Diary Of Mine for Paradise Halls

[Jasmine92]

On 4/22/2022 at 1:41 AM, Jasmine92 said:

I just played the Kimli quest line. Nicely done and quite interesting to get the mechanics known a bit. I had two things, that where a little bit off with my game:

• Kimli had no voiced lines, besides the 2 or 3 sentences before you enslave her. Is that really the case, because her voice type is uncommon, or is that because my voice pack is missing her lines (for whatever reason, which would be a curious bug, because I exported all voice lines the  xEdit Script found

 

So I found the problem why Kimly (and in extension NPCs with unique voices like Maven, Ulfirc or Astrid) are not included in the voicepack I created. (See over here, if you are interested) Question now is, should these unique Voices be added to the Voice pack for the next update? Or is no one enslaving NPCs like Brynjolf and Maven anyhow so it doesn't matter? Adding everyone will add another 41k fuz files to the game and I'm not sure if at some point we will hit a wall like with the maximum number of animations. (Some people are already reporting that voices just stop working mid game, but that might not be related) I will add Vex anyway, since that's Kimlys voice.

 

Voices that are missing are:

Spoiler

Alduin
Odahviing
Paarthurnax
Fura
Valerica
Dexion
Garan
Gelebor
Harkon
Isran
Frea
Adril
Lleril
Modyn
Neloth
Storn
Astrid
Delphine
Elenwen
Karliah
Maven
MirabelleErvine
Vex
Ancano
Arngeir
DelvinMallory
Esbern
Galmar
KodlakWhitemane
MercerFrey
Nazir
Tullius
Ulfric
Hadvar
Brynjolf

 

Edited April 25, 2022 by Jasmine92
clarification

[TrollAutokill]

57 minutes ago, Jasmine92 said:

So I found the problem why Kimly (and in extension NPCs with unique voices like Maven, Ulfirc or Astrid) are not included in the voicepack I created. (See over here, if you are interested) Question now is, should these unique Voices be added to the Voice pack for the next update? Or is no one enslaving NPCs like Brynjolf and Maven anyhow so it doesn't matter? Adding everyone will add another 41k fuz files to the game and I'm not sure if at some point we will hit a wall like with the maximum number of animations. (Some people are already reporting that voices just stop working mid game, but that might not be related) I will add Vex anyway, since that's Kimlys voice.

 

Voices that are missing are:

  Hide contents

Alduin
Odahviing
Paarthurnax
Fura
Valerica
Dexion
Garan
Gelebor
Harkon
Isran
Frea
Adril
Lleril
Modyn
Neloth
Storn
Astrid
Delphine
Elenwen
Karliah
Maven
MirabelleErvine
Vex
Ancano
Arngeir
DelvinMallory
Esbern
Galmar
KodlakWhitemane
MercerFrey
Nazir
Tullius
Ulfric
Hadvar
Brynjolf

 

Vex voice is interesting since it is also Kimli voice. The rest could he in separate packs if there is a demand for them.

[Kharos]

10 hours ago, TrollAutokill said:

I couldn't reproduce the run away faction not removed. If you want to try, you need to turn on the verbose mode from DoM debug menu and look for the end of the run away behaviour state: "end state=run_away". If it is there the faction should be correctly removed, unless there is an error message.

 

I doubt that it can easily be reproduced. I had her try to run away 5 or 6 times, and it only happened once. As I wrote, maybe a race condition, or a lost event or something else acting up.

 

I did make a savegame when she was in that bad state (faction set but not in run_away state). If you add fallback code to catch and fix the issue, e.g. in OnToldOff or anywhere else, I can use that savegame to test if your fix is working.

