Jump to content

InClip - A new microphone circuit for ITC

Andres Ramos

Recommended Posts



Fig 1: The InClip microphone in work


In this article I will describe the working principle, technical details and experimental results of a new type of ITC microphone that I call "InClip", a contraction of the term "Inversed Clipper".

Two years ago I did experiments with a circuit that came out of an idea I had by analyzing EVP signals in the time domain. I observed that with some training I could see where in a signal a possible EVP manifestation hides just by watching the fluctuations of the amplitude extremes in a sound editor. It seemed to me that the very tiny pk-modulation hides in those variations of the amplitude peak values.

Because these variations are very tiny compared to the total strength of the signal the resulting modulation is very poor. My idea was to simply cut out the middle area of an audio signal in a way that just the peak values around a specified limit would survive.

Initially I described the concept here.

In that earlier experiment I described and used a hardware circuit doing this job and I applied it to the noise generated by a germanium diode with some success. Now I thought about the question why I never have tried to apply this concept to other sound sources. So I decided to built a microphone that incorporates this circuit to check out the results of different sound sources being processed this way.


The circuit explained


The electronic schematic is shown here and described now in detail.


Fig.1: Electronic schematic of InClip circuit

The frontend contains a simple electret condenser microphone with a preamplifier. I gained very good results with electret condenser microphones and for this reason I am sticking to them in my designs. The LED D2 is used as an "switched on" indicator and to generate a stabilized 2V voltage source for the condenser microphone that generally requires a low voltage supply.

As you can see I used standard operational amplifiers and if you are familiar with my designs you will see many stages you know from my other devices. TL072a is used to provide a virtual ground for the operational amplifiers. Something you always have to do if you use an unsymmetrical power supply. It has no functional meaning for the working principle. TL072b is a preamplifier for the microphone signal that boosts the microphone output by 46dB (factor 150) at maximum.

The switch S2 provides the use of an auxiliary input instead of the microphone in cae you want to process a recorded signal for example.

The preamplified signal is fed into a stage with two comparators (LM339). These circuits are comparing the current signal level against a value the user can adjust with P1. This level is symmetrically to the signal ground, which means that the positive and negative amplitude peaks are handled the same way.

If the signal level stays below that limit the circuit is muting the signal and there is no output. If the level exceeds the given limit by positive or negative peak values, the comparator logic actuates an electronic switch (OP1) that gates the signal through a final amplifier stage to the output socket.


The working principle

One could think that this circuit works like a simple noise gate circuit but this is not the case. A noise gate opens up a conduit for the whole signal swing as long as the averaged signal level stays above a specified threshold. My circuit instead, more or less works like an inverted clipper circuit. A clipper cuts away the parts of the signal that are above a threshold, keeping the parts below the threshold untouched. My clipper works exactly the other way around. It clips away everything that is below a certain level the user can select. The result is a signal composed just of the extreme values of the amplitude.


Fig.2: A signal recorded with InClip

As you can see in Fig.2 the resulting signal looks normal but a bit "ruffled". The modulation of the signal is very strong. This is because InClip cuts away the signal area that not adds up substantially to the overall modulation.


Fig.3: Closeup on InClip signal

In Fig.3 we have a closer look on the signal and you can see that it is composed of amplitude peaks which are the extreme levels of the signal swing. Please see that the clipping works symmetrically, i.e. that positive and negative peaks are arranged nicely around the ground level.


Operating the InClip microphone

Operating the circuit is rather easy. There is one potentiometer P1 to adjust the threshold and a yellow LED that lights up if the signal is going above the threshold. If the threshold is turned down to minimum the signal goes through the circuit almost unaffected. I say ‘almost’ because still the circuit mutes a very small area around the ground level of the signal. Another interesting effect is that the fast switching causes a very subtle chopping effect that seems to improve the pk-modulation. The result is a signal with a very dedicated kind of distortion and a hollow sounding reverb. Basically this kind of signal is what you see in Fig.2+3.

If you turn up the threshold at the potentiometer the resulting signal may change depending on the sound source. Signals that have spikes riding on their amplitude peak levels or signals that are looking "spiky" themselves already, are showing a characteristic you can see in the next pictures.


Fig.4: Spiky signal with threshold turned up


Fig.5: Trains of impulses


Results of the experiments

I tested the circuit in different conditions. First I used the sound of a self oscillating relay as a source. The result was a very spiky signal like shown in Fig.4. The results were nice but not amazing. I post processed the signals basically by the following functions in NCH Wavepad

  • Applying reverb with the shown parameterization
  • Applying 20% auto spectral substraction
  • Effect filter (Teenager/Kid) occasionally


Fig.5: Parameters for reverb effect


Examples of recorded clips




The voices in both clips are not very intelligible and dull. Then I realized that the circuit was not working the way I expected. It generated large spikes always and I quickly found out the the fast rise time of the switching comparators caused glitches that propagated through the voltage supply rail and contaminated the signal output. I fixed this issue by adding two capacitors at the right place (currently not shown in schematic) and continued my experiments.

In my test runs it became obvious that the sound material the microphone picked up plays a crucial role regarding the voice quality and pk modulation, which is no surprise at all. With the background sound of a self oscillating relay I gained some good results from the rather spiky signal.


My interpretation is "Wir lernen jetzt größer"->"We are learning bigger things"


My interpretation is "Du bedankst dich" -> "You said thank you". Actually I really showed my gratitude for the help of my spirits a minute before the recording.


My interpretation is "Arne". This is a name I heard frequently in my session.


My interpretation is "Ich bin Arne "


At the same day I did a last recording session before closing my work in the evening around 06:00 pm local time on Saturday 2022-02-07. I had the threshold turned down so that the circuit would let pass all signal components together with the overlayed switching process I already described. The acoustical background was the sound from the street in front of our appartment with cars passing by and a mix of voices. To my utter amazement I could record a sequence of 30s length with continously speaking voices. To be more specific it was one monoton speaking rather female voice and some sprinkled fragments of other voices. I value them all as spirit voices because of their tonal character and more important the content of their speech.


Here is my transscript from the voices I heared in that clip

1. Timestamp 02:00s - 04:00s

"Oma sucht den Heibel" -> "Grandma is loking for Heibel"

2. Timestamp 09:00s - 11:00s

"One more oven"

3. Timestamp 11:09s - 13:50s

"Ganz leichte Entscheidung" -> "A very easy decision"

4. Timestamp 15:00s - 15:50s

"Welle" -> "Wave"

5. Timestamp 15:50s - 18:00s

"Ist Malu verlogen" -> "Malu is so mendacious"

6. Timestamp 21:80s - 23:50s

"Bei meinen Gruppen fällt er auf" -> "In my groups he stands out"

7. Timestamp 25:50s - 26:00s

"Wir sind bereit" -> "We are ready"

From the content of those messages we can see a clear association with spirits. It is very often they are talking about waves and groups or are expressing feelings. Also the utterance "We are ready" is very common.



From all results I gained, the last one showed is truly the most impressive for me. I was used to get single words or short sentences in a quality that I could understand rather easy or continous speech in a poor quality. This time I could record a 30s clip in a quality I previously just knew from short ITC messages. However I cannot say if this result is repeatable nor if it is really related to the new circuit I tested. Very often I had to face that a sudden success was followed by a fast decline.

It is necessary to continue the experiments with different sound sources and most important trying to replicate the results represented by the last clip.

If my results will continue to be successfull I will post them in a blog probably.


Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.