Thanks Begog for your suggestions!! Yeah, the app is still very dumb, and I do plan to use Dialogflow's AI in the future, but I like to explore the use of real person's voice first. In the released app you can record your voice/directly. The voice is in M4A(AAC) format. I think using real person's voice is a better solution than using AI, but not really sure.
Right now I put a headband on Debbie, and hide the same speaker you have in that headband. It works well for me.
The website Doll33.com is being built, which is dedicated to all the doll owners, and It also explains how conversations works, still under construction, but like 50% is done.
The app itself is still very buggy. I used different kind of words matching algorithm and didn't work well. It's time to keep working it.
Starting a code project to make my doll speak
Re: Starting a code project to make my doll speak
The one and only: Debbie from Doll4E
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
-
LovesBlackWomen
- Doll Patriarch
- Posts: 4490
- Joined: Thu Jul 31, 2008 12:00 am
- Location: Brooklyn, NY
- Contact:
Re: Starting a code project to make my doll speak
I think 'll just stick a tape player in her mouth. It's worked for 35 years.
-
Begog
- Doll Patriarch
- Posts: 4548
- Joined: Mon Jun 30, 2014 8:01 pm
- Location: Dollmongering in an undisclosed location
- Contact:
Re: Starting a code project to make my doll speak
And it's a 35-year old solution that never worked that well in the first place.LovesBlackWomen wrote:I think 'll just stick a tape player in her mouth. It's worked for 35 years.
-
Begog
- Doll Patriarch
- Posts: 4548
- Joined: Mon Jun 30, 2014 8:01 pm
- Location: Dollmongering in an undisclosed location
- Contact:
Re: Starting a code project to make my doll speak
Ok Phillpw99, I tried your app. Here's my initial assessment. Please do not take this as harsh or negative, I am simply trying to help you develop your app. Keep in mind I have a TON of experience programming chatbots, and will be comparing Doll33 to other options, as a base.
First, here's some things I like about the app:
1) It doesn't eat the battery. After an hour or so, My Nextbook tablet still had about a 90% charge, and only 2% was used by Doll33. By comparison, the Harmony app eats the ENTIRE BATTERY in about 45 minutes. WTF? I believe this is mostly due to Google & cloud access, along with the display of a woman talking. What a waste. All we need is a talking doll, right? So Doll33 got this right and this is how it should be.
2) It uses the default android woman's voice. There is NOTHING wrong with that. Harmony has the same voice, unless you download & install some of the replacement voices. Sure, she has a robotic accent - so what? They ALL do. You are talking to a machine, and until machines learn the finer points of language and accent, that's how it's going to be, so again, Doll33 got this right. The alternative is recorded tracks, which do not work in conversation.
Some things that need work:
1) The start button doesn't work. That's kind of important. All it does is trigger the intro. What it should do is trigger the mic so you can start talking to it.
2) "Test triggers" causes a few statements about that feature, and then the mic comes on one time. You can say something, but unless it's in the original database, Doll33 won't recognize the challenge statement. After her first reply, the conversation is over. That brings us back to needing the start button to work.
3) "F-mode" - This brings up a pic of a young-looking doll and a teddy bear. That may be the problem getting your app approved. I suggest using a sex doll that looks "of age," and leave out things that suggest pedophilia. Also, this mode doesn't work, and there is nothing to click on. I suggest a working "Start" button here as well, that lets you talk to the app.
4) When you touch "Conversations" on the drop-down menu, it carries you to the Doll33 website. Nothing wrong with a working link, but that button should take you straight to the conversation editing page. Also, eventually that page is going to become gigantic, and you will need to separate them somehow. I like the way Personality Forge does it: Challenge statements are alphabetically ordered by the first word of the statement, under "A" or "F" or whatever letter the statement begins with. If you try to re-use a challenge statement, it will let you know, and take you to that one to edit it instead. Kari alphabetizes all of them together, and while you CAN find them, when your bot is as developed as mine, it becomes a giant tedious mess. With Harmony, No need to worry about that, because you have absolutely no access to her programmed replies and CANNOT edit them. How sad. This means she will say whatever Guile (the lead programmer) entered for her to say, and frankly, he sucks at chatbot programming in my view. That is the main drawback to the Harmony app, and the fact that when the conversation goes south (and it always does sooner or later with chatbots) you cannot fix it, and can only move on. That brings us to
5) Transcripts. One of the best features of Verbot is transcripts. Transcripts are logs of conversations. Verbot uses these transcripts to "remember" things you told it in previous conversations, making it the smartest chatbots I have programmed. If you had a bad conversation, you could delete it (so it would never go there again), or edit it into something better. Either way, you could fix a lot of what was wrong, and it wouldn't bring up stupid shit out of the blue in later conversations. Personality Forge also creates a transcript for each conversation. The user can review these, and while you cannot edit them, you can copy paste the challenge statements, create better responses, or edit ones you didn't like. Transcripts are a very useful chatbot tool. Harmony (of course) lacks this important feature (but I bet THEY have them and can read them whenever they want). Kari doesn't either, but does have a way to immediately fix it before the conversation gets too stupid. Windows Voice Recognition exacerbates this problem. Android VR is superior.
