Cloning the Brainworx AMEK EQ 200: Oversampling, FIR Filters, and Distortion

#Audio#Plugins#Cloning#EQ#Oversampling#FIR Filter#Convolution#Harmonic Distortion#Reaper#AMEK EQ 200

The Brainworx AMEK EQ 200 is a popular digital equalizer known for its musicality and analog-inspired workflow. But what’s actually happening under the hood? In this project, I set out to analyze and clone its core sonic characteristics – the EQ curves, the oversampling filter, and the harmonic distortion – using a chain of other plugins.

Watch the full process in the video below:

This post summarizes the analysis, the cloning steps, and the tools involved.

Analyzing the AMEK EQ 200

Before attempting a clone, we need to understand the target. Using analysis tools (like Plugin Doctor), I examined the AMEK EQ 200, focusing on these aspects (excluding TMT, filters, mono maker, etc.):

  1. Harmonic Distortion: The EQ adds predominantly third-order harmonic distortion with a flat frequency response. This is relatively simple to replicate.
  2. Oversampling: By creating an EQ boost near the Nyquist frequency and comparing the curve shape (especially its symmetry) to another EQ with adjustable oversampling (like MeldaProduction’s MEqualizer/StandardEQ), I determined the AMEK EQ 200 likely uses 4x oversampling. Standard digital EQs exhibit curve warping near Nyquist, which oversampling mitigates. Matching the AMEK’s curve symmetry required the 4x setting on the reference EQ.
  3. Oversampling Filter: The frequency response showed a steep filter near Nyquist, the phase response was nearly linear, and the impulse response was symmetrical. These are hallmarks of an FIR (Finite Impulse Response) filter, commonly used for oversampling in high-quality plugins. To accurately clone the EQ’s sound, this filter’s sonic signature needs to be replicated.

Building the Clone Chain in Reaper

The goal was to create a plugin chain that would null as closely as possible against the original AMEK EQ 200 plugin. Here’s how the clone chain was constructed:

  1. Replicating the FIR Filter: Since coding an identical FIR filter isn’t practical for this exercise, I captured the AMEK’s filter characteristic using convolution. An impulse response (IR) recording of the filter was made.
    • This IR file (download here) was loaded into a convolution plugin (like Reaper’s ReaVerb) at the end of the clone chain.
    • Latency Compensation: The IR introduces latency (half the IR length, in this case, 512 samples for a 1024 sample IR), so an equivalent delay had to be added to the track running the original AMEK plugin for accurate null testing.
    • Level Matching: Minor volume adjustments were needed to precisely match the level change introduced by the filter/IR.
  2. Replicating the EQ Bands:
    • I used a flexible parametric EQ plugin (like MEqualizer/StandardEQ) set to 4x oversampling (matching the AMEK’s deduced rate).
    • Matching the specific curve shapes required significant trial and error, adjusting frequency, gain, and Q values.
    • Gain Compensation: Boosting with the AMEK EQ also increases the overall level slightly, so gain compensation was added to the clone chain to match this behavior.
  3. Replicating Harmonic Distortion:
    • The free M Saturator from MeldaProduction was used to add the third-order harmonics.
    • Oversampling M Saturator: To prevent aliasing from the distortion process itself, M Saturator was hosted inside MetaPlugin (also from Melda), allowing internal oversampling of just the saturator.
    • Again, trial and error and gain compensation were needed to match the distortion amount and level generated by the AMEK’s THD circuit.

Null Testing Results

By placing the AMEK EQ 200 on one track and the complete clone chain on another track (with its polarity inverted), we can listen to the “difference” signal. A perfect clone would result in silence (a perfect null).

  • With just the filter IR, delay, and level compensation, the null was already extremely deep.
  • Adding the matched EQ resulted in a difference signal below -90 dBFS LUFS relative to the source.
  • Adding the matched distortion resulted in a final null difference of around -77 dBFS LUFS.

How significant is -77 dB? The video demonstrates this difference is equivalent to a volume change of just 0.0012 dB – practically inaudible and confirming a very successful clone.

Conclusion

While creating a bit-for-bit identical clone often requires custom coding, this exercise shows that by carefully analyzing a plugin’s components (EQ shapes, oversampling, filter characteristics, distortion) and using readily available tools (flexible EQs, convolution, saturation, plugin hosts), it’s possible to replicate its core sonic behavior with remarkable accuracy. It’s a great way to learn more about what makes specific plugins sound the way they do.