 

[InsanityFactor]

1 hour ago, Jasmine92 said:

So I found the problem why Kimly (and in extension NPCs with unique voices like Maven, Ulfirc or Astrid) are not included in the voicepack I created. (See over here, if you are interested) Question now is, should these unique Voices be added to the Voice pack for the next update? Or is no one enslaving NPCs like Brynjolf and Maven anyhow so it doesn't matter? Adding everyone will add another 41k fuz files to the game and I'm not sure if at some point we will hit a wall like with the maximum number of animations. (Some people are already reporting that voices just stop working mid game, but that might not be related) I will add Vex anyway, since that's Kimlys voice.

 

Voices that are missing are:

  Reveal hidden contents

Alduin
Odahviing
Paarthurnax
Fura
Valerica
Dexion
Garan
Gelebor
Harkon
Isran
Frea
Adril
Lleril
Modyn
Neloth
Storn
Astrid
Delphine
Elenwen
Karliah
Maven
MirabelleErvine
Vex
Ancano
Arngeir
DelvinMallory
Esbern
Galmar
KodlakWhitemane
MercerFrey
Nazir
Tullius
Ulfric
Hadvar
Brynjolf

 

Well looking at the list we can just chop a bunch of them since they either shouldn't be enslaved for game stability reasons or no one would really want to.

 

Here's my first pass at it. 

 

Spoiler

1 hour ago, Jasmine92 said:

Alduin: no
Odahviing: no
Paarthurnax: no
Fura: sure
Valerica: probably shouldn't? 
Dexion: probably shouldn't? 
Garan: sure
Gelebor: probably no
Harkon: probably shouldn't 
Isran: maybe
Frea: sure
Adril: probably shouldn't 
Lleril: probably shouldn't 
Modyn: he's a guard so idk if that's safe to enslave? 
Neloth: might break something 
Storn: maybe
Astrid: yes
Delphine: probably fine
Elenwen: probably fine? 
Karliah: yes 
Maven: yeah
MirabelleErvine: sure 
Vex: yes
Ancano: probably gonna break something 
Arngeir: might break something 
DelvinMallory: yeah
Esbern: might break something 
Galmar: sure
KodlakWhitemane: probably no
MercerFrey: sure
Nazir: yeah
Tullius: sure
Ulfric: yeah sure
Hadvar: might break something 
Brynjolf: sure

 

If you only did the ones marked "yes" or "sure", would it be more viable? I only marked them yes or sure if I have enslaved them before without issue or if they can be killed through player choice anyway. (destroy the thieves guild would be needed to enslave them safely I think) 

 

The ones I'm not sure about would probably need some testing. If anyone has enslaved them successfully then I'd say go for it. But if you want to make voice files and prioritize them then I'd say you should start with the most highly demanded ones, and then among those start with the ones that you can kill in the vanilla game, and then move on to the ones you can kill with mods, and then if people can confirm that it's stable you would move on to the ones that you can't kill or have a lot of quest baggage. Like for example if you enslave Mirabelle and then do the College quest where she dies, would the game delete your slave version of her too? Things like that should be considered before deciding if it's even worth it to make the voice files. 

 

I mean you're just doing this pro bono and it's really appreciated, so do it however you want, but I think starting with the girls like Vex and Astrid would be a good start, then maybe some of the boys. 

 

One thing I wanted to mention too is what about the reavers in Solstheim? Can those voices be done as well? 

Edited April 25, 2022 by InsanityFactor

[TrollAutokill]

12 minutes ago, InsanityFactor said:

One thing I wanted to mention too is what about the reavers in Solstheim? Can those voices be done as well? 

I don't think those voices were trained and available. The problem is training a voice from samples takes a lot of ressources (like several GPU cards) and a lot of time. If I could I would train new voices...

[InsanityFactor]

31 minutes ago, TrollAutokill said:

I don't think those voices were trained and available. The problem is training a voice from samples takes a lot of ressources (like several GPU cards) and a lot of time. If I could I would train new voices...

One thing I wonder about too is the accents. Would thick, unique accents like Delvin, Frea, and Mirabelle carry over well? Or would it fail to piece the words together phonetically? More of a question for @Jasmine92 I guess. 

