It would also be nice to remove forced masturbation and orgasm at maximum arousal, or at least limit it to once every 3-5 days. If arousal is high, simply apply a debuff to the amount of resources or stamina restored; the higher the arousal, the higher the debuff. Ideally, the player should be able to control this themselves, for example, by adding a masturbation button: while relaxing at home, at work, or outside in a secluded corner, etc.
- Aug 22, 2019
- 142
- 347
- 157
That's a good point. What I see now is that acquiring arousal is quite hard because there aren't many ways to do it and the ones that exist are random (launched via Streets or Garden with Lea), which gives the user no control in increasing arousal therefore making it an important stat to manage/maintain.
My goal was that arousal wasn't such a hard stat to play with since the number of lewd activities that the elves witness increase as the game goes by, also based on the story.
I'm trying to find some balance there... Either add the "relieve" button action or reduce de grindiness to acquire arousal.
I had a LOT of content to the game since I have made many variations of the same lewd positions, but I am now combining those similar 5-6 second videos into longer ones. So, the amount of videos available is going down although the total nsfw scene time will remain.
I don't understand why you need to combine videos into longer ones. Personally, as a player, I don't like this approach. There are at least two reasons, and I consider them quite important. First, if there's text, it's difficult to read the text and watch the video at the same time. Second, the animations are truly beautiful, and some of them are worth examining more closely and for longer. I think it's better to transfer this point to the code. The game engine is quite flexible. It's best to loop short videos, write a little text in them (dialogue or character thoughts), and then don't have to loop the final video (for example, when they cum). The only catch here is that you'll most likely need to capture frames in the video and even trim it to create a smooth transition from the end of the video to its beginning when looping. I know many games with animations less than 2 seconds long, and they look good; there are many videos that run sequentially.
Last edited:
Game balance is always difficult... I think almost no game has perfect balance. It's all about trial and error, feedback... I think adding a "relief" button is necessary, as it will make the player more engaged and give them more control over the situation. Since as the game develops, there will be more events from which arousal will flow, I suggest reducing the arousal boost per event from 30 to 10, or better yet, to 5 or 3. If you really want to go the extra mile, you can make it different for each character based on their fetishes. Say one likes one thing, another likes another, and give them different amounts of arousal depending on what happens to them. In the early stages of the game, you could add a button, like a cheat, that would give arousal just by pressing it. As mentioned above, you can also add the debauchery stat. It allows for more development options. Do this gradually: initially, when debauchery is at 0, they become very aroused by any "incidents," but there are fewer of them. Later, when debauchery increases, there are more "incidents," but less arousal.
- Aug 22, 2019
- 142
- 347
- 157
Any model in particular? I've been trying to make AI Videos but it's not working out for me
- Aug 22, 2019
- 142
- 347
- 157
NSFW mostly WAN2.2,
SFW mainly the big ones not open source, although LTX-2 just came out yesterday and it's quite impressive, but does not do NSFW... yet.
Out of curiousity: Have you tried LTX-2 already - didn't get around to it yet but I heard you need quite a hefty amount of RAM for it to run properly?
Also...to stay on topic: Any news, eta-wise, on the next update?
I agree with this one. Besides the mentioned points I can imagine, that it is more time consuming to combine several smaller videos together and that it takes time to generate the "right" videos that fit together without making it look akward. The Anneli doggy-style scene that was posted on patreon a few days ago for example looks kind of akward in the second half. At least to me the first ~6 seconds look like a nice animation, but the other half looks weirdly sped up with some akward movements on the transition. In game i would really like it more, if this animation were two seperate looping parts.
Which brings me to the second point that was mentioned, about wanting to see a specific animation for longer. I can imagine, that there are people who prefer a certain sex pose in a scene and wouldnt mind if they can let it loop, or have the option to loop that part in the gallery, but if its just a small part of a longer video, then the player cant do that, as mentioned with the Anneli video. Another example would be the longer videos on patreon. The longer videos that were posted on patreon were good, but i wouldnt watch the whole thing on a loop. I would much rather have them as a short story with short looping videos in the game, so that i could watch specific parts of it in the gallery.
So yeah, since im not a patreon and cant see how the poll about the video length turned out i still wanted to give my oppinion about that here dj03man.
Last edited:
- Aug 22, 2019
- 142
- 347
- 157
Thanks for the feedback, the votes in patron were overwhelming for the combined videos (90+%) but the shared videos were also better glued together than the one with Anneli.
Indeed the example used for Anneli wasn't the best for a combined video as the position is not exactly the same. The idea for the longer videos isn't for 2-3 mins videos, but for 12-15second animation of the same position.
- Aug 22, 2019
- 142
- 347
- 157
I have tried it and it's great for sfw with voice. Also does decent job at lipsync so I'll likely use it in the future for that. Although WAN is better for that, ltx-2 is several times faster. Characters with a voice and expressions add huge immersion to the game. I can run it properly as I am using a 5090 with 96GB ddr5, but not the full model properly yet.
Eta wise I'd like to have a version during the weekend, but things are going slow last two weeks. Today, I was able to finish two full scenes for the game, but before that, quite slow. The good thing is that all reported bugs and most UI issues and negative feedback was already addressed.