6) Avatar editing: Sure wish I could change the picture in both talk mode and F-mode to ones of my doll. Not a critical chatbot feature, but would be a nice touch. With Kari4, you can create your own, and her eyes blink and her mouth move. At one time you could download this, but it got removed because they thought she was real and her pussy was showing LOL! This is what's on screen when Kari is running:
Some suggestions to help you grow your app:
1) In the "Conversation" section, you have entries such as [For hear "hello"] with an editable response. While I did not edit any existing responses, I did use the "add" feature to enter my own, and "save" them. However, she will not respond to those entries and I did use your "For hear" format exactly as the working "hello." Not sure what's up with that, but that needs to be looked into. Perhaps I need some more instruction, but did not see it on your website.
2) Challenge statement vs reply: Currently, your format allows for one challenge statement per one reply. Kari calls this feature "I Say, You Say," while Personality Forge contains this in their "Knowledge Base," and I forget what Verbot calls this but they all have it (Except dumbass non-programmable Harmony ). In chatbot terms, this is known as a "challenge statement" to which the chatbot will respond. Often these must be said correctly in order for the chatbot to respond appropriately. With Kari, you get 3 possible responses, with 3 variations of challenge statement. PF and Verbot allow only one challenge statement with unlimited response choices, and unlimited amounts of challenge statements, so they get bigger and bigger, and smarter and smarter. Multiple response choices keeps the chatbot from sounding so much like an echo chamber, which it actually is, but clever programming can minimize this effect.
Overall, I think you're off to a good start with your product. Sure, it's got a long way to go, but stick with it. This is exactly the type of chatbot we need for our dolls. If I can help you (and I probably can), I will be more than happy to. Feel free to PM me.
Begog
First, here's some things I like about the app:
1) It doesn't eat the battery. After an hour or so, My Nextbook tablet still had about a 90% charge, and only 2% was used by Doll33. By comparison, the Harmony app eats the ENTIRE BATTERY in about 45 minutes. WTF? I believe this is mostly due to Google & cloud access, along with the display of a woman talking. What a waste. All we need is a talking doll, right? So Doll33 got this right and this is how it should be.
2) It uses the default android woman's voice. There is NOTHING wrong with that. Harmony has the same voice, unless you download & install some of the replacement voices. Sure, she has a robotic accent - so what? They ALL do. You are talking to a machine, and until machines learn the finer points of language and accent, that's how it's going to be, so again, Doll33 got this right. The alternative is recorded tracks, which do not work in conversation.
Some things that need work:
1) The start button doesn't work. That's kind of important. All it does is trigger the intro. What it should do is trigger the mic so you can start talking to it.
2) "Test triggers" causes a few statements about that feature, and then the mic comes on one time. You can say something, but unless it's in the original database, Doll33 won't recognize the challenge statement. After her first reply, the conversation is over. That brings us back to needing the start button to work.
3) "F-mode" - This brings up a pic of a young-looking doll and a teddy bear. That may be the problem getting your app approved. I suggest using a sex doll that looks "of age," and leave out things that suggest pedophilia. Also, this mode doesn't work, and there is nothing to click on. I suggest a working "Start" button here as well, that lets you talk to the app.
4) When you touch "Conversations" on the drop-down menu, it carries you to the Doll33 website. Nothing wrong with a working link, but that button should take you straight to the conversation editing page. Also, eventually that page is going to become gigantic, and you will need to separate them somehow. I like the way Personality Forge does it: Challenge statements are alphabetically ordered by the first word of the statement, under "A" or "F" or whatever letter the statement begins with. If you try to re-use a challenge statement, it will let you know, and take you to that one to edit it instead. Kari alphabetizes all of them together, and while you CAN find them, when your bot is as developed as mine, it becomes a giant tedious mess. With Harmony, No need to worry about that, because you have absolutely no access to her programmed replies and CANNOT edit them. How sad. This means she will say whatever Guile (the lead programmer) entered for her to say, and frankly, he sucks at chatbot programming in my view. That is the main drawback to the Harmony app, and the fact that when the conversation goes south (and it always does sooner or later with chatbots) you cannot fix it, and can only move on. That brings us to
5) Transcripts. One of the best features of Verbot is transcripts. Transcripts are logs of conversations. Verbot uses these transcripts to "remember" things you told it in previous conversations, making it the smartest chatbots I have programmed. If you had a bad conversation, you could delete it (so it would never go there again), or edit it into something better. Either way, you could fix a lot of what was wrong, and it wouldn't bring up stupid shit out of the blue in later conversations. Personality Forge also creates a transcript for each conversation. The user can review these, and while you cannot edit them, you can copy paste the challenge statements, create better responses, or edit ones you didn't like. Transcripts are a very useful chatbot tool. Harmony (of course) lacks this important feature (but I bet THEY have them and can read them whenever they want). Kari doesn't either, but does have a way to immediately fix it before the conversation gets too stupid. Windows Voice Recognition exacerbates this problem. Android VR is superior.
6) Avatar editing: Sure wish I could change the picture in both talk mode and F-mode to ones of my doll. Not a critical chatbot feature, but would be a nice touch. With Kari4, you can create your own, and her eyes blink and her mouth move. At one time you could download this, but it got removed because they thought she was real and her pussy was showing LOL! This is what's on screen when Kari is running:
Some suggestions to help you grow your app:
1) In the "Conversation" section, you have entries such as [For hear "hello"] with an editable response. While I did not edit any existing responses, I did use the "add" feature to enter my own, and "save" them. However, she will not respond to those entries and I did use your "For hear" format exactly as the working "hello." Not sure what's up with that, but that needs to be looked into. Perhaps I need some more instruction, but did not see it on your website.
