Adding sound or a voice to your doll

Technician · Post by **Technician** » Mon Jan 27, 2020 2:28 am

iconiQ wrote:@technician

I've been following your project for some time and am intrigued as to the progress that you have made.

I've also been playing around with multiple "Android" apps (for ease of use) and out of them all, I too, have settled on Bot Libre Offline. It does have some potential as well as limitations. The big one is being able to modify/edit/add/delete phrases/keywords without having to redo training.

Were you able to overcome this obstacle? Or are you phonetically testing your phrases, training the bot with it, and if it's not ideal cancelling the training session and starting again?

I'd much rather have a scripted file (like AIML or such) whereas I can modify as I see fit and fork responses but it seems it doesn't exist at this time.

Either way, great job on how far you've come. I'll continue to watch your progress!

Thank you.
I use an app called "say it" to test dial in phonetics, then I have a text file with multiple acceptable responses (phonetics corrected) to given questions. As long as each response is punctuated you can enter many at once during training (cut and paste) and it'll spit them out one at a time.
I think rooting the phone will provide access to a file that can just be edited, and more important swapped for different personalities.

I'm currently working on BotLibre or UltraHal for pc, not able to get the bt microphone working so far so I may be rooting that old phone soon.

timetraveler1 · Post by **timetraveler1** » Tue Jan 28, 2020 11:59 am

when i mention about blue tooth , the reason i mention it is that's what speakers use (wireless) and as far as head phones you can use your pc with the blue tooth . this is what i have have done in the past and even used a dragon naturally speak voice recognition where i talked with my doll and using a ultrahal a.i. program the doll talked back to me .
i talked through the computers built in mic and the doll answered from a nearby speaker (in a open purse beside the doll and or speaker behind the doll. .)

iconiQ · Post by **iconiQ** » Tue Jan 28, 2020 1:58 pm

Technician wrote:
Thank you.
I use an app called "say it" to test dial in phonetics, then I have a text file with multiple acceptable responses (phonetics corrected) to given questions. As long as each response is punctuated you can enter many at once during training (cut and paste) and it'll spit them out one at a time.
I think rooting the phone will provide access to a file that can just be edited, and more important swapped for different personalities.

I'm currently working on BotLibre or UltraHal for pc, not able to get the bt microphone working so far so I may be rooting that old phone soon.

I've rooted my devices and as for a file that can be edited, well, it's not that easy as it's not in plain-text.

I've gone one up and messaged the developer of the app (Bot Libre) to see whether they can implement a chat editing function or a way to import "Bots" from their website so that we can do editing there.
I did get a pretty fast reply back and a promising insight that they are in the works of creating a new app that has an import/export function. As for editing, he really didn't say as it, in my opinion, would make it easier to edit scattered conversations.
Either way I was quite satisfied that a response was attained and hoping a release is sooner than later.

Until then, we wait and hope.

Cheers!

Technician · Post by **Technician** » Tue Jan 28, 2020 2:02 pm

iconiQ wrote:
Technician wrote:
Thank you.
I use an app called "say it" to test dial in phonetics, then I have a text file with multiple acceptable responses (phonetics corrected) to given questions. As long as each response is punctuated you can enter many at once during training (cut and paste) and it'll spit them out one at a time.
I think rooting the phone will provide access to a file that can just be edited, and more important swapped for different personalities.

I'm currently working on BotLibre or UltraHal for pc, not able to get the bt microphone working so far so I may be rooting that old phone soon.
I've rooted my devices and as for a file that can be edited, well, it's not that easy as it's not in plain-text.

I've gone one up and messaged the developer of the app (Bot Libre) to see whether they can implement a chat editing function or a way to import "Bots" from their website so that we can do editing there.
I did get a pretty fast reply back and a promising insight that they are in the works of creating a new app that has an import/export function. As for editing, he really didn't say as it, in my opinion, would make it easier to edit scattered conversations.
Either way I was quite satisfied that a response was attained and hoping a release is sooner than later.

Until then, we wait and hope.

