honk The EVP "Honk"

[Andres Ramos]

This experiment was more of a joke but I also wanted to see how randomly varying low frequency impulses would be shaped while going through a mechanical spring. This is part of my current work to investigate the transformation of sound by spirits.

I found out that a spring adds reverb and a bandpass filtering to a signal going through it, much like the old reverb spirals for guitars back in the 70s.

As a source of impulses I used a self oscillating relay and connected a spring to the vibrating anchor. The other side of the spring was hooked into the mylar membrane of a cardboard tube.

Depending on the amount of mechanical 'bias' tension the spring behaved very different. Also the mechanical tension it imposed on the relay made the whole arrangement very agile. With some tensions I got different standing waves in the spring I assume and I could make it emit strong or very low sound.

After all the pk effect was minor compared to the roaring sound. In another recording I could isolate some voices at least perceivable without post processing but accompanied by strong sound.

I don't think I could make an ITC device with satisfying results from this setup but at least it was interesting how sound can be shaped by only applying mechanical means.

Bedanke.wav

Dankeschön.wav

Es geht.wav

Hier ist nichts neues.wav

 

[Andres Ramos]

I extracted the audio track from the video and could find some fairly good but short voices in it.

"Wir möchten mit Daniel sprechen" -> "We want to talk to Daniel": Wir möchten mit dem Daniel sprechen.mp3

"Mit der NASA" -> "With the NASA": Mit der NASA.mp3

"Lachen" -> "Laughter": Lachen.mp3

"Freiheit" -> "Freedom": Freiheit.mp3

This seems to be an english speaking voce but I could not understand the content. Maybe you can help?: English voice.mp3

[Arizona EVP]

Ahhhh - the old Ring Modulator concept.

Nice

[Andres Ramos]

3 minutes ago, Arizona EVP said:

Ahhhh - the old Ring Modulator concept.

Nice

I didn't know that this effect is entangled with the ring modulator.

[Lance Reed]

Very good experiment Andres!!  

[Andres Ramos]

8 minutes ago, Lance Reed said:

Very good experiment Andres!!  

Honk, the ghost!

[Lance Reed]

Just now, Andres Ramos said:

Honk, the ghost!

[Arizona EVP]

3 hours ago, Andres Ramos said:

I didn't know that this effect is entangled with the ring modulator.

Ooops - My apologies Mr. Ramos.

My old age and mixing memories got me confused 

You are correct. I was thinking of a spring box which does do reverb.

Ring mod does the voice of Dr. Who's Dalek.  Exterminate - Exterminate 

[Andres Ramos]

32 minutes ago, Arizona EVP said:

Ooops - My apologies Mr. Ramos.

My old age and mixing memories got me confused 

You are correct. I was thinking of a spring box which does do reverb.

Ring mod does the voice of Dr. Who's Dalek.  Exterminate - Exterminate 

Now I understand. Thank you for clarifying this Ron!

[Andres Ramos]

A very good idea Jeff. It would just mean to switch on and of the relay randomly since it oscillates by itself. I could try this by hand first to see what results come out before I make up my mind how to do this automatically. Do you have an idea for an analog circuit that produces random impulse with varying lengths? 

[Andres Ramos]

This device is somehow mysterious. I made more tests today with a different acoustic resonator. Basically not very satisfying since it resonates on a very low frequency, thus maybe a tube with decreased diameter and length would be better. I looked for voice at a point where I detected speech rhytm and really got two voices, however they have a very low frequency.

Voice.wav

Du misst.wav

Then I tuned the tension in the spring a bit and the output turned into a humming tone. After analyzing the recording I could faintly hear a slowly singing spirit like a monk singing a prayer.

This is the whole sequence: Singing voice.wav

Singing spirit.wav

I processed it a bit with moderate denoising but it didn't improve much:Singing spirit Processed.wav

The most clear part is at the end of this sequence:Nur im Regen.wav

 

[Andres Ramos]

Now I found a sequence that shows a very faint sound like singing angels.

