EStim sound library and EStimSurprise Tease

[mantrid]

EStim sound library
Because some people announced to write new EStim teases, I decided to start a sound library. This project is a part of the discontinued MPGenerator project (which was used for sound generation). The sound definitions are portable and can also be used to build sound data with more than 2 channels.

Starting from version 0.2 the library is delivered with a testing tease. Details can be found at the end of this post.

I will use this initial post for the description of the current version and therefore modify it from time to time. Release specific notes will be send in the following posts.

The plan is to make the library downward compatible starting from v1.0 for the same major version number, i.e. if you use v1.0 you can upgrade to v1.1 without surprises. But there may be minor changes or adjustments.

Version, Download

• All-in-one package containing the tease, library version 1.10 and 1.11 and GServer Download, Release notes

Playback using a browser
Unpack the .zip archive and open EStimSurprise.html using a browser (Ctrl+O).

Playback using gserver
Unpack the .zip archive, change to the directory where you unpacked it (where the file GServer.jar is) and run

Code: Select all

java -jar GServer.jar -v

on the command line.

Then, open a browser and enter "http://localhost:7980/" or "http://127.0.0.1:7980/" into the URL field. Then, configure the sound by clicking on "Configuration" (can also be done in a separate window/tab while playing the tease). The Tease is started by clicking on "EStimSurprise-GServer".

This method is recommended because sound is generated dynamically and can be configured. For example, it is possible to have more than 2 audio channels.

Sound types and configuration
The files does not work if channels are mixed or if the EStim box re-modulates the signals somehow.
Mixing often occurs automatically (if the sound system assumes the channel assignment is not correct). With stereo files this should not happen.

Pre-generated sound files
The zip archive contains pre-generated sound files for stereo devices in .ogg format. They are used in the standalone (browser only) version.

Spoiler: show

Audio generation parameters are

• Channel group configuration: 2
• Phase offsets: 0.5
• Recognition limit in Hz: 40
• Volume compensation: 0.5
• Carrier frequency in Hz: 600

Dynamic sound files
The dynamically generated files are stored in .mps format (the script format of MPGenerator). Audio generation parameters can be changed by clicking on "Configuration" in the browser window (after starting GServer as mentioned above).

Here are a few tips and tricks for configuration:

• Click on the question mark for an explanation of the parameters
• Setting "Channel group configuration" defines the channel topology. The sum of the comma separated values is equal to the number of audio channels.
• Setting "phase offsets" changes the characteristic of the phase modulation effects. With standard male electrode placement (large common electrode at perineum) smaller values feel softer. That should be tried out.
• The other settings are less important
• You can change the settings during playing a tease. Just open the configuration page in another browser window/tab. Changes of the setting must be committed using the "save setting" button and come into effect when the next sound file is played.

Sound File Naming Convention
Each file name consists in one or more letters and three digits or an 'x'. The letters define the type. Digits are used for grouping and sorting.

Digits

• First digit: Difficulty level. Levels of different types can't be compared. Higher numbers are more difficult.
• Second digit: Group number. Files of different groups should feel different.
• Third digit or 'x': Number within group. Files within one group should feel similar. Files ending with x have a duration of 120 seconds and also have a slow volume increase.

Letters
Files that start with one letter are randomized within group but there is a tendency of increasing intensity.

Files that start with two letters are less randomized and (unless otherwise noted) contain an effect that is increasing with increasing last digit. The difficulty level is the level of the first file in group.

Currently these types exist:

• n: Constant effects. Duration: 20s
• ni: Constant effects with increasing intensity. Duration: 20s
• v: Variations. Duration: 30s
• vi: Variation with increasing intensity. Duration: 30s
• vv: Variation with increasing volume (volume training). Duration: 30s. Volume is increased in waves and minimum volume increases from file to file.
• vp: Transition from normal to painful within 20 files / 2 groups. Duration: 30s.
• s: Signals with short interrupts
• sd: Signals with long interrupts (tease + denial; same as 's' but longer pauses)
• x: Shocks / pain signals

Usage Hints

• As mentioned, normal files (one letter) are quite randomized. If you find a file that is more intense than the other ones within the same group, this may change with the next built. If you need files of different intensity either use different groups or levels or use special effect files (that start with two letters)
• There are a lot of short term volume effects, but the only files that increase the volume for a period longer than very few seconds are the files with 'v' as second letter and the files ending with 'x'.

