New estim tracks for CH videos

[diglet]

I created an actual release that you can just download and run. https://github.com/diglet48/restim/releases

For basic help, see the wiki: https://github.com/diglet48/restim/wiki

Recent changes:

• Single executable
• Select audio device, audio start/stop
• Integrated Funscript converter, in tools menu

The intended use is alongside MultiFunPlayer for video synchronization. I had good success with Cock Hero - Contract, Sex Cauldron Part 2 - The Embiggening, Ultimate Edging Challenge - Massage Edition.

[Kelvinator]

I downloaded your restim release, ran it, and converted a funscript file from 1d to 2d, all without a hitch. I've also installed MultiFunPlayer and am playing around with that for video synchronization. Looking forward to getting it all to work together. Thanks for this nice contribution.

[diglet]

I decided to spend a bit more time reverse engineering exiting stimfiles, because data science is addicting, yo. I'll be looking at ch_eroclip stimfile by BoundSquirrel viewtopic.php?t=22889

The orientation of the diagrams below equals the orientation in the software (neutral = up)

Each circle in this image shows 5 seconds of movement. This gives a nice overview of the script. What's nice to see is that the volume is higher towards the bottom of the circle. I believe this stimfile was generated by copy-pasting small good-feeling segments from other stimfiles. The amount of variety is impressive, especially considering the method of construction.



I found this sequence interesting, the sensation starts low and gradually moves to the head (assuming neutral is head). The graphs show a clear relation between power and alpha, indicating power decreases as the sensation becomes more focused at the head. The volume ramp / tremolo is not as strong as I would expect.





Here is a segment which focuses the sensation towards the bottom (or left and right electrode). The volume is clearly higher than in the previous segments.





Some segments turn into a bit of a mess... audacity gives a closer view of what's going on: this particular segment of the file consists of two separate tracks that were probably never intended to be combined. My algorithm can't make much sense of garbled mess.





Here is one more segment from the end:




This analysis was performed with 500 non-overlapping windows per second, output is not smoothed. Analysis is very fast (1s per 1 hour of audio). It is very feasible to convert these patterns into funscript.

tl;dr: converting audio to funscript and then back into audio with restim works well in most cases.

[senorgif2]

I decided to spend a bit more time reverse engineering exiting stimfiles, because data science is addicting, yo. I'll be looking at ch_eroclip stimfile by BoundSquirrel viewtopic.php?t=22889

The orientation of the diagrams below equals the orientation in the software (neutral = up)

Each circle in this image shows 5 seconds of movement. This gives a nice overview of the script. What's nice to see is that the volume is higher towards the bottom of the circle. I believe this stimfile was generated by copy-pasting small good-feeling segments from other stimfiles. The amount of variety is impressive, especially considering the method of construction.



I found this sequence interesting, the sensation starts low and gradually moves to the head (assuming neutral is head). The graphs show a clear relation between power and alpha, indicating power decreases as the sensation becomes more focused at the head. The volume ramp / tremolo is not as strong as I would expect.





Here is a segment which focuses the sensation towards the bottom (or left and right electrode). The volume is clearly higher than in the previous segments.





Some segments turn into a bit of a mess... audacity gives a closer view of what's going on: this particular segment of the file consists of two separate tracks that were probably never intended to be combined. My algorithm can't make much sense of garbled mess.





Here is one more segment from the end:




This analysis was performed with 500 non-overlapping windows per second, output is not smoothed. Analysis is very fast (1s per 1 hour of audio). It is very feasible to convert these patterns into funscript.

tl;dr: converting audio to funscript and then back into audio with restim works well in most cases.

Interesting stuff, I Honestly don't even understand it, but that file you are looking at is meant to be played as stereostim dual channel, not in triphase. It seems to me that you are analyzing it as if its triphase no?

[diglet]

I'm analyzing it as three phase, users report the file working fine in that configuration. I tried to analyze many different files and the stimfiles from this author were by far the most complicated looking.

