooth, even rise that preserves more of the transient relationship and produces more transparent compression behavior on sustained material. Neither mode changes the attack time parameter itself; both change the shape of how the follower reaches that time.

Dual Release Structure

The release stage runs two independent time constants simultaneously and blends between them. Fast local recovery from short peaks coexists with a slower return to baseline for longer, denser passages. A signal with brief transient peaks followed by sustained body material — snare hits in a mix, words in a vocal phrase — drives the fast release constant upward quickly after each peak while the slow constant holds the compressor’s engagement through the sustained portion. Without dual release, a single time constant either recovers too fast between transients, producing audible pumping, or too slowly after, producing sustained gain reduction that stiffens the transient attack of the next event.

Two modes further shape the release shape in the logarithmic domain. Logarithmic release calculates decay in dB over time, which mirrors perceived loudness changes and produces a release that sounds natural and self-effacing across general program material. Linear release calculates decay in linear amplitude, producing a more direct initial drop that can read as punchier on rhythmic material — the gain recovers faster at the top of the range where the compression was heaviest, then slows as it approaches the baseline. The difference is audible on drum bus and percussion processing more than on sustained sources.

No secondary release speed parameter is exposed directly — the blend between fast and slow constants is configured internally to provide the dual-release behavior without additional controls. This limits direct access to the exact fast/slow ratio, which matters for engineers who want surgical control over the pumping/smoothness tradeoff at specific tempos.

Mid/Side Curve Independence

Three independent transfer curves operate simultaneously — Stereo, Mid, and Side — each with its own envelope detection chain. The domain selector at the bottom of the interface switches the editor between curves. Switching domains does not pause processing; all three curves operate continuously regardless of which is displayed. Stereo curve edits affect the summed stereo signal. Mid curve edits affect the center content. Side curve edits affect the difference content, which governs perceived stereo width.

A Side curve with an expansion shape below a specific input level amplifies low-level ambient content in the stereo field while leaving the center signal untouched. A Mid curve with a compression shape tightens center-channel dynamics without narrowing the stereo image. A Mid→Side sidechain routing option takes the Mid signal and uses it to drive the Side domain’s detection — the Side curve responds to what the center is doing rather than what the sides are doing. The practical application is unconventional spatial processing: a loud center event compresses or gates the Side signal, reducing width during dense center activity and restoring it during gaps.

The limitation is that the three curves share the same envelope detection parameter set per domain — attack, release, and RMS settings are configured per curve independently, but within each domain, a single attack time governs all input levels. A curve that applies expansion at low levels and compression at high levels on the Side domain uses the same attack time for both behaviors. Engineers who want different transient response at different gain change stages within the same domain must route two instances with separate curves in parallel.

Asymmetric Envelope Processing

Asym Mode activates separate envelope followers for the positive and negative halves of the waveform. Conventional dynamic processors follow the full waveform with one detector, treating positive and negative excursions symmetrically. Asymmetric detection introduces a difference between how the compressor responds to upward and downward signal movement — the gain reduction applied during the positive excursion differs from what applies during the negative excursion at the same signal amplitude.

At subtle settings, Asym Mode produces a coloration comparable to the asymmetric clipping behavior of certain analog circuits — a slight tonal thickening that registers as harmonic density without obvious distortion. At aggressive settings, the asymmetry becomes audible as a distinct saturation-like character with the timing and level response of a compressor rather than a waveshaper. The behavior is adjustable through the same attack and release parameters as normal operation, keeping Asym Mode’s character within the same envelope framework rather than introducing uncontrolled distortion.

The category of processing Asym Mode occupies is not cleanly compression or saturation — it applies gain change driven by dynamic detection, asymmetrically, in real time. Engineers reaching for a single tool to address both dynamic control and subtle harmonic coloration on a single insert find the combination efficient. Engineers who want clean, transparent dynamics processing and separate harmonic control should leave Asym Mode off; harmonic content produced by asymmetric detection is not independently adjustable from the dynamic response that generates it.

Multichannel Soundstage Mapping

DynCurve handles multichannel formats — 5.1, 7.1, Atmos, and other spatial audio configurations — by converting the full channel array into a single unified stereo soundstage and deriving a global Mid and Side signal from it. Every channel pair maps into a shared spatial representation, and the same three-curve system that operates on stereo material applies to the entire multichannel bus through this conversion. The Side curve modulates the stereo width dimension of the spatial image regardless of how many physical channels carry that image.

This is the only identified dynamics processor that applies M/S processing to surround and immersive buses through a stereo-derived soundstage model. Conventional multichannel processors apply gain change per channel or per channel group; DynCurve applies it across the spatial Mid/Side axis of the complete image. Center and LFE channels carry toggle options that include or exclude them from the sidechain detection signal — preventing dialogue peaks from triggering compression on the full mix, or preventing sub-bass LFE content from driving excessive detection activity, without removing those channels from the gain change output.

The production context where this system applies is narrow: mix engineers and re-recording mixers working in surround or immersive formats who want width-based dynamic control at the bus level. In stereo-only sessions, the multichannel architecture adds no sonic capability over the three-curve M/S model, which functions identically on stereo material. The stereo use case is the primary one at the current market stage of the plugin — most purchasers will engage the multichannel system occasionally rather than as the daily working mode.

FAQ