Cheers!

Can you tell me what the file name is? I'd like to see if I can copy it onto a PC and get something in Linux to open it.

iconiQ · Post by **iconiQ** » Tue Jan 28, 2020 2:51 pm

Technician wrote:
Can you tell me what the file name is? I'd like to see if I can copy it onto a PC and get something in Linux to open it.

From what I can tell it's based on the initial "bot" you select (ie. Julie, empty, etc). I'll need to play around more with it just to be sure.

But, the file I tried to explore is found here:

root/data/data/com.paphus.botlibre.offline/files/Julie

I'll play around some more when I get some free time (limited as it is).

If you haven't rooted (or can't) just let me know. I can probably destroy my current implementation in favor of a raw "virgin" bot and upload that file here.

Technician · Post by **Technician** » Tue Jan 28, 2020 6:38 pm

Thanks! I'll see if I can copy that file with a root file browser tomorrow.

I have a few plans moving forward with Bubbles' ears. Picking up some 3/8" nylon tubing tomorrow and I had to order some components from my electronics supply for work so I ordered 2 pairs of different impedance microphones while I was in there.
I'll try both sets and test for clarity, then decide how to procede with the ear canal reworking.
I also need to do some internal TPE repair, the mouth oriface was only bonded to the mask and it's now separating so I'll use the opportunity to re-shape it into the right angle for the oral cavity.

Technician · Post by **Technician** » Fri Jan 31, 2020 12:47 am

Ordered the voicebox for my other doll, her skull is smaller and this should be a good fit. I'm not wild about the traditional power switch, can't wait to see how I make that work but it got great reviews and has crazy battery life between charges.

https://www.amazon.com/gp/aw/d/B01HTH3C ... prod_image

iconiQ · Post by **iconiQ** » Fri Jan 31, 2020 8:58 am

Technician wrote:Ordered the voicebox for my other doll, her skull is smaller and this should be a good fit. I'm not wild about the traditional power switch, can't wait to see how I make that work but it got great reviews and has crazy battery life between charges.

https://www.amazon.com/gp/aw/d/B01HTH3C ... prod_image

Not sure what your skills are but you could try replacing the SPST switch with a capacitive touch panel and place it on the back of the skull. On/off just by caressing the back of her head. Just like the real thing. LoL

Good luck!

(still watching from the sidelines)

Technician · Post by **Technician** » Fri Jan 31, 2020 9:16 am

My first thought was to use one of the little PIC10F200 microcontrollers I have to make a latching on/off. But then there's the matter of power consumption since it would have to hold a relay. Really the only way I see to do it without parasitic loss is to just put a switch in. This weekend is still just testing microphones and getting Bubbles' ear canals tuned.
Hopefully next weekend I can make a push towards two talking dolls but I suspect Ill have to order microphones for the other doll.
I'm thinking about hearing tests tonight if all goes well but still need to fix the TPE in the oral orifice.

Technician · Post by **Technician** » Wed Feb 05, 2020 10:56 am

Haven't had much time to work on the ear canals, most of last weekend was yanking an engine from a car. I have to practice splicing TPE together first too so I can re-orient the oral orifice while her head is apart.
Also found that the new voicebox doesn't want to use the microphone so unless an app is out there that I somehow missed the first time I'll have to get a different one for the other doll.

Gundam · Post by **Gundam** » Wed Feb 05, 2020 11:30 am

I love following this thread and other AI/Robotics & Doll integration thoughts. I am an InfoSec consultant and so many of my customers know they need integrate AI into their business. I prefer to think of my ladies as Sex Yoga Action Figures and just Works of Art that make me feel happy to see. I also have Alexa, Siri and Google assistant in my house & office with all kinds of voice activated outlets, lights, tvs, thermostats, etc. I am always paranoid about how easily they could be hacked by the manufacturers and more. Still, I accept the risk.