This wiki page might help understanding the diagram. The labels are intended for three-phase, I'm sure the same math works for dual channel, but I still need to work out the right interpretation of the data. In think in a dual channel setup, left/right controls whether the current flows through the left or the right channel, and up/down controls the amount of current flowing between the channels.

[senorgif2]

I'm analyzing it as three phase, users report the file working fine in that configuration. I tried to analyze many different files and the stimfiles from this author were by far the most complicated looking.

This wiki page might help understanding the diagram. The labels are intended for three-phase, I'm sure the same math works for dual channel, but I still need to work out the right interpretation of the data. In think in a dual channel setup, left/right controls whether the current flows through the left or the right channel, and up/down controls the amount of current flowing between the channels.

I always played this one as a dual Channel, I believe that was the creators intention with the file. AFAIK in my limited understanding of electricity in a dual channel config, the channels are isolated from each-other. So the left track is one set of electrodes, and the right track is another set of electrodes. In practice this seems true because I can turn off one channel, and still have the same exact sensation in the active channel. in other words there should not be any current flowing between the channels in dual channel. Again, that my understanding I'm clearly not as learned as you are in the subject.

And shortly after I posted my first reply i downloaded Restim and I understand the diagrams a lot better

[BoundSquirrel]

I decided to spend a bit more time reverse engineering exiting stimfiles, because data science is addicting, yo. I'll be looking at ch_eroclip stimfile by BoundSquirrel viewtopic.php?t=22889

The orientation of the diagrams below equals the orientation in the software (neutral = up)

Each circle in this image shows 5 seconds of movement. This gives a nice overview of the script. What's nice to see is that the volume is higher towards the bottom of the circle. I believe this stimfile was generated by copy-pasting small good-feeling segments from other stimfiles. The amount of variety is impressive, especially considering the method of construction.



I found this sequence interesting, the sensation starts low and gradually moves to the head (assuming neutral is head). The graphs show a clear relation between power and alpha, indicating power decreases as the sensation becomes more focused at the head. The volume ramp / tremolo is not as strong as I would expect.





Here is a segment which focuses the sensation towards the bottom (or left and right electrode). The volume is clearly higher than in the previous segments.





Some segments turn into a bit of a mess... audacity gives a closer view of what's going on: this particular segment of the file consists of two separate tracks that were probably never intended to be combined. My algorithm can't make much sense of garbled mess.





Here is one more segment from the end:




This analysis was performed with 500 non-overlapping windows per second, output is not smoothed. Analysis is very fast (1s per 1 hour of audio). It is very feasible to convert these patterns into funscript.

tl;dr: converting audio to funscript and then back into audio with restim works well in most cases.

I just caught up on this thread.
1: holy crap, I had no idea my stim files were still being used for much of anything anymore. I'm really flattered that you're using my creations.
2: you're partially right - the right (bottom) channel in the clip shown was copy/pasted from one of digitalparkinglot's creations with his permission. I can't remember which one now, but they're green, yellow, or blue sections of one of his RL/GL series. The top beats were my own creations - I had a custom Nyquist formula that I used to generate these. It's been a long time since I've done this so I can't remember what the syntax was. Somewhere I've got a template file with this kind of beat from 0.1s all the way to 1.0s in 0.01s intervals. I used that template library for all of my creations. If that would help your work, I'll see if I have that file and post it if there's interest.
3: I have no idea what you're doing with the mapping and the analysis, but it looks really impressive and I'm happy my stuff is able to help!

[diglet]

Thanks for the kind words . To summarize, I use some math to find what the freedom space of three-phase audio is, and then I generate audio by moving around in that freedom space. This makes it much easier to generate audio that utilize the full range of three-phase capabilities. In theory.


I tried some action VR and higher carrier frequencies today. The results were interesting. I found that funscript files that aren't as enjoyable with a stroker style device work quite well with restim. With stroker style devices it's quite hard to find funscripts with the right tempo and stroke depth so suit your tastes, e-stim seems to be less sensitive to this.

