[sdiy] Best & fastest envelope follower schematic.. anybody ?

[Jared Anderson]

Hello All,

Seeings as different transforms have come up, I've been meaning to try
using a Hilbert Transform for doing envelope recovery as described in this
Electronotes Application Note:

https://electronotes.netfirms.com/AN355.pdf

I got as far as coding up an example iir based hilbert transformer in
Octave:

https://github.com/jazamatronic/iir_hilbert

Cheers,
Jared


On Sat, 18 Dec 2021 at 13:06, Guy McCusker via Synth-diy <
synth-diy at synth-diy.org> wrote:

> Can you give us a clue on how a wavelet transform would be useful in an
> envelope follower? I don't know much about the area but at first glance it
> does not seem like the right tool for the job.
>
> On Fri, Dec 17, 2021 at 11:25 PM Mike Bryant <mbryant at futurehorizons.com>
> wrote:
>
>> If anybody is thinking of doing this digitally, you might want to read up
>> on the Continuous Wavelet Transform.   Then once you realise your DSP isn’t
>> up to it, have a look at the Discrete Sub Band Transform, aka Wavelet
>> Packet Decomposition.  This gives as accurate results as the Discrete
>> Wavelet Transform with less group delay problems, and given enough
>> processing power is able to decompose even the human voice in real time.
>>
>>
>>
>>
>>
>> *From:* Synth-diy [mailto:synth-diy-bounces at synth-diy.org] *On Behalf Of
>> *Mr&MrsAccount
>> *Sent:* 17 December 2021 21:41
>> *To:* Jean-Pierre Desrochers; 'Tom Wiltshire'
>> *Cc:* synth-diy at synth-diy.org
>> *Subject:* Re: [sdiy] Best & fastest envelope follower schematic..
>> anybody ?
>>
>>
>>
>> Remember that high order filters don't come without their own baggage...
>>
>>
>>
>> They suffer from greater group delay (which could be a problem) and poor
>> transient response (also a problem).
>>
>>
>>
>> Especially in a use where the input could start at zero or suddenly
>> return to zero, the overshoots can be a killer.
>>
>>
>>
>> I've been making filters to isolate the fundamental of Guitars for many
>> (now many mnay) years.   The settling times are longer than the delay of
>> the guitar signal itself, therefore the delays...
>>
>>
>>
>> ------------------------------
>>
>> *From: *Jean-Pierre <jpdesroc at oricom.ca>
>> *To: *'Tom <tom at electricdruid.net>
>> *Cc: *synth-diy <synth-diy at synth-diy.org>
>> *Date: *Friday, 17 December 2021 10:39 AM EST
>> *Subject: *Re: [sdiy] Best & fastest envelope follower schematic..
>> anybody ?
>>
>> > This would be the digital equivalent to the single pole RC. It should
>> be possible to get better ripple suppression by using higher order filters.
>> Harald showed that he got a better response using a fourth-order filter.
>>
>>
>> So.. how can I modify the C1 section to get for example a fourth-order
>> filter ?
>> To do it the way the signal flows across C1 is a bit confusing..
>> The signal comes from the left side but the clamping comes from D1 at
>> right side.
>>
>> I know how to implement high order filters in circuits but only
>> if I know the signal flow direction..
>>
>>
>>
>>
>>
>> *De :* Synth-diy <synth-diy-bounces at synth-diy.org> *De la part de* Tom
>> Wiltshire
>> *Envoyé :* 17 décembre 2021 08:37
>> *À :* Martin Klang <mars at pingdynasty.com>
>> *Cc :* synth-diy at synth-diy.org
>> *Objet :* Re: [sdiy] Best & fastest envelope follower schematic..
>> anybody ?
>>
>>
>>
>> This would be the digital equivalent to the single pole RC. It should be
>> possible to get better ripple suppression by using higher order filters.
>> Harald showed that he got a better response using a fourth-order filter.
>>
>>
>>
>>
>>
>> On 16 Dec 2021, at 16:26, Martin Klang <mars at pingdynasty.com> wrote:
>>
>>
>>
>> I think an envelope follower with no lag is called a rectifier.
>>
>> Jokes aside, there are a number of options available but ultimately it
>> comes down to filtering.
>>
>> For digital envelope followers, I like to use a simple exponential
>> average (aka leaky integrator, a first-order IIR filter) on the fully
>> rectified signal and tune the time constant to the expected input signal.
>> Generally you want it to be fast, but not so fast as to track the waveform
>> oscillations, even at say 20 Hz. This makes a computationally very
>> efficient follower with predictable results.
>>
>>
>>
>> best,
>>
>>
>>
>> Martin
>>
>>
>>
>> On 16/12/2021 01:36, Jean-Pierre Desrochers wrote:
>>
>> Here is my goal..
>>
>> I need a circuit that will take an AC signal and will shift level it from
>> a DC offset extracted from its min & max Peaks..
>> Ex: Suppose you have steady +/- 5v peak sine wave coming in this circuit,
>> The output would be the same waveform but shifted up to +5vdc.
>>
>> So 0 to +10vdc.
>> That final waveform would be much usable for a 0-5vdc control voltage
>> input
>> than the original AC waveform.
>> Now the challenge is to do the same with a level changing AC signal.
>>
>> In the case of a *steady waveform* this ‘level shifter’ could be an
>> opamp that would add half of the peak to peak voltage as an offset DC
>> voltage,
>> But here I’m looking for something that will be more versatile
>> to handle varying level waveform.
>>
>>
>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at synth-diy.org
>> http://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace at synth-diy.org
>>
>>
>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at synth-diy.org
>> http://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace at synth-diy.org
>>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace at synth-diy.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20211218/2087fda7/attachment.htm>