Hints for electrode placement
On this external page there is a good tutorial for electrode placement.

Tease: EStimSurprise
A tease is provided with the library. The sessions are stolen from Brycis EStim Experience, see this post.

There are different levels of difficulty. By pressing the skip button you can go to the next higher level. But you will receive some pain with two exceptions: The first level can be skipped without getting pain, in the highest level you only get pain.

If you beg for mercy or a ask for a break, you go downstairs.

The story is explained by the chief nurse.

Session texts where only minimally adapted but sound is totally new.

Spoiler: show

Cheaters uncoment the line

Code: Select all

var enableDebug=true;

in EstimSurprise.html

[mantrid]

For a general description, please read the initial post. This message contain the
Release notes for version 0.1
This is a preview release which is still incomplete and there may be major changes. Currently it only contains constants effects and variations. All kinds of non-continuous effects (strokes, pain files, stop+go ) are missing.

All sound files are defined by a few scripts which can be parameterized. This allows bulk generation of a large amount of files with limited effort. It would be easy the to build a library with 10000 files. But it's much more difficult find a collection of a few hundred files with all relevant effects that are not to similar.

Currently I'm a little bit numb from all the testing. Therefore any help/feedback is welcome:

• Are there groups, that feel to similar / are to boring / are to easy or to difficult for a certain level?
• Do yo want to have more of a certain effect (single file or group)?
• Do you miss any continuous effects you know from EStim Tower or EStim Distraction? (I do not mean non-continuous effects like strokes or pain files that have not been implemented yet.)
• Do you want any other special effects, e.g. extra long files with increasing effects or transitions?

Hints for Testing:

• Files within group should be similar. Its sufficient if you test 2 or 3 files per group.
• Only test the electrode configuration (sound file type, channel assignment) that works best for you. Test it with 'calib2.ogg' as mentioned in the initial post.

[mantrid]

For a general description, please read the initial post. This message contains the
Release notes for version 0.2

The library should be almost complete but certain things will probably resorted, see known issues below. If you miss something, just inform me.

New features and changes

• There is a tease now which uses the library, see initial post
• Signals with interrupts (I call them "strokes") and shocks have been implemented

Known issues (library and tease)

• Pain files have a random phase shift, more exact: either full phase shift or no phase shift. With 3-pase setups the phase shifted signals are softer than the non-phase shifted ones. The effect is small for the easy pain files, but large for the harsh pain files. (These effects can also be seen in the EStimTower). In future I will split that, i.e. there will be really painful pain files and "pleasant" pain files.
• There is a similar issue for the some variation files (starting with 'v') which contain a shock element.
• Some sound files (lib) / girls (tease) need to be uprated/downrated (you know what I mean if you meet Ash in the 4th floor )

Feedback/help requests

• As in the past: any feedback for sound files (missing effects, to easy/to difficult for a certain level, ...) is welcome
• I recommend to enable debug mode in GuideMe. This shows the image name in to top right corner which can be used to identify sound files, erroneous pages, ...

  Spoiler: show

  Furthermore it allows to skip pages.

• I'm better in programming computers than in programming girls (and maybe I'm also a little bit lazy): I ran out of normal sessions (between 13 and 18 pages) and pain sessions (between 11 and 15 pages) that I can steal. I need them because I plan to split a few effect groups as described above. If someone can write one or a few sessions that would help a lot (just images + text are sufficient).
• I tried to integrate the calibration into the story. If this is not understandable (also read the sound section of initial post) or if there are other issues with the language, please inform me or change this yourself. The Script Engine scripts (.scr files) are easy to read and to edit.

[lolol2]

Tried the tease today a bit with different settings, using my own build stereo device.
First the signals felt very interesting in the calibration phase, I tried both and when I play the 4 channel version with my 2 channel device, I quess the both signals are just playing over each other, it feels better.