Original sound:Angel.wav

Processed: Angel processed.wav

[Keith J. Clark]

I hear the "voice" clip as English. 

My interpretation is "excellent, you can hear us."

[Andres Ramos]

16 hours ago, Keith J. Clark said:

I hear the "voice" clip as English. 

My interpretation is "excellent, you can hear us."

Thank you Keith for your interpretation! This could very well be correct.

[Andres Ramos]

More experiments today with a different tube I made from a tin can where I had cut off the bottom and glued in a membrane. Will put photos later. The device has different modes of oscillation. With a very low tension it emits something really close to a deep human voice. It seems to be my first successful realization of an artificial pharynx excited by glottal impulses. I never heard something comparable.

Gaga.wav

Du hast gleich den grossen.wav

Und ist gleich dein Lehrling.wav

I got some very excited messages from our friends. They said now we will kick of a revolution. We'll see.

[Andres Ramos]

Thanks Jeff. I am confident that there is more to come. I am working on a simple prototype now. Of course I could do some more optimization with the tube which is in fact a Helmholtz resonator, but I see that the careful setup of the interconnecting spring is the crucial factor and I want a setup where I can investigate all modes of operations specified by the spring tension most exactly.

Moreover this apparatus is very close to my vision of a machine built from occult but simple techniques, that does have that connection with the human soul  we already talked about.

[Andres Ramos]

My prototype of the "Honk" is finished. It works more or less as expected. here are some pictures

Fig.1: View on complete device

The brown extension is the outer part of the resonance tube. On the left is a screw handle to adjust the tension of the spring.

 

Fig.2: Inner view

Here you can see the screw with the knob on the left side. It's moving a block with the self oscillating relay on it. The spring conveyes the oscillations to the mylar membrane. You can see that the resonator is a drink can from lipton's ice tea

First test gave me some good results depending on the tension adjustment.

• "Wer zahlt die Zeche?" -S "Who is going to pay the bill?": Wer zahlt die Zeche.mp3
• A longer sequence with a talking voice: Speech Sequence.mp3

Both audios where processed with slight automatic denoising in WavePad.

[Sfay]

Great results.  Definitely hear the cadence of speech.   Like the prototype too!

[Karyn]

I admire your quick work, Andres. Impressive indeed. Happy to see you used almost 100% eco water. Perhaps you could come to Australia. Under my house is an unlimited supply of recyclable materials, even down to the silver flashing that was "kept in case" when we removed the fireplace.  (It is silver and heavy). My daughter was the DIY, and she must have got it from a sperm donor. He was a University-trained engineer. When Emma mentioned "my hoarding," I told her I considered it a precious asset or two - well, actually more than two for her future. 

As she was the major ecologist in our home, who wished we could go back to Amish values/living and regularly clean the sides of the roads we did on our walk, the apple, I can say, did not fall far from the tree, lol.

[Andres Ramos]

Thank you all for your kind comments! Yes, I learned to use 'waste' as a resource for materials in my childhood. My father taught me that. He was a skilful man who always produced solutions from thin air. He seldomly bought materials. I learned from him and found out that trash is a well of prefabricated tools that can be utilized in many ways, especially tin cans can serve as housings or containers for materials, all kind of electronic components can be recycled from old PCBs, all kind of wood like beams and boards serve as base material for mechanical setups a.s.o.

Despite the good results I came to the conclusion that a cardboard tube gave better results. Hence I will remove the tin can and replace it with cardboard tubes recycled from toilet paper or household cleaning paper. And I will prepare the tube for making it longer by stacking more tubes. This was a proposal from Jeff to find out the optimum tube length to reach a resonance as close as possible to the human pharynx.