[DieHans]

On 4/25/2022 at 9:15 AM, TrollAutokill said:

The idea is to have the master and trainers work to control the slaves orgasms, not a random algorithm deciding.

Now I may understand thats a fantastic idea. If i am correct the final goal is to have a working permission system with slaves orgasm control, punishmemt reason and so on? Else I probably misunderstood the word "control". That would be a huge deal and probably a lot of work.

 

But what you want in the end is to use the seperated orgasm option in the SexLab MCM? Whether you modify SLSO scripts to work with PAHe or do your own "framework". Because you don't want them to come at every sex event. I don't know how SLSO works but if SLSO can sent an orgasm event to SL, DoM shoulb be able to right? And than deactivate SLSO for PAHe+ slaves. And some bug fixes because activading the seperated orgasm feature will certainly cause problems for no-SLSO users.. those peasants Maybe you can just give non PAHe+ slaves their orgasm at the end if the scene back.

 

Edit:

On 4/25/2022 at 10:31 AM, TrollAutokill said:

 - Updated fix for SLSO slave orgasm event. The event to check for orgasm chance was not called properly.

Tested SLSO event seems to work as intended now. The slaves orgasm (maximum) one time per sex Scene. Also tested what DocClox said about the masturbation function and indeed well trained slaves come every time you click the bottom which makes them come as fast as you can navigate the menu and as often as you want.

Edited April 26, 2022 by DieHans

[CliftonJD]

2 hours ago, Jasmine92 said:

So I found the problem why Kimly (and in extension NPCs with unique voices like Maven, Ulfirc or Astrid) are not included in the voicepack I created. (See over here, if you are interested) Question now is, should these unique Voices be added to the Voice pack for the next update? Or is no one enslaving NPCs like Brynjolf and Maven anyhow so it doesn't matter? Adding everyone will add another 41k fuz files to the game and I'm not sure if at some point we will hit a wall like with the maximum number of animations. (Some people are already reporting that voices just stop working mid game, but that might not be related) I will add Vex anyway, since that's Kimlys voice.

 

Voices that are missing are:

  Reveal hidden contents

Alduin
Odahviing
Paarthurnax
Fura
Valerica
Dexion
Garan
Gelebor
Harkon
Isran
Frea
Adril
Lleril
Modyn
Neloth
Storn
Astrid
Delphine
Elenwen
Karliah
Maven
MirabelleErvine
Vex
Ancano
Arngeir
DelvinMallory
Esbern
Galmar
KodlakWhitemane
MercerFrey
Nazir
Tullius
Ulfric
Hadvar
Brynjolf

 

uniques, dlc actors and unique actor mods will always run into that issue when not using the shared dialogues. can't say if its worth the effort attempting to fix either of those, maybe choose some actors you would enslave yourself and consider adding others by request

1 hour ago, Kharos said:

 

I doubt that it can easily be reproduced. I had her try to run away 5 or 6 times, and it only happened once. As I wrote, maybe a race condition, or a lost event or something else acting up.

 

I did make a savegame when she was in that bad state (faction set but not in run_away state). If you add fallback code to catch and fix the issue, e.g. in OnToldOff or anywhere else, I can use that savegame to test if your fix is working.

 

the problem with trying anything related to it is until you can reproduce cause and effect, there's no way knowing if what you're attempting to fix it will effect it. 1 example is if the runaway is a clone of a unique and you have the original or another clone of the same unique, the game will treat both as if running away until you catch the correct run away.

 

another example could be if something has your papyrus running slow, the command to stop the runaway won't be issued to the slave until the engine catches up

[DieHans]

12 hours ago, DocClox said:

In fact, the first thing I do with SLA is disable player arousal since it does nothing for me but print annoying notifications, give me a 20% malus on all xp, and make most of the skill books no longer work.