But in the first 5min the signals got harsher and harsher... I guess there are to much peaks and the frequency is to low.
What frequencies you are using?

The pain files were way to quite to give me any pain and my device has no signal transformation, so it can hit you very hard and unpleasant normally.

Just a little feedback from around 15-20min.
But nice work!

[mantrid]

Tried the tease today a bit with different settings, using my own build stereo device.
First the signals felt very interesting in the calibration phase, I tried both and when I play the 4 channel version with my 2 channel device, I quess the both signals are just playing over each other, it feels better.

But in the first 5min the signals got harsher and harsher... I guess there are to much peaks and the frequency is to low.
What frequencies you are using?

The 4 channels are usually down mixed to 2 channels if they are played back on a 2 channel sound card. The same may happen if you use a surround sound card but only connect one output (some sound cards detect this situation), thats why I recommend a dummy load, see the initial post.

The sounds have a complex carrier waveform which is power saving, allow phase modulation and avoids interferences if different channels are mixed (I plan to document the technical details in future) The last characteristic should ensure that the signal is even predicable if it is down mixed.

In the first level (most likely first two levels) there should not be any signal that feels hasher or more painful than the calibration signal that is used for channel setup. Unless you cheat you can't reach a higher level.

So either there is a bug in the tease (which girl was to harsh) or your sound card does something odd (volume normalization, interrupts due to software down mixing, ...)

Do you have the same issues with the 4 channels sounds for EStim Tower or Brycis Estim Experience ?

The pain files were way to quite to give me any pain and my device has no signal transformation, so it can hit you very hard and unpleasant normally.

There are several levels of pain signals. The first ones appear between levels 2 and 3 and should not generate much pain. In any case the pain signal should be more painful than the normal signal appeared before.

[sweden123]

This might be a bit too advanced for me. Where do I find sound directories 6us' and '2' ?

I also tried running mpgenerator but only get an error:

Error: Could not find or load main class MPGenerator

[mantrid]

I now also built the redundant 3-phase-2-channel sounds (redundant because they are contained in the 4 channel files) and an alternatitve 3-phase signal which may work better if you are sensitive to currents between the non-common electrodes.

If anyone prefers the alternative signal I will add it to the main distribution in future.

Here is the download link

Unpack the contents of the .zip archive into the directory 'EStim' of the tease such that 'calib.scr' is overwritten.

This might be a bit too advanced for me. Where do I find sound directories 6us' and '2' ?

The sound files are distributed independently. First, unpack the contents of the tease archive into your tease location (by default the directory 'Guides' in your GuideMe installation). Then, change to the directory 'EStimLib' and create the directory 'Audio'. Unpack the contents of the sound library into this directory. You now should have the directories 'ESimLib/Audio/6us' and 'ESimLib/Audio/2' (relative to your tease location).

Download links can be found in the initial post.

I also tried running mpgenerator but only get an error:

Error: Could not find or load main class MPGenerator

Just call

Code: Select all

java -jar MPGenerator.jar

or

Code: Select all

java -cp MPGenerator.jar mpgenerator.MPGenerator

as described in Readme.md (outdated). That prints an error message because no script is defined plus a help text.

[lolol2]

In the first level (most likely first two levels) there should not be any signal that feels hasher or more painful than the calibration signal that is used for channel setup. Unless you cheat you can't reach a higher level.

I mean that the signal typ feels kind of harsher to me than the most estim signals.
Hard to describe but It feels a bit spicky and has not that smooth constant feeling like most of the other signals you can find.
I had fun for around 5min than the signal typ got more and more uncomfortable for me.
I tried both the 2 channel and 4 channel version, the 4 channel version feels like the 2 channel with a little extra variations.
I don't have a 4 channel setup and never tried any other files with 4 channels.

But maybe just personal impression.
Just wanted to give a feedback to your hard work, could not leave that thread uncommented any longer.
Maybe this will start the conversation a bit.