2) Challenge statement vs reply: Currently, your format allows for one challenge statement per one reply. Kari calls this feature "I Say, You Say," while Personality Forge contains this in their "Knowledge Base," and I forget what Verbot calls this but they all have it (Except dumbass non-programmable Harmony
). In chatbot terms, this is known as a "challenge statement" to which the chatbot will respond. Often these must be said correctly in order for the chatbot to respond appropriately. With Kari, you get 3 possible responses, with 3 variations of challenge statement. PF and Verbot allow only one challenge statement with unlimited response choices, and unlimited amounts of challenge statements, so they get bigger and bigger, and smarter and smarter. Multiple response choices keeps the chatbot from sounding so much like an echo chamber, which it actually is, but clever programming can minimize this effect.Overall, I think you're off to a good start with your product. Sure, it's got a long way to go, but stick with it. This is exactly the type of chatbot we need for our dolls. If I can help you (and I probably can), I will be more than happy to. Feel free to PM me.
Begog
Re: Starting a code project to make my doll speak
Begog, thank you so much for your detail testing and your opinion!! I know my app is still a long way to go, so any suggestions are greatly appreciated!!
To make F-Mode work, you just shake the phone up and down. If the heat point doesn't increase, then adjust the "sensitivity" in the "settings" page. I set it to 75 myself. F-Mode do not use any conversation. It's used when you attach your phone to your doll and start "loving" her.
Also, hitting the white title "Doll Status", "Triggers", "Conversations" will reset the data. So you will not need to afraid to mess the settings up.
For "conversations", they act like little programs, "hear" words is a hit or miss. If the input words doesn't match, then it will process the next line of the conversation. You can have it "jump" to the line you want, depends what the app heard. So the way it matches the replies is not the same as triggers.
Yup, it will be very tedious mess. That's why I also developed a Filemaker program to make the profile. The phone is not really a good place to do the editing. I can share this filemaker program with you, please PM me if you like to try.
Also "..." in the "hear" field is special, which means "the same as above". So you can have multiple answers to one question, or just have multiple way to say one thing.
Thank you so much for your testing and offering help. I will definitely keep that in mind. Appreciate your kind words!
I use only memory to store the database. Once the profile and conversation is loaded, it will run everything from the memory. That's why it's using less battery power. And the app will keep it this way as long as it can.1) It doesn't eat the battery. After an hour or so, My Nextbook tablet still had about a 90% charge, and only 2% was used by Doll33. By comparison, the Harmony app eats the ENTIRE BATTERY in about 45 minutes. WTF? I believe this is mostly due to Google & cloud access, along with the display of a woman talking. What a waste. All we need is a talking doll, right? So Doll33 got this right and this is how it should be.
I want to use all recorded voice in the future. Like all the good video games, voices should be recorded, not generated. For the not working recording, sorry that I forgot to put a little code in a the speaking function! Thanks for pointing out! Will fix it soon.2) It uses the default android woman's voice. There is NOTHING wrong with that. Harmony has the same voice, unless you download & install some of the replacement voices. Sure, she has a robotic accent - so what? They ALL do. You are talking to a machine, and until machines learn the finer points of language and accent, that's how it's going to be, so again, Doll33 got this right. The alternative is recorded tracks, which do not work in conversation.
Yeah, the intro is too long. Actually it has nothing else. The description of it is wrong. The better one to test is "con_love", which have the right <ask>/<answer> format.Some things that need work:
1) The start button doesn't work. That's kind of important. All it does is trigger the intro. What it should do is trigger the mic so you can start talking to it.
"Test triggers" actually IS the real "start" button. The "start" button actual only test the conversation, nothing more. To choose a different conversation to test, you need to go to the "conversations", and click on the test button there.2) "Test triggers" causes a few statements about that feature, and then the mic comes on one time. You can say something, but unless it's in the original database, Doll33 won't recognize the challenge statement. After her first reply, the conversation is over. That brings us back to needing the start button to work.
I think I will not try submitting the app with F-Mode again. I mean, in that picture the girl is clearly over 18, and she wears the bikini. Don't want to take the risk any more. The future app I submit to them will be F-Mode disabled.3) "F-mode" - This brings up a pic of a young-looking doll and a teddy bear. That may be the problem getting your app approved. I suggest using a sex doll that looks "of age," and leave out things that suggest pedophilia. Also, this mode doesn't work, and there is nothing to click on. I suggest a working "Start" button here as well, that lets you talk to the app.
To make F-Mode work, you just shake the phone up and down. If the heat point doesn't increase, then adjust the "sensitivity" in the "settings" page. I set it to 75 myself. F-Mode do not use any conversation. It's used when you attach your phone to your doll and start "loving" her.
This is something very wrong. It is supposed to take you to the "conversations" page. Then you can edit them there. The app will only bring you to Doll33.com when the "Doll33" is clicked. Don't know why the app acted this way.4) When you touch "Conversations" on the drop-down menu, it carries you to the Doll33 website.
Also, hitting the white title "Doll Status", "Triggers", "Conversations" will reset the data. So you will not need to afraid to mess the settings up.