[Andres Ramos]

5 hours ago, Karyn said:

I admire your quick work, Andres. Impressive indeed. Happy to see you used almost 100% eco water. Perhaps you could come to Australia. Under my house is an unlimited supply of recyclable materials, even down to the silver flashing that was "kept in case" when we removed the fireplace.  (It is silver and heavy). My daughter was the DIY, and she must have got it from a sperm donor. He was a University-trained engineer. When Emma mentioned "my hoarding," I told her I considered it a precious asset or two - well, actually more than two for her future. 

As she was the major ecologist in our home, who wished we could go back to Amish values/living and regularly clean the sides of the roads we did on our walk, the apple, I can say, did not fall far from the tree, lol.

I feel we will come back to this more conscious way of living our lives and managing our resources, just because we can't go on like this anymore. And I think Corona was the first milestone on this way. This virus was a punch in the face for global trade. Yesterday I looked for the postage to send a parcel to Canada. I read that Corona disrupted so much of the delivery chains that sending by air freight can't be provided anymore. The only way is delivering parcels as sea freight and this takes many weeks AND the postage is above 50€ for a 2kg parcel! Logistically we are back on our way to the middle ages. Soon we will start to consume less and to produce what we need more and more in our own countries again or import it from our direct neighborhood. It's the same with the holidays, more and more germans now are passing in their own country again and loving it.

[Andres Ramos]

I trained myself to utilize every fraction of time I can grab for doing this work and if there is no time I still try to prepare the next step by providing the right tools, ordering material a.s.o. Sometimes there are just minutes I am using to quickly solder a cable or glue something together. It's proceedings in fragments.

I had to learn this because the physical plane is burying me in works and duties to keep me away from my ITC research. There you can learn who are the masters of this world and how they are trying to neutralize us.

[Andres Ramos]

Before I leave for vacation I want to post this outview here. Jeff gave me a nice idea how improve the Honk circuit. Since the oscillating relay that generates those glottal sound goes into a weird state of metastability and produces voice like patterns, that behavior could be improved by additionally modulating it with noise or mix it with other sounds. This is possible because in the metastable state the relay behaves more like a crazy going loudspeaker instead like a digital component that only knows the states ON and OFF.

New schematic for the Honk

A modulation voltage is amplified by the transistor T1 that controls the current for the oscillating relay Rel-1. The generated sounds are transformed by the mechanical spring ME2, going through the screw ME4 that does the mechanical bias for relay circuit breaker and finally into the membrane in the resonator tube ME3. The whole arrangement consisting of the oscillating relay, the spring and the membrane with the tube is in fact a metastable system with fragile points of operation where different frequencies are generated that get bounced forth and back between the end points made up by the relay and the resonator tube and are susceptible to any changes.

[CanadaKim]

As per usual Andres, AMAZING. The actual construction of the unit is second to none. You really should try just a tiny bit of "echo to extend the voices a bit.

Did you just use what we in North America as "Saran Wrap" for the Membrane (modulated end of the tube) ?? 

Edited August 16, 2021 by CanadaKim

[Andres Ramos]

14 hours ago, CanadaKim said:

As per usual Andres, AMAZING. The actual construction of the unit is second to none. You really should try just a tiny bit of "echo to extend the voices a bit.

Did you just use what we in North America as "Saran Wrap" for the Membrane (modulated end of the tube) ?? 

Echo or reverb is a very good idea Kim!  It could add up as an improvement to the oscillations in the spring.

Saran wrap I don't know. Need to google this.

[Andres Ramos]

After some discussions with Jeff I tried two improvements today.

1. Changing the relay driver from voltage control to current control. I got better results with this technique with the coherer and wanted to see if it pays off with the Honk too.
2. Modulating the relay with an audio signal, especially with the signal from a microphone in order to establish an acoustical feedback loop.

My first goal was easy to achieve and the results are very good.  While adjusting current and the spring tension the device is going through different modes. It seems to be favorable to adjust the spring in a way that there is no oscillation on a single tone or in other terms, the use as a horn has to be avoided. The voices are hiding in the areas where the systems is in a noise mode and provides a good modulation. Voices have a good pk modulation and a remarkable voice characteristic. I exported some samples. They are structured by first showing the raw signal and after 1s pause the same signal processed. Processing was done by applying 30dB noise reduction by using auto spectral substraction in WavePad and occasionally a 250 Hz high pass. In the following table I provide the decosing translated into english.

