New estim tracks for CH videos

[digitalparkinglot]

Wow, this is a great thread!!!

I just spent the last couple hours reading every post and doing my best to understand what is being said. A lot is very close to going over my head. I know or understand very little about coding. I can rewire a house, but shaping how electricity feels make my head spin. I am an industrial designer by day, so I am all about how something looks and feels. That probably goes a long way toward describing my video releases.

So, let's call out the elephant in the room. Funscript wasn't designed for estim. It's a bunch of position based commands to tell your stroking device where to move to. It is beyond my comprehension how someone can take that info and turn it into audio that my DIY estim device makes feel like stroking.

My name has been invoked several times in this thread. I think some of you like my estim tracks, so it might be helpful if I share how I create them. I have three main ways of creating estim tracks.

The first and easiest is to take the beat track from a CH video and using LondonGents tutorial viewtopic.php?f=25&t=24040&hilit=funscript+10+minutes turn the beats into a funscript and take that to the cfs6t08p online converter and output an estim file. Those can be unbalanced and boring. I'll usually go into OpenFunScripter (OFS) and clean and tart up the funscript a bit before converting. I read that a lot of this thread is devoted to making the conversion from funscript feel a lot better. I applaud that heartily and will be installing and trying out restim as soon as I'm done writing this.

The second method is motion synchronized with a video in OFS. This is very tedious and results in a funscript. This is really awesome for a stoking machine, but we are again left converting that funcript into an estim track. I'm super curious how well some of these new conversion methods will handle synchronized motion.

The third method, also very tedious, is me sitting in front of my computer, all wired up, creating, borrowing, stealing, reworking, testing, trying out, swapping, re-testing, replacing existing estim signal files until I find the feeling that fits with scene I am pairing it with. This takes hours and hours. I'm wicked numb by the end. In the end I can usually get something that feels really good to me and is completely customized to the video. But, it's not backward compatible and can't be made into a funscript for stroking machines (so far anyway).

I really love this discussion. I wish Milovana would make estim it's own forum topic...but that's beside the point. There is a giant world of funscripts already created. I love that there is discussion on how to get better estim conversions from them and I'm excited to try out some of these new methods.

For me, as a creator, ultimately I want a new tool (a piece of software) that would make creating estim tracks faster, easier, and way more controllable. I want to be able to say I want this kind of sensation on this part of the cock and I want that sensation to travel to this part of the cock and then spread out and begin to feel warm all over. Ideally there would even be multiple sensation at the same time...for example, a tickly sensation in the balls, a licky sensation on the head, and a spiky sensation that travels up and down the shaft. I don't know if any of this is possible, but I'm putting this out there in front of people that I believe might be able to make it real.

-digitalparkinglot

[edger477]

The third method, also very tedious, is me sitting in front of my computer, all wired up, creating, borrowing, stealing, reworking, testing, trying out, swapping, re-testing, replacing existing estim signal files until I find the feeling that fits with scene I am pairing it with. This takes hours and hours. I'm wicked numb by the end. In the end I can usually get something that feels really good to me and is completely customized to the video. But, it's not backward compatible and can't be made into a funscript for stroking machines (so far anyway).

And thank you for this.

I am software engineer and I understand what diglet is doing, but this part of your work is what stands out and reason why I am not fully sold into giving up mp3 files for funscript. Having just tested your latest work, and also tested several funscripts with restim same evening, I have to say that what you did is not yet possible to experience with any of the funscript->estim converters. Maybe, it is possible for one scene, or one break, but to generate whole file that stacks like that from scene to scene, we don't have format to describe it.

In my opinion, your works are unique and I have never experienced anything similar from raw funscript conversion. But having some experience in Audacity I know how tedious that work can be. I think a tool could be developed to describe the signals that would make it easier to change them because you would not have to generate all from scratch every time, but have some kind of parametrized generator.

I fully agree with you that funscript is not meant for this, and I will try to analyze that file now that I have describet sections of with words, to see how many additional axes are needed to describe those variations.