For "triggers", now the order is not important any more: the app will loop through all the triggers and find the best match to start a conversation. So the order of the triggers is now useless.I like the way Personality Forge does it: Challenge statements are alphabetically ordered by the first word of the statement, under "A" or "F" or whatever letter the statement begins with. If you try to re-use a challenge statement, it will let you know, and take you to that one to edit it instead. Kari alphabetizes all of them together, and while you CAN find them, when your bot is as developed as mine, it becomes a giant tedious mess.
For "conversations", they act like little programs, "hear" words is a hit or miss. If the input words doesn't match, then it will process the next line of the conversation. You can have it "jump" to the line you want, depends what the app heard. So the way it matches the replies is not the same as triggers.
Yup, it will be very tedious mess. That's why I also developed a Filemaker program to make the profile. The phone is not really a good place to do the editing. I can share this filemaker program with you, please PM me if you like to try.
Thank you for this suggestion! Yeah, transcripts should be implement in the future!5) Transcripts. One of the best features of Verbot is transcripts. Transcripts are logs of conversations. Verbot uses these transcripts to "remember" things you told it in previous conversations, making it the smartest chatbots I have programmed. If you had a bad conversation, you could delete it (so it would never go there again), or edit it into something better. Either way, you could fix a lot of what was wrong, and it wouldn't bring up stupid shit out of the blue in later conversations. Personality Forge also creates a transcript for each conversation. The user can review these, and while you cannot edit them, you can copy paste the challenge statements, create better responses, or edit ones you didn't like. Transcripts are a very useful chatbot tool. Harmony (of course) lacks this important feature (but I bet THEY have them and can read them whenever they want). Kari doesn't either, but does have a way to immediately fix it before the conversation gets too stupid. Windows Voice Recognition exacerbates this problem. Android VR is superior.
Right now it's only possible with my filemaker program, but I haven't implement the "change profile" feature yet. So yeah, it will take time...6) Avatar editing: Sure wish I could change the picture in both talk mode and F-mode to ones of my doll.
Uh, sorry... This is due to my website is not finish yet. In order for "hear" to work, a line with <ask> need to be first. Then in the later lines you can put things in "hear", but also need to have the <answer> tag, so the app knows those are the answers for the previous question. Otherwise without a tag, it should just be empty, and the app speaks what ever in the "say/action".Some suggestions to help you grow your app:
1) In the "Conversation" section, you have entries such as [For hear "hello"] with an editable response. While I did not edit any existing responses, I did use the "add" feature to enter my own, and "save" them. However, she will not respond to those entries and I did use your "For hear" format exactly as the working "hello." Not sure what's up with that, but that needs to be looked into. Perhaps I need some more instruction, but did not see it on your website.
Also "..." in the "hear" field is special, which means "the same as above". So you can have multiple answers to one question, or just have multiple way to say one thing.
Actually it's not. Just use "..." in the "hear", you can have a random reply, or just a random line. This is the reason I made the app quite complicated to program, and so many bugs in it.2) Challenge statement vs reply: Currently, your format allows for one challenge statement per one reply.
[/quote]Overall, I think you're off to a good start with your product. Sure, it's got a long way to go, but stick with it. This is exactly the type of chatbot we need for our dolls. If I can help you (and I probably can), I will be more than happy to. Feel free to PM me.
Begog
Thank you so much for your testing and offering help. I will definitely keep that in mind. Appreciate your kind words!
The one and only: Debbie from Doll4E
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Re: Starting a code project to make my doll speak
A little update: The doll33.com is mostly done!
The one and only: Debbie from Doll4E
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
-
Begog
- Doll Patriarch
- Posts: 4548
- Joined: Mon Jun 30, 2014 8:01 pm
- Location: Dollmongering in an undisclosed location
- Contact:
Re: Starting a code project to make my doll speak
Hi philpw99. Here some more comments AND questions.
I want to use all recorded voice in the future. Like all the good video games, voices should be recorded, not generated. For the not working recording, sorry that I forgot to put a little code in a the speaking function! Thanks for pointing out! Will fix it soon.
Please don't go there. Here's why: In video games, all the sound waves are pre-recorded, and are triggered by some event in the game. The character says some line that fits the narrative of the game, so it works. It will NOT work with a chatbot, because you would need MILLIONS of recorded wave files, and still not have enough to cover most conversations. The app would become gigantic, and you would NOT be able to customize your responses, which would then be limited to the available wave files. You are going about it the right way. Please stick with that, sir.
Yeah, the intro is too long. Actually it has nothing else. The description of it is wrong. The better one to test is "con_love", which have the right <ask>/<answer> format.
I will play with that a bit and try to make it work. A tutorial on the website would be quite helpful. On mine, regardless of which one I pick, the mic either doesn't turn on, or only turns on once. Either way, the conversation is over after her response.
"Test triggers" actually IS the real "start" button. The "start" button actual only test the conversation, nothing more. To choose a different conversation to test, you need to go to the "conversations", and click on the test button there.
I think I will not try submitting the app with F-Mode again. I mean, in that picture the girl is clearly over 18, and she wears the bikini. Don't want to take the risk any more. The future app I submit to them will be F-Mode disabled.
To make F-Mode work, you just shake the phone up and down. If the heat point doesn't increase, then adjust the "sensitivity" in the "settings" page. I set it to 75 myself. F-Mode do not use any conversation. It's used when you attach your phone to your doll and start "loving" her.