• "The building is complete now": Das Gebäude ist komplett.mp3
• "He should..(unintelligible": Soll er.mp3
• "This is a major breakthrough": Das sehen wir einen Riesenerfolg.mp3
• "This is one of..(unintelligible)": Das ist so ein.mp3
• Unknown: Unknown.mp3

The second goal was interesting but gave no usable results in respect of ITC. In fact I got some nice feedback patterns that proved my theory, that the relay, in this mode, is more acting like a loudspeaker. Here are two sample exports, however as I said without voices.

Microphone Feedback.mp3

Microphone Feedback-2.mp3

What I want to try now is to use more low frequency signals that give more interruptions to the oscillating relay in order to form a better speech pattern (segmentation of words).

Also I decided to give the "Honk" another name and I came to the term "VISPRE", standing for "Vibrating Spring and Relay" that also shows a phonetic similarity with "Whisper" what I think is a cool relation.

 

[Andres Ramos]

And here some results with my latest hack. I am exciting the osillating relay with pulses from a NE555 with adjustable duty cycle. The impulses are mixing with the inertia from the VISPRE givin a fairly good entropy. I would not say that the results are really better but there is headroom open in my eyes.

• Vati hat Rabatt.mp3
• Abziehen.mp3
• Als ob das egal wäre.mp3
• Angenäht.mp3
• Ist schon brauchbarer.mp3
• Long sentence.mp3
• Sie wollten Player.mp3
• Und Herr Ramos.mp3

 

[Andres Ramos]

I worked like crazy to finish the first prototype of VISPRE unit 1.

It shows endless modes depending on the adjustable spring tension and the driving current. An additional feature is a microphone that can pick up sounds (also VISPRE's own sound) and modulate the relay with it.

Basically the device has two main modes, the intermittent mode and the continuous mode. Here some examples with intermittent mode. Files with only one content are  NOT processed. Files with two contents separated by a pause are showing the raw signal at the start and the 30dB auto substraction denoised signal after the pause.

 

Da braucht man garnichts zu machen.mp3

Da wird dein Test noch nachwirken.mp3 Das müsste jetzt zumindest ausreichen.mp3 Das werdet ihr bald im Projekt darstellen.mp3 Das wird euch der Bergmann selbst noch zeigen.mp3 Euch wird ja bald der Parkplatz bestellt werden.mp3 Fragen.mp3 Sind bereit.mp3

The 2nd mode is characterized by a steady noisy signal. Here some examples.

Sind Reisen.mp3
Wir sehen Probleme.mp3 Wird besser.mp3

At the end of the session I got one last message. A spirit said "Wahnsinn" -> "This is crazy"

Wahnsinn.mp3

 

[Andres Ramos]

You are certainly right, Jeff. I purchased a set of different springs to try them out and I am making up my mind about a random pulsing of the rely. The intermittent mode turns out to be favorable if the relay can be excited by random pulses mimicking speech patterns. I need to make another test setup for this development path, something where I can easily change and adjust the spring and feed into the relay whatever I want. Something like  a test bench.

The prototype I made is just for "freezing" the current design and to have something to work with and it will serve me as a reference design for the next development steps.

Keith was the one who occasionally came up with the idea of an artificial larynx and pharynx to give the spirits something as close to the human "speech apparatus" as possible. I think he now is more looking for a software solution but it turns out that his idea was right. I'm confident that in the future we will realize a continuous talking apparatus the spirits can use.

[CanadaKim]

Of course, you know that I will want a schematic, altho I do have a 555 board with adjustable on and off times....just to cheat.

555 board.html

[CanadaKim]

How do you keep the spring from tearing through the Diaphragm material ?, and what id the material used for the Diaphragm?