I don't know how well-thought-out the future plot is at this point. But I have another idea for a more logical way to play out the introduction of the combat system and training. As I understand it, the queen "falls," and (whether she was forced or willingly, it doesn't matter) she orders all castle employees to undergo "special" training/exercise. Next, it would be a good idea to slightly rework the scene in the library with Lea, when she begins training... It would be better to make some animation of her awkward seductive dance or something similar; this would certainly not suit our green friends, and they will take over her training. Anneli's main mission is to search for information. She can't get much out of simply working at the bar, which causes the other girls to press her a bit to try harder. So she goes to the bar master to ask him to teach her how to serve customers better, and then I begin training her. It would be better to slightly expand the training so that they don't become pros in everything in a couple of days. This way, you can make the combat more interesting. For winning, or rather, for satisfying your opponent, you can gain valuable information; when the goblin finishes, he'll spill it all. For example, in one battle, satisfying a goblin once will earn you the usual gossip, twice more interesting information, and so on. It's better to include plot information here, then everything will "stick" together. The main thing is to distribute the arousal of both the goblins and the heroines in the battle. If the heroine's arousal reaches its maximum, she loses, while the goblins can get their arousal up to 3-5 times in a single battle. Again, you can play around with preferences and poses for both the goblins and heroines, and award different amounts of arousal per action. And most importantly, as I mentioned earlier, add goblin "attacks."
- Aug 22, 2019
- 142
- 347
- 157
The main story topics are written for 3 chapters, main story events and characters.
For the Library I initially thought of a "translation" minigame where the goblin gives Lea the task to translate runes/glyphs and correct or false translations would trigger certain scenes, but that would take ages to implement and one more mechanic to the game, so I forgot that idea.
Regardless of that the Library is currently under full redo, where every training step will be a corruptions scenes with some lore to it, with just "proceed/cancel" choices where "proceed" requires some minimum stats (arousal, etc). When the event proceeds to the end, Lea unlocks new recipes for Anneli brewing or progress paths for Valya's meditation. Next visit to the Library starts a new corruption scene, in the end unlocking all skills for her. So, in the end there will be a full story between Lea and Gumul and how they helped each other against their will, where she got the info required to potentially or not help her city and where he potentially or not got his way with her while getting to his goals of elf dominion. A bit more of a VN approach where the player is a spectator, I think.
For Anneli, you're quite close to my plan. Anneli will need something from the bar master as part of the main story and the trade for that is - indeed - learning her SKILLS. This is key for last chapter.
Information / Intel is supposed to be a shared value that goes in steps. So, the girls can gather intel from 0-100, then at 100 a new step in the story emerges and new scenes and quests are given, this way the player is forced to keep acquiring information and intel throughout the game.
For combat, winning the game is one of the actions planned to increase intel, exactly as you mention.
Where I am struggling is in the actual combat game... currently only available for Lea's, but the goal is to have it triggered for all (Valya in the castle and Anneli in the bar). My concept is a game that re-uses the animations and gives the player power to chose which, but I haven't found a decent way to implement it without being too "rapy"... that's the reason why the minigame is untouched since v0.2. Lots of players liked the idea of having a lewd "combat" minigame, but they complain that the implementation is shit (and I agree) and that creates some division. I don't know if I should create a new concept or clean&consolidate the current concept, like "clean the shit". Community feedback hasn't been clear on this also.
Finish the meditation route. Just follow the yellow path.
yawt
Newbie
- Aug 29, 2019
- 47
- 123
- 162
- Aug 22, 2019
- 142
- 347
- 157
I rented between 4090 and 5090 on Runpod until last week, then last week I bought a second hand pc with those specs, supporters helped a lot <3do you have that monster locally ??
next scene : Anneli raped by a RTX 5090ti
With the DRAM prices going up, a lot of gamers decided to sell their high end computers since 5090 plus anything higher than 32GB is kind of overkill for any games out there. So, I searched for months to find someone interested in selling at a decent price. I drove 6 hours and spent a full day out of the house in snowy conditions to get it, but so far it was worth it as it's faster to run it locally than on cloud, everything from test, trial and error, etc... This is one of the reasons why Escaping the Goblins development was slow, it took me a few days to have the full system running (installing new disks, ubuntu from scratch, separating data disks from OS disks, nvde for speed, swaps, moving ~1TB of data from content, workflows, models and loras, setting up the 50 nvidia series is harder to setup than the 30-40).
Then I just tested some generations in HD to see it all working well and after that I haven't spend actually time generating anything new with it yet (only tested LTX-2 for lip sync and SFW body animations for some game scenes), this week I've been focusing on adding already generated content to the game whenever I have time.
- Aug 22, 2019
- 142
- 347
- 157
There are no proper local LLMs for code assistance yet, at least not without extra context or trainning (as far as I know). DeepSeek R1 is great but heavy on DRAM.It's great to hear that you got the local equipment to produce more content !!
With this kind of setup you can probably have a (NSFW) LLM analyze your code and fix it automagically
The best for that - I found - is Claude or Gemini via API and I have been doing that already for some time, but it doesn't fix the code correctly always and even for renpy I find that it creates new code to do things that Renpy already does as base... which is stupid because it makes it harder to debug. Anyways... enough ranting and back to Valya's scenes.