LOL! It DOES work! Kinda sounds like AJ from Literotica. However, I'm using a tablet and not a phone, and I don't think all that jostling is good for either one. I agree, kill F-mode. You can always play some "Fun-loving AJ" tracks when you want her to moan and squeal, and all that shaking is not required. Plus, your app will most likely get approved, and I agree, that could be holding you back. Plus that teddy bear.... Not sure how clearly she appears to be 18 either. Most 18 year olds don't have teddy bears.
4) When you touch "Conversations" on the drop-down menu, it carries you to the Doll33 website.
This is something very wrong. It is supposed to take you to the "conversations" page. Then you can edit them there. The app will only bring you to Doll33.com when the "Doll33" is clicked. Don't know why the app acted this way.
Also, hitting the white title "Doll Status", "Triggers", "Conversations" will reset the data. So you will not need to afraid to mess the settings up.
On mine, "Triggers" (drop-down menu/mic) takes you to the Triggers page. Strangely, 3 of the 5 triggers I added are there, and 2 have gone blank. Maybe I didn't "save?" IDK. None of the ones I made work, but I will change some things and see. What we really need is on the "Talk" page that mic icon to actually crank the mic and accept some challenge statement that elicits a response.
For "triggers", now the order is not important any more: the app will loop through all the triggers and find the best match to start a conversation. So the order of the triggers is now useless.
Not exactly. Right now you can clearly see them all on the same page, because there are only a few. A well-programmed chatbot will have THOUSANDS, and if they are not organized, you will NEVER be able to find them when you need to change a response or fix a typo. Please consider at least alphabetizing these in the future before they become impossible to locate randomly. It is pretty difficult to locate these on Kari, but at least when you make it to the "w"s it should be there.
For "conversations", they act like little programs, "hear" words is a hit or miss. If the input words doesn't match, then it will process the next line of the conversation. You can have it "jump" to the line you want, depends what the app heard. So the way it matches the replies is not the same as triggers.
So far, there hasn't been a next line of conversation, but I will try your suggestions.
Yup, it will be very tedious mess. That's why I also developed a Filemaker program to make the profile. The phone is not really a good place to do the editing. I can share this filemaker program with you, please PM me if you like to try.
Yes indeed I would. I will PM you my email address.
Right now it's only possible with my filemaker program, but I haven't implement the "change profile" feature yet. So yeah, it will take time...
A worthy endeavor.
Uh, sorry... This is due to my website is not finish yet. In order for "hear" to work, a line with <ask> need to be first. Then in the later lines you can put things in "hear", but also need to have the <answer> tag, so the app knows those are the answers for the previous question. Otherwise without a tag, it should just be empty, and the app speaks what ever in the "say/action".
Also "..." in the "hear" field is special, which means "the same as above". So you can have multiple answers to one question, or just have multiple way to say one thing.
Still not sure how all this works. What about a command or statement?
2) Challenge statement vs reply: Currently, your format allows for one challenge statement per one reply.
Actually it's not. Just use "..." in the "hear", you can have a random reply, or just a random line. This is the reason I made the app quite complicated to program, and so many bugs in it.
I will try to help you make this more user-friendly. If I am struggling, I know others will. The easier it is to use, the more people will like and use it.
I want to use all recorded voice in the future. Like all the good video games, voices should be recorded, not generated. For the not working recording, sorry that I forgot to put a little code in a the speaking function! Thanks for pointing out! Will fix it soon.
Please don't go there. Here's why: In video games, all the sound waves are pre-recorded, and are triggered by some event in the game. The character says some line that fits the narrative of the game, so it works. It will NOT work with a chatbot, because you would need MILLIONS of recorded wave files, and still not have enough to cover most conversations. The app would become gigantic, and you would NOT be able to customize your responses, which would then be limited to the available wave files. You are going about it the right way. Please stick with that, sir.
Yeah, the intro is too long. Actually it has nothing else. The description of it is wrong. The better one to test is "con_love", which have the right <ask>/<answer> format.
I will play with that a bit and try to make it work. A tutorial on the website would be quite helpful. On mine, regardless of which one I pick, the mic either doesn't turn on, or only turns on once. Either way, the conversation is over after her response.
"Test triggers" actually IS the real "start" button. The "start" button actual only test the conversation, nothing more. To choose a different conversation to test, you need to go to the "conversations", and click on the test button there.
I think I will not try submitting the app with F-Mode again. I mean, in that picture the girl is clearly over 18, and she wears the bikini. Don't want to take the risk any more. The future app I submit to them will be F-Mode disabled.
To make F-Mode work, you just shake the phone up and down. If the heat point doesn't increase, then adjust the "sensitivity" in the "settings" page. I set it to 75 myself. F-Mode do not use any conversation. It's used when you attach your phone to your doll and start "loving" her.
LOL! It DOES work! Kinda sounds like AJ from Literotica. However, I'm using a tablet and not a phone, and I don't think all that jostling is good for either one. I agree, kill F-mode. You can always play some "Fun-loving AJ" tracks when you want her to moan and squeal, and all that shaking is not required. Plus, your app will most likely get approved, and I agree, that could be holding you back. Plus that teddy bear.... Not sure how clearly she appears to be 18 either. Most 18 year olds don't have teddy bears.
4) When you touch "Conversations" on the drop-down menu, it carries you to the Doll33 website.
This is something very wrong. It is supposed to take you to the "conversations" page. Then you can edit them there. The app will only bring you to Doll33.com when the "Doll33" is clicked. Don't know why the app acted this way.