For example, you might have a file format that describes one or more "baselines" i.e. carrier frequency (the one that we enter in funscript converters), and i.e. volume ramp per minute, and then sections of file with start/end timestamps in which you choose which baseline is used, then there could be things we could described similar to how funscript describes strokes (timestamp/position), like frequency modulation of carrier wave, tremolo (amplitude changes) in each channel (+2 axes), and maybe some other things depending on what effects you use in your workflow. And of course phase offset for triphase configuration (the thing that we use funscript now for, to change where the most of sensations is). I think developing an editor with gui to edit such file would be most of the work as generating sound from such format is not hard.

[diglet]

I implemented a few new features.


Configurable network ports. Settings are stored in %appdata%/restim/


More options for choosing audio interface. I found that WDM-KS with 'low' latency has the lowest latency on my machine, but it doesn't always work when there are other programs utilizing the same audio interface. If you experience pops, choose a higher latency or different device.


Latency adjustment for the UI (only affects the UI, not the generated audio).


TCode mapping for volume, carrier frequency, modulation. The range can be adjusted. Not sure if these parameters are a good choice, will probably change in the future. The modulation frequency uses the modulation strength set in the GUI.

If you configure MFP like this:


Then you can script with all of these axes in OFS and test the result on your estim box immediately.
https://github.com/diglet48/restim/releases


Re funscript format, I think the problem isn't the file format but the lack of understanding of how the signals translate to sensations. And lack of tools. I think it would really help if scripters try these tools and give some feedback. I'm particularly looking for feedback on (threephase) sensations the tools cannot yet generate.

[diglet]

I was thinking, gee, I really need to create 5-axes funscripts for a bunch of video's so we can test these new features, that sounds boring. But it turns out we don't need to because of a few features in MFP.

I setup L2 (volume) like so:
Pattern, sine wave with range 99-100.
Smart limit dependent on L3, linear ramp from 100% to 82% (target value: 0%), so the volume decreased automatically when the frequency increases.

And L3 (frequency) like so:
Pattern, triangle with range 0-100. Speed maybe 20%

And L4 (modulation) like so:
Pattern, triangle with range 35-100. Speed maybe 30%

..now play your favorite cock hero on L0 and L1.

Personally I don't notice very much difference as the carrier frequency changes. It feels maybe a bit more focused at higher frequencies, and more diffuse at lower frequencies? It's hard to tell whether this is due to variations in carrier frequency or volume. Modulation feels a bit like a slow vibrator at the low end, smoother at the high end.

[satn00b]

The third method, also very tedious, is me sitting in front of my computer, all wired up, creating, borrowing, stealing, reworking, testing, trying out, swapping, re-testing, replacing existing estim signal files until I find the feeling that fits with scene I am pairing it with. This takes hours and hours. I'm wicked numb by the end. In the end I can usually get something that feels really good to me and is completely customized to the video. But, it's not backward compatible and can't be made into a funscript for stroking machines (so far anyway).

that's what makes them so great!


about the tedious part: the whole task sounds to me as if it would be perfect for an a.i.

however, i'm very aware that a.i. needs huge chunks of data to be trained on and this would require A LOT of video, estim material and user ratings for each moment of these videos to make an a.i. understand what feels good when used with certain video sequences. Heck, I think given enough data there would be a lot of discoveries of patterns we don't even know about now.

However, as I said, I think this is not feasible now because the data collection is not possible in such detail and also in such amount. That would have to be done automatically, by collecting the users response on the stime file at each moment automatically. there are projects that try to do that, but i don't know how far they've come. I found something about orgasm detection and continuously edging (https://github.com/nogasm/nogasm) but dont't know if that works. If it does, that would be a great source of data. Also people are collecting muscle pressure data (https://blog.lioness.io/i-dped-myself-t ... c22e1b4d7a) and that could be a great source, too. If someone's interested in more such stuff: https://discord.gg/uRzbRdZG. Sadly, with 3 children I don't have the time to develop anything in that regard. But since I'm not the only computer scientist around and most of them will notice that a.i. would be perfect for that job, I have hope we'll see something in the future

[diglet]

I was thinking about adding current sensors to the stim box to measure skin resistance. With a bit of luck the skin resistance is correlated with muscle contractions or hardness. Can probably use the same method to detect heart beats. No additional wires to the body needed. This data can be used to modify the estim signal, for example stroking synchronized to your own heart beat.

I tried automatic video-to-funscript conversion for stroker devices with computer vision algorithms last year, but it's very hard and required tons of manual labor to cleanup. I think this might be a bit easier with estim as there are no constraints on the maximum movement speed. With estim (generated by restim) it doesn't really matter how fast of slow you move through the parameter space. With stroker devices, only actions of a certain length and speed feel good.