Try the new OSL Aroused it will fix your problem by having a completely overhauled and much more logical mechanic compared to SL Aroused. Its an upgraded version from the Baka version so you have no downsides, everything included.

 

12 hours ago, DocClox said:

SLSO isn't perfect, but the minigame means that the sex part of the training isn't limited to grunt-grunt-grunt-HEURRG! Next girl please! And with some work you can improve your own skills and condition the slave so that controlling (or denying) their orgasms becomes easier.

thx for pointing that out

 

It makes the game so much better. Which doesn't mean that SLSO the ultimate way to go.

Edited April 25, 2022 by DieHans

[Jasmine92]

3 hours ago, InsanityFactor said:

One thing I wonder about too is the accents. Would thick, unique accents like Delvin, Frea, and Mirabelle carry over well? Or would it fail to piece the words together phonetically? More of a question for @Jasmine92 I guess. 

The accents and pronunciation of words gets carried over relative well. It differs from voice to voice though depending on a) the amount of voicelinces actually available in the game and b) the version of the trained model. Serana for example has lots of lines in the game and the data was created with the newest AI model so her voice is quite good. Vex voicemodel on the other hand is not that good, since she has not that many lines spoken and still uses the old AI model. Brynjolf is a good example where his accent gets carried over quite well and he sounds nearly natural. Delvins "nasal"  speech type is clearly audible and even Freas "from above" / "to the point" pronunciation and melody remains. Conveniently for those of us who mainly enslave Bandits: since they use generic voices which have the most spoken lines, their AI models tend to be quite good.

 

But you can try it for yourself if you want:

PAH - Diary Of Mine v2.9.x - Voicepack v1.1.7z (Mega Link, includes all non unique voicetypes + Seranas new 2.1 model)

PAH - Diary Of Mine v2.9.x - Voicepack Addon I.7z (Mega Link, includes many unique voicetypes (see Spoiler) + Kimli (aka Vex))

Spoiler

Voicetypes in the Addon I are:

- Fura

- Garan

- Frea

- Storn

- Astrid

- Delphine

- Elenwen

- Karliah

- Maven

- Mirabelle Ervine

- Vex (Kimli)

- Brynjolf

- Delvin Mallory

- Galmar

- Mercer Frey

- Nazir

- Tullius

- Ulfric

 

Be advised: I haven't tested the addon extensively jet, only checked some random lines and a quick talk to Kimli and it seemed to work. So if something is off, please do tell.

 

Both packs are BSA packed for SE, so LE users just have to unpack them. (You can ignore the overwriting files, that is just one line per voice that got updated because of a spelling error that TrollAutokill corrected in DoM.)

 

@TrollAutokill you can add the links to the download site of the mod or even re-upload them if you want. Oh, and right now I "abuse" the auto-loading of "espfilename.bsa" and "espfilename - textures.bsa" of skyrim to let the game load the bsa files with DoM, but if one day you decide to pack DoM into a BSA, give me a heads up, because then this would break the mod.

 

 

2 hours ago, CliftonJD said:

uniques, dlc actors and unique actor mods will always run into that issue when not using the shared dialogues. can't say if its worth the effort attempting to fix either of those, maybe choose some actors you would enslave yourself and consider adding others by request

I have thought about that recently: PAHEs reuse of shared lines where applicable is quite clever and helped the mod to even have some voiced lines, even though limited since not many lines that could be used around the topic of slavery come with the base game.

Since the DoM slaves are quite talk active when training, scolding and praising them it adds a tremendous amount of immersion experience, that those lines are actually voiced through the use of xvasynth (at least for me, I was never the type that liked to read a lot of text when playing a game and always preferred to listen to the characters)

So would it be possible to abandon the "reuse shared lines" approach of PAHE and change to the DoM "just create the conversation and  let xvasynth sort out the dubbing"? And would that maybe even reduce the workload when adding new lines, since you no longer need to add it to every shared voicetype but only once for every line spoken? (I have no inside in how one actually creates a conversation in the CK so I'm totally speaking out of my ass here)

Edited April 25, 2022 by Jasmine92

[sidfu1]

osl aroused doesnt work with all mods yet. the one i have that dont work with it is  aroused nipples. due to  lag from it not working when i did intial testing wasnt able to check for other mods

there is no reason to disable aroused and xp is a dime a dozen in teh game. dont stand around fucking all day and your arousal will be fine.