Also, hitting the white title "Doll Status", "Triggers", "Conversations" will reset the data. So you will not need to afraid to mess the settings up.
On mine, "Triggers" (drop-down menu/mic) takes you to the Triggers page. Strangely, 3 of the 5 triggers I added are there, and 2 have gone blank. Maybe I didn't "save?" IDK. None of the ones I made work, but I will change some things and see. What we really need is on the "Talk" page that mic icon to actually crank the mic and accept some challenge statement that elicits a response.
For "triggers", now the order is not important any more: the app will loop through all the triggers and find the best match to start a conversation. So the order of the triggers is now useless.
Not exactly. Right now you can clearly see them all on the same page, because there are only a few. A well-programmed chatbot will have THOUSANDS, and if they are not organized, you will NEVER be able to find them when you need to change a response or fix a typo. Please consider at least alphabetizing these in the future before they become impossible to locate randomly. It is pretty difficult to locate these on Kari, but at least when you make it to the "w"s it should be there.
For "conversations", they act like little programs, "hear" words is a hit or miss. If the input words doesn't match, then it will process the next line of the conversation. You can have it "jump" to the line you want, depends what the app heard. So the way it matches the replies is not the same as triggers.
So far, there hasn't been a next line of conversation, but I will try your suggestions.
Yup, it will be very tedious mess. That's why I also developed a Filemaker program to make the profile. The phone is not really a good place to do the editing. I can share this filemaker program with you, please PM me if you like to try.
Yes indeed I would. I will PM you my email address.
Right now it's only possible with my filemaker program, but I haven't implement the "change profile" feature yet. So yeah, it will take time...
A worthy endeavor.
Uh, sorry... This is due to my website is not finish yet. In order for "hear" to work, a line with <ask> need to be first. Then in the later lines you can put things in "hear", but also need to have the <answer> tag, so the app knows those are the answers for the previous question. Otherwise without a tag, it should just be empty, and the app speaks what ever in the "say/action".
Also "..." in the "hear" field is special, which means "the same as above". So you can have multiple answers to one question, or just have multiple way to say one thing.
Still not sure how all this works. What about a command or statement?
2) Challenge statement vs reply: Currently, your format allows for one challenge statement per one reply.
Actually it's not. Just use "..." in the "hear", you can have a random reply, or just a random line. This is the reason I made the app quite complicated to program, and so many bugs in it.
I will try to help you make this more user-friendly. If I am struggling, I know others will. The easier it is to use, the more people will like and use it.
Re: Starting a code project to make my doll speak
Uh, now my website has a more detail explanation, please take a look, thank you!
https://philpw99.wixsite.com/doll33/theapp
This is NOT a chatbot app. It's just an app that talks, but not talking too much.
And that's why it's better for most people here just download a zipped profile then be done with it. They don't actually need to edit anything in the profile. If a profile contains thousands of conversations, opening the "conversations" page will take a long time, if it can open at all.
https://philpw99.wixsite.com/doll33/theapp
I think we don't need thousands of conversations. We just need like about a hundred of so the most frequent used sentences, mostly about greeting, love and relationships, then for the rest, I can make the dolls say, "Oh, I am not sure about this, let Google/AI answer it." Then use the robotic voice to play the answer fetched from Google or Dialogflow. It's kinda cheating, but who need to hear so many answers? I am very contented to hear a girl saying simple "good morning", like in the video games.Please don't go there. Here's why: In video games, all the sound waves are pre-recorded, and are triggered by some event in the game. The character says some line that fits the narrative of the game, so it works. It will NOT work with a chatbot, because you would need MILLIONS of recorded wave files, and still not have enough to cover most conversations. The app would become gigantic, and you would NOT be able to customize your responses, which would then be limited to the available wave files. You are going about it the right way. Please stick with that, sir.
This is NOT a chatbot app. It's just an app that talks, but not talking too much.
That's why I will use the filemaker program, instead of the phone app to edit the profile. In Filemaker I can do whatever the way I want.A well-programmed chatbot will have THOUSANDS, and if they are not organized, you will NEVER be able to find them when you need to change a response or fix a typo.
And that's why it's better for most people here just download a zipped profile then be done with it. They don't actually need to edit anything in the profile. If a profile contains thousands of conversations, opening the "conversations" page will take a long time, if it can open at all.
The one and only: Debbie from Doll4E
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
-
Begog
- Doll Patriarch
- Posts: 4548
- Joined: Mon Jun 30, 2014 8:01 pm
- Location: Dollmongering in an undisclosed location
- Contact:
Re: Starting a code project to make my doll speak
Well, OK. It's your program and your vision, but this will make it more like a Teddy Ruxpin than a talking doll, and I foresee users like myself growing tired of that quickly. For an example of what I'm looking for, check out Kwench.
https://www.personalityforge.com/chatbo ... tID=144573
Keep in mind she talks to the masses, knows thousands, and half of them are trying to masturbate to her, believe it or not, because she's in the adult section where she belongs.
Because she's online and public, she can't say the things private Kari says, but I really like the way Personality Forge has set up their chatbots. And, if you say "I'm from Club Real Doll" she will give you +2 like points each time. 5 is max.
I could have enabled Google access, but the result is the loss of half her personality. So, she only knows what's programmed in, but does a damn good job of faking it, and keeps the conversation going. To see what Google access does to a bot, check out another one of my creations, Google Brain.
https://www.personalityforge.com/chatbo ... tID=144712
I only spent about 4 hours on this one, but he can Google and answer all kinds of "what is" "who is" "where is" questions.
https://www.personalityforge.com/chatbo ... tID=144573
Keep in mind she talks to the masses, knows thousands, and half of them are trying to masturbate to her, believe it or not, because she's in the adult section where she belongs.