When it comes to my Sex Yoga Assistants, I wonder if it isn't better they see and hear nothing. I love a smart woman but prefer a dumb Tantra Action Figure. I am certain the world will see huge developments in our Doll world and their will surely be wonderful attempts and perhaps successes, creating a love doll robot Jarvis, who you can trust with all your secrets/passwords/SS#/etc. I am just as sure their will be breaches and people will likely look at today as a sort of "Golden Age" of unhackable dolls.

Oh, believe me. I am looking forward to a day, say in the 2045 area, where I will have a trusted AI/Robot assistant who is also built like an OR156H. And when that happens, I believe we will have, to some extent the contributions on this thread to thank. In 25 years, I think the TDF forum will be looked at like the early days of Hollywood.

Grateful to be here. Thank you!

Technician · Post by **Technician** » Sun Mar 01, 2020 3:12 pm

Got the second bt voicebox working and I actually like it better than the Monster now. It's already been disassembled and tested after stripping. Before I figured it out I ordered another that will be here tomorrow so maybe I'll experiment with thisone too and decide which voicebox is going into which doll.

Technician · Post by **Technician** » Wed Mar 04, 2020 11:21 am

Pretty cool stuff today, I decided I'm definitely replacing the Monster with the Soundcore!
A while back I got some microphones and tested them yesterday, one works particularly well with the Soundcore when two are in parallel. It's generally not thought to be good practice to put two microphones in parallel but I wanted to see if I could give her two ears.
These things can hear me perfectly from a distance speaking at a normal to low volume level. They can hear me whisper from a couple feet away! I threw tongue twisters at it and a whole lot of phrases and over 90% of it was interpreted accurately.
Mind you this was also with them taped to a table to muffle the backsides, and simulated ear canals pressed onto them!
I spoke toward them, away from them and 90* off from them at all volumes, she should be able to hear every bit as well as me.
I also took some time to fiddle with the Acapela voice modifiers, I was having her say sentences and speeding/ slowing key words as well as modifying the pitch of certain words. With the right voice tag in the right spot she can really say some sexy stuff! Straight up pillow talk. There's also a way to enter a word in phonetically but I'm not good at it yet, by specifying the phonetics she can do some accentuated syllables and even whisper.

I'm dropping the links to the BT speaker, and the microphones here for reference.

https://www.amazon.com/gp/aw/d/B01HTH3C ... asin_title

https://www.mouser.com/ProductDetail/CU ... pf9w%3D%3D

: 20200304_114840-01.jpeg (1.27 MiB) Viewed 1458 times

iconiQ · Post by **iconiQ** » Wed Mar 04, 2020 12:03 pm

Technician wrote:Pretty cool stuff today, I decided I'm definitely replacing the Monster with the Soundcore!
A while back I got some microphones and tested them yesterday, one works particularly well with the Soundcore when two are in parallel. It's generally not thought to be good practice to put two microphones in parallel but I wanted to see if I could give her two ears.
These things ...

Excellent progress indeed!

Still following your project

Have you completely settled on a single app for voice recognition? There's many available and I've reached out to a few of them in hopes of finding that "all in one" solution. They all have pros and cons but just don't seem to have it all.

Like you, I've been experimenting with Bot Libre but find it's extremely slow when trying to have a conversation. Nobody likes waiting 2-5 seconds when trying to have a decent conversation and then having to wait for that annoying "beep" to let me know when to talk! LoL

I've also been trying many different voices and have yet to settle on any "one" particular voice. The Lisa voice from NaturalTextReader has some very good tones to it but it would seem it's not available anywhere offline for purchase (looks to be part of Amazon Polly project). Curious as to the avenue you've taken

Keep us all updated!

Technician · Post by **Technician** » Wed Mar 04, 2020 2:08 pm

For now I'm still listening to the beeps and pauses from bot Libre. Ideally I would like to go to a computer with this instead of a phone because that would open the door for so much more including being able to change files and have many personalities. About Libra is powered by Yandex, and I have to say it's voice recognition is much better than my phone when I'm sending a text message or leaving a post on the Forum like I am right now.
Thank you for following along! And any others that also do follow my progress....