[mantrid]

I mean that the signal typ feels kind of harsher to me than the most estim signals.
Hard to describe but It feels a bit spicky and has not that smooth constant feeling like most of the other signals you can find.
I had fun for around 5min than the signal typ got more and more uncomfortable for me.
I tried both the 2 channel and 4 channel version, the 4 channel version feels like the 2 channel with a little extra variations.
I don't have a 4 channel setup and never tried any other files with 4 channels.

The 4 channel sound can be used with 2 channel devices or with 4 channel devices, but requires a surround sound card in any case in order to avoid down mixing, see the initial post. Yesterday I also uploaded the redundant 3-phase-2-channel files. That can at least help to identify problems with down mixing. The new sound package (see my post from yesterday) also contains an alternative 3-phase sound.

My aim is to build a library that contain sound that work for everyone. Thats why I try to find the reason for the issues you describe.

• It could be a simple bug, e.g. a girl that accidentally uses wrong files. It would be helpful if you know the girl which misbehaved. (enable debug mode in order to see picture names)
• It could be an issue with the estim device. Due to the special carrier waveform my sounds are less compatible than other sound files. A device the passes the frequency range from 100 Hz to 5000 Hz without modifications is required. In the easiest and best case this is a audio amplifier with transformer and optionally some passive components for maximum safety. (TVS diode, resistor, capacitor).
• It could be an issue with the sound file (girl or picture name would be helpful). But I cant image it because there are only very soft effects and all transitions from from low to high are slow. In the first there levels are no sounds that contain spikes or so.

[mantrid]

For a general description, please read the initial post. This message contains the
Release notes for version 1.0

Changes and notes - Library

• Phase shift/intensity issues fixed
• Several median and minor adjustments
• Electrode placing hints in initial post
• Additional stereo file pack (redundant because included in the 4 channel files, see begin of this post)

Changes and notes - Tease

• Tease has a name now: EStimSurprise
• Reload without repeating girls (requires GuideMe 0.4.0 or later, see initial post)
• Many sessions are probably too long. I won't fix it the near future because I'll focus on the software development. If someone wants to continue the tease, please contact me per PM.

[cfs6t08p]

I had the same experience as lolol2 playing through the tease, looking at the audio files (2 channel version) it looks like it is using a low duty cycle square wave at 175Hz? I'd be interested to know what kind of setup you have, this is the kind of signal I would expect from a low-end powerbox (simple circuitry & pulse transformer), on a pure stereostim device it doesn't feel all that great

I really like the idea of being able to generate stim files programmatically though, I'm guessing it's just a matter of configuration. If I can find the time I might have a look at MPGenerator to try to come up with something that fits my setup better

[mantrid]

I had the same experience as lolol2 playing through the tease, looking at the audio files (2 channel version) it looks like it is using a low duty cycle square wave at 175Hz? I'd be interested to know what kind of setup you have, this is the kind of signal I would expect from a low-end powerbox (simple circuitry & pulse transformer), on a pure stereostim device it doesn't feel all that great

In the initial post, section "Hints for electrode placement", there is a recommendation for the standard 3-phase setup. This is what I use. For all other setups you have to try out what works best using the calibration pages. My estim box consists in a USB surround sound card, two stereo audio amplifiers, 1:10 audio transformers for impedance conversion and decoupling and a few passive components for additional safety.

It's correct, a low duty cycle carrier waveform is used. Theoretically (see below) and practically it feels the same as continuous sinusoidal waveforms. You can verify it by comparing the modded Estim Tower sound files in directory 'AAudio.3c' with the original ones. They should feel the same at the same audio level as the original ones. (In practice there is a minimal difference due to compression artifacts in form of small overshots which occur irregularly. If I compare directly the powers-saving waveform feels a little bit noisier.)

If you feel a difference your estim box alters the sound files. If the problem occurs with all files that seems to be likely (but this is not what lolol2 reported). I can also build a file set with continuous carrier waveform but a estim box that does not alter the sound is better.