[Niceguy_LondonUK]

Hey All,

The knowledge in this thread is truly outstanding. I thought I'd feedback on my experience with Restim so far.

I managed to get it to connect to MFP and sync with funscript files and video... really amazing feeling very realistic stroking actions and lots of buzzy vibe-like sensations too.

I then manage to get it to connect with VAM using Yoooi's plugins... I tried his ToySerialContoller.. and that was fun... (If anyone hasn't tried VAM its getting to a point where some really good content creators are building close to photo-quality models and scenes). I also managed to connect it to the BusDriver plugin where the VAM model performed perfectly in sync with the real-life model in the video loaded into MFP and MPV. I found an interesting plugin that lets you possess a VAM model and go into a multi-user VR environment with other er... players.. for interactive you know what... I can't wait to connect my Estim to that... Honestly with innovation in converging technologies, like AI, VR tech like Quest Pro/3 - Unreal engine ver.5 and Stable Diffusion... this could be the future of porn.

Finally I managed to connect Restim to this which was also fun... A really nice GUI and simple to use... https://www.ayva-stroker-lite.io/

One thing I wanted to bring back and possibly question... in the old days of estim.. from the smartsim website back then... a sort of universal 'trode configuration was beginning to become established... Right on the head, Common in the middle and left lower.. i.e. testicles... this in my brain was really simple to understand, because well it was sort of linear... when i played a stim file.. if i wanted more feeling.. on the head... i would use balance control to nudge the sensation further to the right and the head.. if i wanted more on my testicles i would nudge it further to the left... as i used a slider type balance control... i could literally feel the sensation moving in realtime along the shaft... turn that slider 90º and you have a graphical representation of a dick.. really easy to understand...

Why did we move away from that configuration to 'common on the head'... is this due to the way funscript files are constructed and then used to create the audio? Its so much harder for me to understand how i can affect the 'feeling' of the file and move it to the parts I want.

Niceguy.

[digitalparkinglot]

Finally I managed to connect Restim to this which was also fun... A really nice GUI and simple to use... https://www.ayva-stroker-lite.io/



Niceguy.

Hey Niceguy_LondonUK,

Thanks for pointing out Ayva Stroker Lite. It looks super cool. I tried connecting restim with it and can't get it to connect. Will you share how you did it?

-digitalparkinglot

[edger477]

Thanks for pointing out Ayva Stroker Lite. It looks super cool. I tried connecting restim with it and can't get it to connect. Will you share how you did it?

it seems there is option to choose output settings when you click on gear symbol next to the usb symbol in top left. Then you can choose web socket and enter address and port of the other side.

[diglet]

One thing I wanted to bring back and possibly question...

The problem with threephase is that there are three independent degrees of freedom. Volume, alpha and beta. Controlling the volume is easy: use the master volume knob on your stimbox. Controlling beta (left/right) is also easy: you can do that with the left/right volume knobs. But now there is a problem: there is no way to control the volume at neutral relative to L/R (alpha axis), no matter how you twist the volume knobs, it's impossible.

If you put the common at head, due to various hardware problems the intensity at the head is much higher than in the shaft/balls. There are two ways to solve this, you can move the neutral electrode to the least sensitive spot (the shaft), now you can control the balance between the head and the balls with either the left or the right volume knob, but now you can't control the balance between the shaft and the rest. Or you can modify the audio to reduce the volume output at the neutral electrode. The second approach is the one that restim takes. Reducing the Neutral power calibration parameter is like twisting a new volume knob at your stimbox that only reduces the output power to the neutral electrode. The wiki documents the underlying math. I personally don't use the L/R balance knobs on my hardware since the adjustments in the software do the same.

Since the volume is corrected in software, the electrode orientation becomes irrelevant. No matter how you hook up the electrodes, you can modify the calibration parameters to get the sensation exactly the same. So you should hook up the electrodes in a way which you prefer. The scripting assumes neutral is the top position of the stroke, there will be a way to modify this in the future.

I think Ayva Stroker Lite works better if you configure restim to use L1 and L2 for the alpha/beta axis.