[Antiope_Apollonia]

4 hours ago, InsanityFactor said:

More of a question for @Jasmine92 I guess. 

*coughs*

  

4 hours ago, InsanityFactor said:

One thing I wonder about too is the accents. Would thick, unique accents like Delvin, Frea, and Mirabelle carry over well? Or would it fail to piece the words together phonetically?

There's really no such thing as an accent.  It's just a difference of perspective.  What you call an accent is really just a deviation from the prestige standard most familiar to you.  What you describe as "thick" is really just more divergent from that familiar standard.  But someone to someone who speaks that dialect, it follows just as readily discernible phonological rules as your dialect does, and what sounds "normal" to you sounds "thickly accented" to such a person.  

 

But the AI doesn't have any such biases.  It's just a matter of sufficient data input.  Given equally large samples to train the AI on, it will learn to replicate one accent pretty much equally as well as the next.  That's not quite strictly true—there are some features that can make one language or dialect more difficult than another—but especially when we're simply talking about different dialects/idiolects of the same language, the differences are essentially negligible.  

 

The challenge with certain voice types would simply be a paucity of data to train the AI on.  The more input, the better the results will be.  Even if the initial results are worse, though, a lot of that could likely be corrected for with good IPA(/ARPAbet) transcriptions.

Edited April 25, 2022 by Antiope_Apollonia

[InsanityFactor]

56 minutes ago, Antiope_Apollonia said:

*coughs*

  

There's really no such thing as an accent.  It's just a difference of perspective.  What you call an accent is really just a deviation from the prestige standard most familiar to you.  What you describe as "thick" is really just more divergent from that familiar standard.  But someone to someone who speaks that dialect, it follows just as readily discernible phonological rules as your dialect does, and what sounds "normal" to you sounds "thickly accented" to such a person.  

 

But the AI doesn't have any such biases.  It's just a matter of sufficient data input.  Given equally large samples to train the AI on, it will learn to replicate one accent pretty much equally as well as the next.  That's not quite strictly true—there are some features that can make one language or dialect more difficult than another—but especially when we're simply talking about different dialects/idiolects of the same language, the differences are essentially negligible.  

 

The challenge with certain voice types would simply be a paucity of data to train the AI on.  The more input, the better the results will be.  Even if the initial results are worse, though, a lot of that could likely be corrected for with good IPA(/ARPAbet) transcriptions.

Right, of course. I guess what I was trying to get at was the "data input" as you put it.

 

Which I think I misunderstood how the voice synthesizer thing works. I was thinking it would work like a text-to-speech type thing, but it sounds like it's more advanced than I was giving it credit for. So like it would take the word "far" and look inside the NPC's dialogue to see how it makes an "f" sound and an "ar" sounds, and then smash those together, rather than trying to make Delvin's voice say "far" with a hard American r sound, for example. Or even better it finds the word "far" and just uses that? Is that how it works? 

[Jasmine92]

6 hours ago, InsanityFactor said:

Right, of course. I guess what I was trying to get at was the "data input" as you put it.

 

Which I think I misunderstood how the voice synthesizer thing works. I was thinking it would work like a text-to-speech type thing, but it sounds like it's more advanced than I was giving it credit for. So like it would take the word "far" and look inside the NPC's dialogue to see how it makes an "f" sound and an "ar" sounds, and then smash those together, rather than trying to make Delvin's voice say "far" with a hard American r sound, for example. Or even better it finds the word "far" and just uses that? Is that how it works? 