[CanadaKim]

Why dont you use the actual Diaphragm of a small speaker directly?

[Andres Ramos]

15 hours ago, CanadaKim said:

Of course, you know that I will want a schematic, altho I do have a 555 board with adjustable on and off times....just to cheat.

555 board.html 607.24 kB · 1 download

I got the message Schematic is in work.

[Andres Ramos]

21 hours ago, CanadaKim said:

How do you keep the spring from tearing through the Diaphragm material ?, and what id the material used for the Diaphragm?

I used a small sheet of metal behind the diaphragm with two tiny holes in it. The end of the spring is shaped like a hook. The pin of the hook is piercing the diaphragm and metal sheet and goes around it. Thus the metal sheet is pressed against the diaphragm and not the spring itself. The diaphragm is mylar from a zip bag.

[Andres Ramos]

More experiments today. I could achieve the reception of some snarling voices. I used the neon bulb receiver to generate nice random speech patterns. A bit of electronic is shaping impulses for the relay from them.

• 202-09-08 Jetzt hast du mehr gelernt.wav
• 2021-09-08 Bewährt.wav
• 2021-09-08 Der Bruder.wav
• 2021-09-08 Der Harry hat kein Herz.wav
• 2021-09-08 Für Barbara.wav
• 2021-09-08 Meine Erker sind voll.wav
• 2021-09-08 Projekt werden viele Leute begehren.wav
• 2021-09-08 Wir sind da.wav

Hear the audio clips.

[Andres Ramos]

Thanks Jeff. Tonight I had another idea for improvements. We'll see if I can push the borders a little more.

[Andres Ramos]

Some more improvements. I guess now it's time for an updated schematic.

• Bleibst du eine Waise.mp3
• Du hast nah gedrängt danke.mp3
• Jetzt werde ich das bewerten.mp3
• Preis wird weiter steigen.mp3
• Werde Vergleich Morgen.mp3

Here is the new schematic. It will be the base of my new prototype. I will need a bigger box to put this all in since it now includes also the neon bulb with its electronic components.

 

[Andres Ramos]

The weekend is reserved for the lite bridge. So today I worked as hard as I could to proceed with new prototype. It incorporates everything I learned from the Honk and I will make it look really good.

As a teaser here is my work in progress.

[Andres Ramos]

Basically the device is finished but I am still doing optimizations. Today I recorded a longer sequence to become more familiar with the sound and evaluate the capabiity for generating continous speech. Basically the speech is still too fast and sounds like coming from a tin can but I never made a device before that could generate a continous stream of voices with almost no noise.

The following sequence was slowed down by 20%, 250-Hz high pass filtered and 20% denoised. Moreover I inserted 100ms silence gaps between the coherent statements. You will not comprehend anything by hearing the clip in one piece but certainly if you select shorter sequences and play them back in a loop.

Recording 2021-09-20 Processed.mp3

Then I got the idea to route the signal picked up from the internal microphone in the VISPRE through the linguaphone speech filter. The result is very weird.

Lingue 2021-09-20.mp3

[Andres Ramos]

New schematic. I am making small improvements every day.

VISPRE electronic schematic

It turned out to be useful to make the amplification of the impulses adjustable. As a result I don't need thee adjustable threshold for the impulse shaping anymore and the relay driver is now a darlington transistor.

 

[Andres Ramos]

I made a promo video of the VISPRE for one of the upcoming newsletters.

https://youtu.be/oPp0OuE3nIw

 

[Andres Ramos]

Some more refinements today with amazing results. I checked the voice generation with maximum spring elongation of 15mm. In this extreme setting the relay oscillations are shortly before tapering of. This configuration gave very good results too. Another improvement is to not use the external microphone to face the tube outlet directly. Instead it is favorable to place VISPRE and microphone both facing a wall where the sound waves of the VISPRE are reflected from before entering the microphone.

The following audio clips are all in german but I hope you get the quality level and the excellent SNR.

