Implementing a Harmonic Exciter / Treble Boost on ADAU1701

Hello LiuWuXiaobin,

I am not 100% certain what you have in mind. Like many things there will be a few ways to implement it so it will depend on what you like.

This DSP will not be powerful enough to synthesize frequencies to add to frequencies in the signal to add some "air". You will be able to do things using other methods.

The way I see it you have two ways to do it. Either boost only the top end when there is a signal there and let it decline more slowly than the input signal.

So this would be using compressors/expanders.

Another method would be to use an index selectable filter to boost the high frequency only when an input signal is present. Something like that. This would have to be better defined.

I do not have a project that does this and do not have the time right now to experiment with some ways to do this.

I am attaching a project that basically did the opposite. It takes the audio and splits it into four frequency bands and then compresses each band separately which allows you to have different time constants. Then it re-sums the audio and sends it out.

You could experiment with having the compressor on the top end boost instead of compress.

You will want to shut off the noise source I have in this project. This project was to reduce noise not add in some high-end. So I had a noise source for testing.

See if this helps get you started.

Dave T

ADAU1701 Noise Reduction with 4-bands.dspproj

Hi Dave,

Thank you very much for your detailed reply and the project file! I really appreciate your help.

To clarify what I'm looking for: I want to implement an exciter effect as defined in Apple's Logic Pro documentation:

"An exciter generates high-frequency components that are not part of the original signal. It does this by using a nonlinear distortion process, similar to the process used to generate overdrive and distortion effects. However, unlike that process, an exciter passes the input signal through a high-pass filter before feeding it to the harmonic (distortion) generator. This adds artificial harmonics to the original signal. The distorted signal is then mixed with the original dry signal."

This creates new high-frequency harmonics that weren't in the original signal, adding "air" and clarity.

I understand from your previous reply that the ADAU1701 may not have enough power for frequency synthesis. But based on this definition, do you think it's possible to approximate this effect using:

A high-pass filter to isolate high frequencies

An overdrive/distortion block to generate harmonics

A low-pass filter to clean up the distortion

Then mix back with the original?

Or is there a better architecture you'd recommend?

Thanks again for your time and expertise!

Hello,

For the second part, I recommend looking at the non-linear processors available in SigmaStudio.

https://wiki.analog.com/resources/tools-software/sigmastudio/toolbox/nonlinearprocessors

The "Standard Cubic" might be useful for creating tube-style distortion. I researched the transfer function of typical tubes, and since the formulae are cubic, they translate well - provided that you derive the correct coefficients.

You might also get a lot of ideas by looking at the patents for the Aphex Aural Exciter, which no doubt inspired the Logic Pro exciter effect.

See United States patent 2,866,849 "Apparatus for Improving Sounds of Music and Speech," patent 4,150,253 "Signal Distortion Circuit and Method of Use", and patent 5,424,488 "Transient Discriminate Harmonics Generator"

The high-pass filter and low-pass filter sections should be obvious to implement in Sigma.