I will try to explain a little from the IT point of view and what i understand of the system used, however I'm no expert for speech and all the things concerning the human language like @Antiope_Apollonia, all my expertise in that field only comes from using speech myself. So take that side of my explanation with a grain of salt:

 

It is really more like the former. The voicemodel, when trained, has no concept of language, it doesn't even now the difference between vowel and a consonant. All it learns is, like you said, which kind of letter combination leads to which kind of sound. Which is the beauty of an AI model really, since you don't have to create all the rules yourself which tell the system "this is a t, and it sounds like this, but when its followed by an h it sounds like this" and all that, which, depending on the problem that needs to be solved, could become quite unfeasible really fast. So what you do is, you take all the data you have, in this case the spoken voicelines, pair them with a good transcription and just throw them at the AI model to learn. And depending on the quality of the voicefiles, the transcription and the number of data points you have, it can get you quite far.

 

However as you try to raise the quality, you soon realize that the beauty of an AI is also its greatest weakness, since it does know nothing about the world surrounding its system like emotions or all the other things humans convey with their speech (questions sound different than exclamations, a fearful response is different than a strong one). So what you do is, you try to break down the problem even more and give the AI more information about the datapoints it learns from.

In the case of the 2.0 voicemodels they took the transcription and translated it to a phonetic one (is this case arpabet) because with that, you can convey the differences between different sounding versions of the same letter combinations (the u in butt is different from the u in dude and live and live are two different things) And that leads to quite a bit of quality improvement, but unfortunately can not really convey emotions either.

I don't know if there is a phonetic transcription code, that although includes all the different things like melody and emotion (maybe @Antiope_Apollonia knows more), but from what I understand, that would be the next step, to train the AI all the different ways one could say the world "stop" (pleading, demanding...). A little bit of that can already be done because the AI understands that words that are followed by a question mark a pronounced different from ones followed by a exclamation mark but that really only encompasses the one word at the moment.

 

So I just realized, I ended up writing quite a lot, but I hope it will help to explain, how xvasynth and the voicemodels works.

 

TLDR: Its not that complicated. Not at first, at least.

[anjuchan]

I finally shift to SE version, and translated all of these mod. but I have a few problems.

 

I enslaved a bandit, and this mod works perfectly. but after save and load, dialogue does not working.

 

is this a script problem? or caused by a translation stuff (it edits .esp.)?  sorry for my skeever brain

Edited April 26, 2022 by anjuchan

[sidfu1]

troll i found some unique npc to check out

they are rare event npc that you will hardly ever see. no idea if they have dialogue blockers or such as if you enslave them and dont clone them then after they enslaved you cant talk to them once collared they just keep the base game dialogue. if you clone they be fine. thought you might want to check it just in case its a cause for a future bug  with any other npc.

 

[Antiope_Apollonia]

4 hours ago, Jasmine92 said:

I don't know if there is a phonetic transcription code, that although includes all the different things like melody and emotion (maybe @Antiope_Apollonia knows more), but from what I understand, that would be the next step, to train the AI all the different ways one could say the world "stop" (pleading, demanding...).

"Emotion" isn't really a feature of language.  It's something that listeners interpret from cultural conventions about how certain features of language are used.  What we interpret as emotion is really just particular variations in volume, prosodic features like pitch or the length of certain sounds, and phonations like creaky voice or breathy voice, etc.  For example, if you increase the volume, raise the pitch, and shorten the vowel length, you'll make an English speaker sound more panicked.  Or if you want an English speaker to sound like they're pleading, the main thing is to increase vowel length, as well as to modify the intonation.  English speakers often use creaky voice to convey sympathy.

 

From my limited Wikipedia knowledge of ARPAbet, it looks like a much less powerful tool than the IPA that it mostly mimics, lacking ways to represent a lot of the subtler features that IPA handles with diacritics, so I'm not sure if it's possible to transcribe all of these features in a way that xVASynth can read them, but you can represent all of that sort of information in IPA.  Some of the phonation stuff is still very much an ongoing area of research, though, so there isn't always universal agreement on exactly what qualifies, which makes using some of those diacritic representations a little more problematic than the core phonemic elements of the IPA.  