• "Ich bin hier Plakta" -> "I'm here, Plakta": Ich bin hier Plakta.mp3
• "Ich stümperte" -> "I was failing": Ich stümperte.mp3
• "Jetzt bin ich da" -> "Now I'm here": Jetzt bin ich da.mp3
• "Jetzt kriegt er Ärger" -> "Now he's in trouble": Jetzt kriegt er Ärger.mp3
• "Zork verbirgt es" -> "Zork is hiding it": Zork verbirgt es.mp3
• "Er darf - bewilligt" -> "He can do it - allowed": Er darf - bewilligt.mp3
• "Für Vater ist gesorgt" -> ""Father is taken care of": Für Vater ist gesorgt.mp3

First I started the recording with the activated Lingua in order to adjust recording levels. After switching it on I got the message "Du machst das Filter" -> "You're using the filter now": Du machst das Filter.mp3

I think I can state that these are the best spirit messages I ever yould get on a regular base.

[Karyn]

well done Andres

 

[Andres Ramos]

I stole myself some time today because the work with the lite bridge is proceeding very well and thus I realized an idea for improvements with the VISPRE.

I wanted to replace the oscillating relay with a loudspeaker. The sound of the relay is very rich but also a bit dull since it was not designed to create sound.

So I employed a loudspeaker with a piece of perfboard to connect the spring. First impression was that the sound had more granularity but was lacking speech. Then I considered that the relay was self oscillating and the loudspeaker not. Thus it emitted more or less only the envelope of speech.

So I made a little oscillation circuit with an NE555 that generates a square wave signal that can be tuned between 150Hz and 400Hz. The neon lamp impulses are switching the NE555 just on and off. The results are much better now. Here the only english sample I got.

We've been all built with the stars:

We have been all built with the stars.mp3

The fidelity is not bad and the cool thing is that I can tune the pitch electronically now by adjusting a potentiometer on the NE555 circuit. In the current VISPRE the distance between relay and diaphragm must be properly adjusted to pick the right pitch. For the new design I won't need the mechanical screw arrangement anymore and the loudspeaker can be placed fixed.

While granularity is improved the sound is very "thin" now. However I already have ideas for an alternative oscillation circuit. I guess the next version of the VISPRE will be easier in mechanics and better in performance

I use the current VISPRE just for testing and will reestablish the previous state after my tests are completed. The new design will give birth to a new prototype.

 

 

 

 

[Andres Ramos]

Fascinating how the content changes if you apply other processing methods! Now I hear "Geh nach auf dem Wege",  what means "Follow the way".

Your description of the complexity of the interaction of the mechanical parts is very true. I made some attempts today to replace the components by electronic circuits like reverb but it doesn't work. There is something very special going on between loudspeaker/relay, spring and diaphragm. Will try two ways now, first applying a bigger diaphragm and a softer spring with the speaker. Second replacing the self oscillating relay by one without self oscillation that gets excited by the NE555 impulses. This will be a competition between relay and loudspeaker.

[Andres Ramos]

The last days I spent fathoming the possible improvements for the VISPRE. I tried,

• Using a loudspeaker as a replacement for the relay
• Replacing the oscillating relay by a non-oscillating
• Using the neon lamp to trigger electronically generated trains of impulses with adjustable pitch
• Elongating very short impulses in the above mentioned setup
• Adding reverb electronically

In the end I really was pleased to see that all those ideas made the result worse and not better. So, in the end the magic is in the mechanical parts and the oscillation patterns they are generating. I am very pleased that a Edison-style technique outperforms all the electronic stuff.

I have put the VISPRE back into it's original state again. The only thing I added is to another softer spring in series to the original one. It really would be an improvement to make a small hook at the diaphragm in order to connect different springs.

With the softer spring the signal has more reverb and thus has a more richfull sound.

Another improvement is in the setup. It turned out to be favorable to let the acoustical waves being reflected from a wall and to catch this signal with a microphone. Therefor I implemented an electret condenser microphone beneath the resonating tube facing outside.