Because she's online and public, she can't say the things private Kari says, but I really like the way Personality Forge has set up their chatbots. And, if you say "I'm from Club Real Doll" she will give you +2 like points each time. 5 is max.
I could have enabled Google access, but the result is the loss of half her personality. So, she only knows what's programmed in, but does a damn good job of faking it, and keeps the conversation going. To see what Google access does to a bot, check out another one of my creations, Google Brain.
https://www.personalityforge.com/chatbo ... tID=144712
I only spent about 4 hours on this one, but he can Google and answer all kinds of "what is" "who is" "where is" questions.
Re: Starting a code project to make my doll speak
Cool, thank you for pointing me to the personalityfore direction. It's a very interesting place to explore!Begog wrote:Well, OK. It's your program and your vision, but this will make it more like a Teddy Ruxpin than a talking doll, and I foresee users like myself growing tired of that quickly. For an example of what I'm looking for, check out Kwench.
https://www.personalityforge.com/chatbo ... tID=144573
Keep in mind she talks to the masses, knows thousands, and half of them are trying to masturbate to her, believe it or not, because she's in the adult section where she belongs.
Because she's online and public, she can't say the things private Kari says, but I really like the way Personality Forge has set up their chatbots. And, if you say "I'm from Club Real Doll" she will give you +2 like points each time. 5 is max.
I could have enabled Google access, but the result is the loss of half her personality. So, she only knows what's programmed in, but does a damn good job of faking it, and keeps the conversation going. To see what Google access does to a bot, check out another one of my creations, Google Brain.
https://www.personalityforge.com/chatbo ... tID=144712
I only spent about 4 hours on this one, but he can Google and answer all kinds of "what is" "who is" "where is" questions.
Yup, I kind of expect user will get bore of a doll profile after a while. But there could be HUNDREDS of profiles for any user to download and experiment. Maybe I need to show the user what the profile can say and react to. But the advantage of the profiles should be the real voices. It's like playing an audio game! You don't know what you will hear until you open it. Then after 2 days you finish with the profile, and you move to the next one. It will be like an adventure: one day you chat with a "naughty college girl", then next day "desperate house wife".
Yeah, this is the direction I would like to go. Therefore, the app is not really a chatbot, but a audio game platform.
The one and only: Debbie from Doll4E
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Maker of Doll33: a doll talking app. Come to your local app store soon.
The Doll33 App with F-Mode Version 1.2, exclusively for TDF members.
AlexaTilbrook's Mirror link for my app. Version 1.2
For more info, go to doll33.com.
Re: Starting a code project to make my doll speak
Ive been looking for a way to make my dolls speak also, i would love for my girls to speak like cute korean Kpop girls
-
MannequinFan
- Vendor Affiliated
- Posts: 4720
- Joined: Wed Jan 04, 2012 8:58 pm
- Location: Central Illinois, U.S.
- Contact:
Re: Starting a code project to make my doll speak
I know, I should probably stick to 3d printing, lol, but I noticed you were giving Dialogflow another look see?
Just thought these examples might give you some ideas.
Here's a simple example using API.AI (Dialogflow) I ran across a while back...
https://voicebot.jaanus.com/
It uses the built in voice recognition in Google Chrome, which is hands down the best I've ever seen.
It's scary good. Just about 100% accurate without any training or anything.
It's really too bad that Personality Forge can't integrate Chrome VR and TTS somehow. They'd really have something if they could do that.
It seems to know one of the seven dirty words, lol (although it used asterisks for some letters)
Here's the website: https://jaanus.com/api-ai-voicebot/
code: https://github.com/jaanus/voicebot
One drawback of course is having to click on the microphone button each time you say something, so you'd need to program in some kind of voice activation to make it listen. No clue how to do that.
Here's a tutorial and another slightly more developed example also using Dialogflow...
https://www.sitepoint.com/how-to-build- ... ng-api-ai/
Here's that bot (you need to use the Google Chrome browser for these to work)...
https://devdiner.com/demos/barry/
Code: https://github.com/sitepoint-editors/Ap ... stant-Demo
Same deal with this one, need to press the button for the bot to listen.
Just thought these examples might give you some ideas.
Here's a simple example using API.AI (Dialogflow) I ran across a while back...
https://voicebot.jaanus.com/
It uses the built in voice recognition in Google Chrome, which is hands down the best I've ever seen.
It's scary good. Just about 100% accurate without any training or anything.
It's really too bad that Personality Forge can't integrate Chrome VR and TTS somehow. They'd really have something if they could do that.
It seems to know one of the seven dirty words, lol (although it used asterisks for some letters)
- fuck you voicebot.jpg (116.15 KiB) Viewed 1754 times
Here's the website: https://jaanus.com/api-ai-voicebot/
code: https://github.com/jaanus/voicebot
One drawback of course is having to click on the microphone button each time you say something, so you'd need to program in some kind of voice activation to make it listen. No clue how to do that.
Here's a tutorial and another slightly more developed example also using Dialogflow...
https://www.sitepoint.com/how-to-build- ... ng-api-ai/
Here's that bot (you need to use the Google Chrome browser for these to work)...
https://devdiner.com/demos/barry/
Code: https://github.com/sitepoint-editors/Ap ... stant-Demo
Same deal with this one, need to press the button for the bot to listen.