 

If you're interested in this stuff, this article might tickle your pickle.

 

As for machine learning handling this stuff, as always, it really just comes down to the input.  If we had 100,000 unique lines spoken by each voice actor, you could train the algorithm to perceive all kinds of subtleties without needing to fine tune things by hand.  Since Skyrim already pairs dialogues with emotions, that could be part of the training data, too.  The AI would be able to extract patterns from how the actor reads certain words in lines flagged "Neutral" vs. those same words in lines flagged "Fear" and so on.  I rather doubt we have enough input data available for that level of sophistication, though.

 

That said, I think in principle it should be possible for the AI to learn some things as generalisations not tied to specific voice types, and that might help.  I have no idea whether the current setup does this, but conceivably, it could compare dialogues flagged for different emotions, and then pool those insights to learn general insights about the features of language that correspond to different emotions.  Then it could be able to apply what it has learnt about how to make a speaker sound afraid even to a voice type that doesn't have any voice acted dialogues with the "Fear" flag.  Whether the current generation of the software is able to do that, I have no idea, but in principle, it can be done.

Edited April 26, 2022 by Antiope_Apollonia

[DieHans]

13 hours ago, sidfu1 said:

osl aroused doesnt work with all mods yet. the one i have that dont work with it is  aroused nipples. due to  lag from it not working when i did intial testing wasnt able to check for other mods

there is no reason to disable aroused and xp is a dime a dozen in teh game. dont stand around fucking all day and your arousal will be fine.

Aroused nipples? Never heard of it. Why don't you use ABBA, can be customly adjusted to work on every morph slider of every body there is kinda. I use it with newly added sliders of 3BAv2. Works splendid. And when a new update comes out like 3BAv3 you can just add the new slider in the config. If they are even interesting for arousal.

As I know OSL Aroused is 100% backwards compatible.

[TrollAutokill]

43 minutes ago, sidfu1 said:

troll i found some unique npc to check out

they are rare event npc that you will hardly ever see. no idea if they have dialogue blockers or such as if you enslave them and dont clone them then after they enslaved you cant talk to them once collared they just keep the base game dialogue. if you clone they be fine. thought you might want to check it just in case its a cause for a future bug  with any other npc.

 

For this kind of NPC there is also the possibility to end their quest by brute force before enslaving. You need to open the CK, find the NPC, use information to find the corresponding quest, open the quest and write down the end stage number. Then load the game, go to the NPC, in the console use "setstage <name-of-the-quest> <end-stage-number>". It might free them from the quest controlling their behaviour. Then enslave them.

 

Or you can try to enslave them at the end of their quest. For Salonia and her husband that is when they are dressed in rags and complaining about Skyrim after you ruined the wedding with the DB quest to kill Vittoria.

 

If those solutions don't work and cleaning their factions doesn't help neither, then I am afraid there is no other way but to clone them. If NPCs are part of a quest, and this quest is active, controlling their dialogue options and changing their package, PAHE mechanism will be broken.

 

Edited April 26, 2022 by TrollAutokill

[TrollAutokill]

Version 2.9.7 is out, mostly bug fixes. Slowly moving toward a stable version.

 - Added fighting class bonus to personality traits

 - Kimli enslavement is forced to no cloning to avoid duplicates.

 - Running away faction is cleared a second time

 - ClearExpression now checks for Is3dLoaded to avoid script errors

 - Slavers will stop punishing their slaves if 100% trained.

 - HSH SLA script skip just captured slaves to avoid script errors.

 - Sex doesn't register as a punishment if not aggressive.

 - Comfort sex should not register as rape, even if always aggressive is set in PAHE. As a consequence it doesn't register as a punishment neither.