 

I routed the signal through the Lingua, recorded it in WavePad and applied a high-pass filter (250Hz 6dB) and 20% denoising (auto spectral substraction). Hear the following audio clips. A person named "Ben" was mentioned multiple times.

• "Erik meldet" -> "Erik is reporting": Erik meldet.mp3
• "Es sind Wellen" -> "They are waves": Es sind Wellen.mp3
• "Meldet bitte" -> "Reporting please..": Meldet bitte.mp3
• "We can't stand...":We cant stand.mp3
• "Bitte der Ben" -> "Now Ben please": Bitte der Ben - Ben.mp3
• "Meldet Ben" -> "Announcing Ben": Meldet Ben.mp3

I think a new test setup would be appropriate to optimize diaphragm diameter and spring type.

Addendum: The new recordings with increased lingua processing level and the new spring the results are really good. Got a lot of messages in bavarian dialect.

• "Bringt uns der untere her" -> "The lower one brings us here":Bringt uns der untere Ehr.mp3
• "Brüder Jahn" -> "Brothers Jahn": Brüller ja.mp3
• "Delta hoffen" -> "Hoping delta": Delta hoffen.mp3
• "Du böser" -> "You nasty one": Du Böser.mp3
• "Du kriegst deine Compliance" -> "You will get your compliance":Du kriegst Deine Compliance.mp3

The mechanical oscillations are the only ones that are generating this snaring sounds the spirits use to make formants from.

 

 

 

[Andres Ramos]

With much help from Michael I managed to set up his Spirit Communicator programm on my PC today. Thank you so much for this Michael Lee!

It works great! Basically I get a stream of continous voices but voices still are is not optimized and dull due to the low frequencies in the signal. Here some exports:

• "Könnt ihr denn nicht Programme nutzen" -> "Why you can't use software?":Könnt ihr denn nicht Programme nutzen.mp3
• "Würde dir was vorlesen" -> "Would like to read something for you":Würde dir was vorlesen.mp3
• "Ihr macht so feine Versuche" -> "You all are making such fine experiments": Ihr mach so feine Versuche.mp3
• "Weckt er Welt, erfolgte maßvoll" -> "Awakened the world in due course":Weckt er Welt - erfolgte maßvoll.mp3
• "Er würde dir noch mehr Fragen fragen in der Hölle" -> "He would ask you more questions in hell": Er würde dir noch mehr Fragen fragen in der Hölle.mp3

 

Here is the complete recording: ML 2021-10-12.mp3

[Michael Lee]

The program has a few machine learning models to reverse things like additive noise, and quantization artifacts. Also, it has a primitive "speech model" that adds formants it finds to a "buzzing" glottal pulse. The machine learning models were trained on a 30 min speech of Gen. David Petraeus speaking to Congress, so it's funny when German is trying to be spoken through it.

We'll try to get you "hooked up" with the program. It requires a good CPU. It requires a few steps to install, including Anaconda or Miniconda (A Python environment), and may take about 5-10 GB of disk space - not because my program is particularly large, it's just that Python and end-users don't mix

[Andres Ramos]

Yes Jeff, I talked to Michael Lee yesterday and said I want to help him to promote this great piece of sw-engineering he made. I mean this is a professional tool especially designed for making spirit voices audible that are deeply buried in interference and noise you get for free!

The setup is a little complicated but not that hard to do. I will write a small how-to as a one-pager how to do it and later I will make a video too and share it.

@Michael LeeMaybe you can give me some descriptions for the user parameter settings represented by all those sliders. Then I would work that into my one-pager. I also would like to put your Spirit-Communicator files as an zip archive into the sw section together with my "manual". Would that be ok?

[Andres Ramos]

Michael Lee already wrote a small manual about the parameter settings. I can add an installation guide and this should be enough for a start. We now need to find a place in the forum where we can put the SW for download and everyone can find it.