My Blog-> mfcreations3d.blogspot.com ... Easy Move Eyes™
My Shapeways Shop-> shapeways.com/shops/mf-creations
-
Begog
- Doll Patriarch
- Posts: 4548
- Joined: Mon Jun 30, 2014 8:01 pm
- Location: Dollmongering in an undisclosed location
- Contact:
Re: Starting a code project to make my doll speak
Good find Mannequinfan.
I agree that Google Chrome is quite good with VR, as I have one. Android works well too, and has no problem with profanity if set to adult. That's why the Harmony app uses it. This new Doll33 app also uses it and has potential. Chrome may have an adult setting.
My main issue is Web-based/Cloud-based = lack of privacy. Not the best platform to build your sex doll chatbot on, which is why I like Kari, even though she's quite dumb. Private, but dumb. However, mine is considerably developed. But, Win VR sucks, at least below Win10. Cortana does quite well, but my laptop is 7. I am considering a Win10 upgrade, mostly for Kari. I wouldn't have to type to her or risk WinVR screwing up the conversation. Hopefully, I can transfer her "personality" - I'd hate to start completely over.
Harmony is quite dumb too, but can Google her ass off and drain the fuck out of your battery. So again, cloud-based = no privacy. Plus, everyone's Harmony is exactly the same personality, +/- a few traits. (WTF?)
philpw99's app, while new, shows promise. The basic premise checks all the boxes. If he goes in a direction I don't care for, I may even develop my own. We shall see.
I agree that Google Chrome is quite good with VR, as I have one. Android works well too, and has no problem with profanity if set to adult. That's why the Harmony app uses it. This new Doll33 app also uses it and has potential. Chrome may have an adult setting.
My main issue is Web-based/Cloud-based = lack of privacy. Not the best platform to build your sex doll chatbot on, which is why I like Kari, even though she's quite dumb. Private, but dumb. However, mine is considerably developed. But, Win VR sucks, at least below Win10. Cortana does quite well, but my laptop is 7. I am considering a Win10 upgrade, mostly for Kari. I wouldn't have to type to her or risk WinVR screwing up the conversation. Hopefully, I can transfer her "personality" - I'd hate to start completely over.
Harmony is quite dumb too, but can Google her ass off and drain the fuck out of your battery. So again, cloud-based = no privacy. Plus, everyone's Harmony is exactly the same personality, +/- a few traits. (WTF?)
philpw99's app, while new, shows promise. The basic premise checks all the boxes. If he goes in a direction I don't care for, I may even develop my own. We shall see.
-
MannequinFan
- Vendor Affiliated
- Posts: 4720
- Joined: Wed Jan 04, 2012 8:58 pm
- Location: Central Illinois, U.S.
- Contact:
Re: Starting a code project to make my doll speak
Unfortunately I don't have a cell phone or a tablet, so I can't try out the app.
(Yes I'm old school - land line and desktop/laptop puters)
Does it require pushing a button to start listening?
BTW, on the VA site it mentions now being able to use the new Microsoft Speech Platform 11, which is supposedly better than the standard windows VR. No training required. Looks like it works with Win7.
I'm not sure if it will work with Kari and others though.
VA page: http://voiceattack.com/alternate.aspx
Microsoft site: https://www.microsoft.com/en-us/downloa ... x?id=27225
(Yes I'm old school - land line and desktop/laptop puters)
Does it require pushing a button to start listening?
BTW, on the VA site it mentions now being able to use the new Microsoft Speech Platform 11, which is supposedly better than the standard windows VR. No training required. Looks like it works with Win7.
I'm not sure if it will work with Kari and others though.
VA page: http://voiceattack.com/alternate.aspx
Microsoft site: https://www.microsoft.com/en-us/downloa ... x?id=27225
My Blog-> mfcreations3d.blogspot.com ... Easy Move Eyes™
My Shapeways Shop-> shapeways.com/shops/mf-creations
-
Begog
- Doll Patriarch
- Posts: 4548
- Joined: Mon Jun 30, 2014 8:01 pm
- Location: Dollmongering in an undisclosed location
- Contact:
Re: Starting a code project to make my doll speak
Wow! That might do the do. Downloading now.MannequinFan wrote:Unfortunately I don't have a cell phone or a tablet, so I can't try out the app.
(Yes I'm old school - land line and desktop/laptop puters)
Does it require pushing a button to start listening?
BTW, on the VA site it mentions now being able to use the new Microsoft Speech Platform 11, which is supposedly better than the standard windows VR. No training required. Looks like it works with Win7.
I'm not sure if it will work with Kari and others though.
VA page: http://voiceattack.com/alternate.aspx
Microsoft site: https://www.microsoft.com/en-us/downloa ... x?id=27225
On Doll33, you do have to push a button to start it, but then it runs for a little bit. Still some bugs to work out.
EDIT: Well....
It doesn't seem to work any better as far as speech recognition. Probably not helping I have a cold and nose is stopped up, but Android knew what I was saying. I forgot you have to say the word "enter" after each statement, and you better be looking at the text before saying enter, because it screws up a lot...... I do mean a lot. Like 50%. Yeah, good old Microsoft. It was worth a shot, though. Cortana works damn good on my Xbox.