VR seems to work a lot better when you're not awkwardly trying to position yourself under the stroker.

As for the higher frequencies, I tried 1000-1100hz, I usually run at 700hz. I heard elsewhere that higher frequencies don't penetrate the skin as well, therefore in theory the 3D effect is less. I wasn't able to reproduce this in testing, If anything, the 3D effect was improved and pain effect when directly stimming nerves decreased (needs more testing).

The 1100hz signal also felt a bit smoother, at 700hz and ~50hz modulation frequency, I feel a slight buzz that doesn't go away once the modulation frequency is increased. But it does go away when the carrier frequency is increased. Maybe it has something to do with the way I modulate the signal... I'll put it on the list of research topics

I had to modify the calibration parameters and volume a bit when changing the frequency. It looks like the nerves in the glans area are a bit more sensitive to high frequency signals. That's too bad because it will make it almost impossible to control the frequency via funscripts...



I also tried to analyze a few more stimfiles, hopefully finding some interesting patterns, but I'm not having any success. I don't see any files with what I believe to be properly calibrated volume or "clean" movement in my freedom space. Not sure if I want to try them or if I should build my own patterns with funscripts from scratch...

[Kelvinator]

After much trial and error, poking around on the web, and dealing with laptop administrative issues, I FINALLY got restim, multifunplayer, and MPV to connect and play well with each other (sort of). In MFP, I was able to get the following to run simultaneously:
1. restim running, calibrated, and audio outputted to laptop speakers.
2, MPV running and video file loaded and controlled by MFP.
3. MFP websocket connected to restim (localhost:12346),
4. MFP: Alpha and Beta funscript files (created via restim) loaded into MFP and playing ,
5. Audio output from the two script files, plus the video playing on my laptop speakers (which connect to my DIY MidiStim box). MPV can independently mute the audio coming from the video file, without muting the audio from restim.
I plan to test this setup tonight.
However, I think I'm missing something. As per diglet's MFP setup section on his restim wiki, I tried (unsuccessfully) to addI need restim as a device in MFP version 1.24. When I click on Settings icon in the top menu bar, and select Device, the only two options I get are Tcode-03 and Tcode-02. The edit icon is grayed out. I would like to add restim, but can't figure out how to do it. I'm stumped. Help, anyone?

[diglet]

However, I think I'm missing something. As per diglet's MFP setup section on his restim wiki, I tried (unsuccessfully) to addI need restim as a device in MFP version 1.24. When I click on Settings icon in the top menu bar, and select Device, the only two options I get are Tcode-03 and Tcode-02. The edit icon is grayed out. I would like to add restim, but can't figure out how to do it. I'm stumped. Help, anyone?

You need to clone one of the tcode devices using the button immediately right of the device dropdown, rename it, and then add/remove axis until the device configuration is as in the image on the wiki. https://github.com/diglet48/restim/wiki/mfp-setup

[Kelvinator]

You need to clone one of the tcode devices using the button immediately right of the device dropdown, rename it, and then add/remove axis until the device configuration is as in the image on the wiki. https://github.com/diglet48/restim/wiki/mfp-setup

That worked, Diglet. Thanks for the help and quick response. I'm going to give this restim / MFP / MPV integrated system a test run tomorrow. I'll let you know how it goes.

[diglet]

I wrote some code to test different modes of amplitude modulation, allowing for specific ramp/duration of the modulation signal. In my testing, this turned out to be not very useful, signal always feels the same or slightly worse than sine modulation. Sine modulation is, at least according to my testing, just the best.

Also played around with motion controls. Was able to get the audio latency down from ~300ms to ~50ms, not stable enough for production but good enough for some testing. Wrote some motion detection code for the joycon, strapped to my leg, to detect hip movements. It was a struggle to detect the motion accurately (This is a really hard problem), but it actually works quite well. Not sure if this is viable long-term, but it's fun to play around with. Also, attaching the motion sensor to different body parts (i.e. feet) is weird.