The theory: Neurons have a maximum fire rate of up to about 100 Hz. If the carrier waveform is is a 1000 Hz signal the neurons fired every 10th pulse or more. If you omit 4 of 5 pulses, you feel the same. (Due to the sampling theorem you cannon omit 9 of 10 pulses). The Shape of the pulses does not matter -- thats far beyond the recognition limit, but the pulse width matters.

I really like the idea of being able to generate stim files programmatically though, I'm guessing it's just a matter of configuration. If I can find the time I might have a look at MPGenerator to try to come up with something that fits my setup better

The script files are not setup dependent. In order to port the lib to another setup it just needs to be rebuild with different command line parameters.

[lolol2]

Just for interest, what is the benefit of the power save wave forms?
Battery saving? Box could get higher output? You get less desensitize over time?

I'm not sure if this is the problem, but as I has written before I also don't like the signals because it's spiky, noisy how ever this is called, they just don't feel good over a longer time for me.

I have a DIY stereostim device which using an audio amp, so there should be no alter of the sound files.
I will try the tower files of you and see if there is the same problem.

[cfs6t08p]

I tried the comparison you suggested, while the sensations are fairly similar, the original audio is silky smooth while the files from 'AAudio.3c' feel prickly and harsh.

While the theory you've presented sounds reasonable, the frequency content, and by extension the pulse shape, definitely does matter. The hard cuts in these files introduce both higher and lower frequency components and I believe it's those lower frequencies that are causing issues here. If I run these files through a 300Hz high-pass filter they improve massively, they still feel "busier" than the original but the uncomfortable sensations are all gone.

A set of files with continuous waveforms would make this pack A+ in my book But if you're really keen on keeping the power down you could use a pulse shaping filter to limit the bandwidth to where it feels more like the original.

[mantrid]

I rebuild the 3-phase extra sound files sound files with continuous carrier waveform Download link. For the standard 3-phase electrode placement sound set 3u (button "unsymmetrical") with channel 1 to the top should work best.

Just for interest, what is the benefit of the power save wave forms?
Battery saving? Box could get higher output? You get less desensitize over time?

1. I would would not induce more energy than necessary.
2. It allows to have different time windows. This makes channel combining more predictable. Without this the intensity of two mixed signals is (a+b)². With windowing it is max(|a|,|b|)². This is relevant if you combine multiple 2- or 3-phase channel sets.

I'm not sure if this is the problem, but as I has written before I also don't like the signals because it's spiky, noisy how ever this is called, they just don't feel good over a longer time for me.

If all signals, feels spiky that could be the reason. If only a few signals are spiky, which files / girls are it. If everything works as intended there should only be soft transitions in the beginning. The girls in the higher levels can be quite rough.

I have a DIY stereostim device which using an audio amp, so there should be no alter of the sound files.
I will try the tower files of you and see if there is the same problem.

What kind of transformers (frequency range) are you using.

I tried the comparison you suggested, while the sensations are fairly similar, the original audio is silky smooth while the files from 'AAudio.3c' feel prickly and harsh.

While the theory you've presented sounds reasonable, the frequency content, and by extension the pulse shape, definitely does matter. The hard cuts in these files introduce both higher and lower frequency components and I believe it's those lower frequencies that are causing issues here. If I run these files through a 300Hz high-pass filter they improve massively, they still feel "busier" than the original but the uncomfortable sensations are all gone.

A set of files with continuous waveforms would make this pack A+ in my book But if you're really keen on keeping the power down you could use a pulse shaping filter to limit the bandwidth to where it feels more like the original.

The lowest frequency is defined by the period of the pulses. This is 175 Hz. The pulse shape influences the higher frequency components (next peak is at 525 Hz for the 2x2-phase signals). Modifying these components, i.e. changing the pulse shape, does not trigger more or less neurons.

Shape only has an effect at lower frequencies: After a pause of at least a few ms the trigger level of the neurons is reduced (inhibition is deactivated) . That's why shocks are shocking.

If you high pass filter this signal you generate something totally new. But I don't think that you can't preserve the phase modulation this way.

What kind of estim box are you using. Maybe the device samples somehow with a similar frequency to 175 Hz.

Again, without the compression artifacts I can't feel a difference between 'AAudio.3c' and the original files.