[Niceguy_LondonUK]

Finally I managed to connect Restim to this which was also fun... A really nice GUI and simple to use... https://www.ayva-stroker-lite.io/



Niceguy.

Hey Niceguy_LondonUK,

Thanks for pointing out Ayva Stroker Lite. It looks super cool. I tried connecting restim with it and can't get it to connect. Will you share how you did it?

-digitalparkinglot

Yeah Sure... so you click on the cog icon next to the usb symbol, in the drop down select websocket and put in the values for restim as supplied by Diglet. Host = localhost, Port = 12346. You'll see the green dot responding to the virtual stroker, you can just click on Freeplay for some randomised play, or if you look on the right there are about 25 x different patterns to select whichever you like.

[Niceguy_LondonUK]

One thing I wanted to bring back and possibly question...

The problem with threephase is that there are three independent degrees of freedom. Volume, alpha and beta. Controlling the volume is easy: use the master volume knob on your stimbox. Controlling beta (left/right) is also easy: you can do that with the left/right volume knobs. But now there is a problem: there is no way to control the volume at neutral relative to L/R (alpha axis), no matter how you twist the volume knobs, it's impossible.

If you put the common at head, due to various hardware problems the intensity at the head is much higher than in the shaft/balls. There are two ways to solve this, you can move the neutral electrode to the least sensitive spot (the shaft), now you can control the balance between the head and the balls with either the left or the right volume knob, but now you can't control the balance between the shaft and the rest. Or you can modify the audio to reduce the volume output at the neutral electrode. The second approach is the one that restim takes. Reducing the Neutral power calibration parameter is like twisting a new volume knob at your stimbox that only reduces the output power to the neutral electrode. The wiki documents the underlying math. I personally don't use the L/R balance knobs on my hardware since the adjustments in the software do the same.

Since the volume is corrected in software, the electrode orientation becomes irrelevant. No matter how you hook up the electrodes, you can modify the calibration parameters to get the sensation exactly the same. So you should hook up the electrodes in a way which you prefer. The scripting assumes neutral is the top position of the stroke, there will be a way to modify this in the future.

I think Ayva Stroker Lite works better if you configure restim to use L1 and L2 for the alpha/beta axis.

Thank you for such a great explanation in layman's terms.. maybe i just need to reframe the labels in my head (er.. brain).. stop thinking about right = head, left = balls... and think electrode 1, 2 and 3. Once again thank for updating restim to v.1.3

Niceguy.

[digitalparkinglot]

Thanks Niceguy,

I had been doing that, that's why I asked. It still wouldn't connect. Then I thought, I'm using Brave web browser, I wonder if that is the problem. I tried it with Chrome and it connected with no issues. I also couldn't get my OSR2 to connect with Brave but connected easily with Chrome. I love Brave, it's really fast, but this is not the first time I've run into issues.

Thanks for your help and your quick response. Man, I can't wait to try this out.

-dpl

[edger477]

I love Brave, it's really fast, but this is not the first time I've run into issues.

Tried with Brave Shield down? When is turned off, should practically behave identical to Chrome. And what is great is when you turn it off, it is only turned off for that site.

[Niceguy_LondonUK]

Hey Guys,

A quick update: Im still enjoying the new lease of estim life Restim has given me, and combined with VAM it a whole new world. I pressed the button and ordered my self a Meta Quest.. to see what that opens up... I have seen a full VR room scale plugin on VAM that lets you import a model into your own living room..(uses passthrough) for you to er... do what you want with her... that'll be fun.

@ Diglet, I wonder if you could add a serial connection to Restim (you said you had the code 'lying around' on the other forum)... I'd love to continue my exploration... adding serial means we'll be able to use Tempest's code for his robots and convert them to estim...

Screen Shot 2023-04-05 at 13.40.22.png (402.56 KiB) Viewed 1659 times

It'll also hopefully mean we can connect to FapTap (https://faptap.net/) and enjoy streaming videos and funscripts for free without the need for downloads.... I tried to connect Restim to Faptap using the websocket but it won't connect, I'm assuming its because its serial... Let me know if you have the time...

IMG_7104.JPG (113.6 KiB) Viewed 1659 times

Screen Shot 2023-04-05 at 13.40.43.png (105.49 KiB) Viewed 1659 times

Screen Shot 2023-04-05 at 13.40.55.png (100.6 KiB) Viewed 1659 times