I have a ring-con (ring fit adventure) lying around, this is basically a large force sensor in the shape of a ring, compress the ring to stroke. If your arms get tired you can place it between your knees. It is practical? No. It it fun? Yes.

If anyone wants to play around with this, I can release the required scripts.

[Kelvinator]

I've been a busy beaver lately. I tried and liked zebbg69's soft triphase setup. I saw that Joanne has come up with the "Stairway to Heaven" holder for conductive rubber CR loops that is perfect for this electrode setup with balls and shaft as common. Check it out on HappyStim-USA.com or Joanne's store https://www.sexmachinereviews.co.uk/ I've also been playing around with MultifunPlayer and diglet's restim program for the past couple of weeks. Tonight I had a really great session with the following setup:
1. Electrodes:
a. Common: Balls and Shaft a la zebbg69 using my Stairway to Heaven.
b. Left - Head CR loop.
c. Right - Prostate PES Prostate Stimulator.
2. Videos:
a. NumberOneFan's Amyl Zombie. Using alpha and beta funscript files converted from the original funscript files using diglet's restim > tool > funscript converter.
Did you know there are about 30 popper hit prompts? Way to much for my tender body. I find that one hit every 8-10 minutes gives me enough time to recover and really enjoy the rush of the next one. Anyway, that's me.
b. DigitalParkingLot's Breathe Harder, again using diglet's funstim converter.
3. Restim settings: Following diglet's instructions, I set the initial calibration settings as follows:
a. Carrier frequency: 700 Hz.
b. Modulation 1 frequency: 70 Hz, at 99.99 %
c. Modulation 1 frequency: 5 Hz, at 10.00 %
I selected the circle calibration pattern, and adjusted the calibration parameters until the signals were close to equal strength at all "3" electrodes:
d. Neutral: Since the balls - shaft common is a fairly large surface area, I was happy with something close or a little below 0.
e. Right: Around 2, as my Head is sensitive.
f. Center: 0.
4. MultiFunPlayer: I used MVP video player to play the videos. MFP synchronized the funscript files and ported them to restim.
5. Result: Amazing !! Suffice it to say, that I felt pleasant / intense / subtle, etc. sensations rolling up and down my head, shaft and balls that I've never felt or noticed before. It was extremely satisfying. I've always said, that to make something better, you have to have good stuff to begin with, e.g. #1Fan and DPL's funscripts. Diglet's restim program does a fantastic job of maximizing the funscript to estim conversion. Another thing I like, is being able to adjust the calibration parameters / settings on the fly while the video and estim signal are playing. It adds a whole new level of interactivity to the cock hero estim experience. I hope others will check out restim and comment on their experiences. Happy stimming. K

[diglet]

I added a volume control feature. It is very simple, volume increases linearly until the target is reached.

It's also possible to adjust the volume with the keyboard (arrow keys, pgup/pgdn for larger steps). This is more useful than I expected, making small adjustments with the keyboard is much easier than with the hardware potentiometers.



Be careful, don't put yourselves in a situation where fat-fingering results in dangerously high volume.

This feature will be included in the next release, or get it now by running from source.

[Niceguy_LondonUK]

I’m reasonably new to Milovana…

But not to estim…

I only really post when I have something useful to say or to add… or when something game-changing comes along…

This is one of those times…

I’ve just stumbled upon this thread. I haven’t had a chance to download restim yet… (great name btw)

Am I correct in thinking this converts funscripts to audio for estim on the fly and in real-time?

It also integrates with MFP to connect to a video player?

I’ve also just discovered VAM and the possibilities of interactive experiences using that system…

Yooi (the creator of MFP) has a couple of plugins that sync the action in a scene to funscripts… either by using the action in a scene to output a funscript to drive and control a stroker toy or robot, alternatively you can use a funscript in MFP to drive the models action in a scene… is there a way to bring restim into the mix instead of a video player so that we (estim fans) can experience the delights of interactive stimming…?

This would be a major leap forward, for estim.

Niceguy.