 - Fixed a few typos

Edited April 26, 2022 by TrollAutokill

[DocClox]

15 hours ago, sidfu1 said:

dont stand around fucking all day and your arousal will be fine.

 

I tend to spend a lot of time leading several naked slaves across the countryside. That can take a toll on your arousal quite quickly.

[pinky6225]

1 hour ago, TrollAutokill said:

Version 2.9.7 is out, mostly bug fixes. Slowly moving toward a stable version.

 - Added fighting class bonus to personality traits

 - Kimli enslavement is forced to no cloning to avoid duplicates.

 - Running away faction is cleared a second time

 - ClearExpression now checks for Is3dLoaded to avoid script errors

 - Slavers will stop punishing their slaves if 100% trained.

 - HSH SLA script skip just captured slaves to avoid script errors.

 - Sex doesn't register as a punishment if not aggressive.

 - Comfort sex should not register as rape, even if always aggressive is set in PAHE. As a consequence it doesn't register as a punishment neither.

 - Fixed a few typos

 

lol that was quick, only just put 2.9.5 in

 

Will you be adding to the "how are you feeling" dialogue response for whether a slave is maid or combat trained? not sure what you'd put for the maid one but the combat one could have varying enthusiasm for fighting as you go from 0 to the cap of 127 (i.e. 0 - i'll fight if i have to , 127 - I'll kill anyone you desire)

[CliftonJD]

7 hours ago, anjuchan said:

I finally shift to SE version, and translated all of these mod. but I have a few problems.

 

I enslaved a bandit, and this mod works perfectly. but after save and load, dialogue does not working.

 

is this a script problem? or caused by a translation stuff (it edits .esp.)?  sorry for my skeever brain

sorry, only way to know that for sure is to try the mod in it original form(without the translation)

[pinky6225]

New question, i tried using the restore persona spell on my long term slave as i wanted to see how her stats had changed (think sensual and submissive has gone up) but she's now resisting the spell, does that mean her stats are at default?

 

Seemed bit odd to have 3 stats at 100 exactly but was only reason i could think of she'd resist the spell

[CliftonJD]

17 hours ago, Jasmine92 said:

I have thought about that recently: PAHEs reuse of shared lines where applicable is quite clever and helped the mod to even have some voiced lines, even though limited since not many lines that could be used around the topic of slavery come with the base game.

Since the DoM slaves are quite talk active when training, scolding and praising them it adds a tremendous amount of immersion experience, that those lines are actually voiced through the use of xvasynth (at least for me, I was never the type that liked to read a lot of text when playing a game and always preferred to listen to the characters)

So would it be possible to abandon the "reuse shared lines" approach of PAHE and change to the DoM "just create the conversation and  let xvasynth sort out the dubbing"? And would that maybe even reduce the workload when adding new lines, since you no longer need to add it to every shared voicetype but only once for every line spoken? (I have no inside in how one actually creates a conversation in the CK so I'm totally speaking out of my ass here)

pah slaves have the base game shared voice lines where applicable and layam's shared voice lines for some of the other dialogues. base game dialogues have the benefit of being already voiced, but at the effect of using the wrong voice type on most slaves since other voice lines will be eventoned.

 

layam's method works great for the dialogues it applies to, but when used in the newer voice types added by pahe, those lines won't use the correct voice type either. however i think i've noticed some changes when using the xva synthpack made for pahe, so its possible those voice lines have been corrected by xvasynth

 

layam's method could be described as adding it to every shared voice type as you say, but that work load has already been taken care of unless or until i decide to make another voice type available to pahe. if or when the time comes to expand pahe to another voice type, it will be easier to replicate what's already done for other voices and modify it to the new voice type...once that's done it will be upto the xva pack author to adjust the voice lines to match the new voice type(sorry don't remember if that was also you). creating new lines from scratch is when it gets to be more difficult, but if you have any suggestions for xva to improve pahe, send any you find to a message