IllustratorsLeak
Athreos
Athreos

patreon


Visual enhancements, rambling and goals

I haven't posted any in-between-videos posts in a while. One reason is that I was away, but another reason is that when I have the time, I work on learning and improving the quality of the scripts, audio, images, visuals, and composition. I didn't want to share things that weren't going to be, because to be honest, I've been consistently failing to break through for the past I don't know how many weeks.

I think I've come up with something different that I quite like, and I'd like to share it with you and get your feedback.

I am somewhat happy with the quality of the images that I am using, but that's it, it goes up to "somewhat"; this is my attempt at improving the visuals in 3 main things.

First, I feel like I haven't found the style I want the visuals to be yet, but this is the closest. I don't like only the flat anime color as it feels too 2D, too flat, for example: (although this image is better than the "improved" version imo, experimentation will tell in time)


I also don't like the more realistic anime version as it feels really "oily" to me, for example:


So I am on the never-ending hunt for trying out different things to find the golden middle ground between the feeling of realism and the aesthetics of anime, and I'd say this is the closest to it so far. Hopefully, I can improve it further in time. (Attached images at the top)

The style is somewhat inconsistent between different prompts, so I guess an added benefit will be some variance in the aesthetics of the images. Hope you like it, and I'll be glad to hear your input.

The other few things that will get consistent improvement are eyes, they're regularly kind of unfocused and not drawn enough.

And finally, lightning, I think this I suffered with the most, the previous images are extremely bright, one of the goals was to bring this brightness down and have better lighting and shadows that will bring better aesthetics. For example, the image below is "too orange" for me. Still, it's quite good, as it's one of the earlier versions of the new style.


I am also working on better audio effects. I have not finalized it, but I hope to start adding better trigger sound effects - stronger or weaker voice effects, subliminals, binaurals, and some brainwave techniques to enhance trance states, etc., after at most 1 or 2 videos.

Another long-term goal is to improve image composition - I want to create images that complement the content in the headphones, aligning audio and video. This is my biggest goal when it comes to visuals, but it's extremely challenging to achieve. However, I'll be working on it. One of the main issues is that it's extremely challenging to produce femdom content. Once you add "BDSM, bondage, ropes", there's a female subject that gets bound, and that's it. You can't simply run away from it; it's biased towards drawing pretty women. And of course, this I have to automate, I can't prompt 1,2,3 hour script sentence by sentence.

Another goal would be to introduce a new voice, I will probably also be able to generate a single audio with multiple voices which will bring things to the next level, but I'll have to chip away at it as well, it's not a 1 day thing and I want to have it make sense with the universe we've created so far.


There's also the never-ending goal of improving and writing better scripts, which is the hardest one because it's more challenging to establish a good feedback loop on the scripts. I generate an image, I see it, and I like it, but the hypnosis script itself is not that simple; you have to be in the right mindset for being hypnotized, step out of the writer's hat and get into the listener's hat, see where the flow breaks, etc., which is a lot harder and time-consuming.

Although it seems I am not around, I am, I am just trying to bring the quality up, I don't enjoy doing the same thing over and over so I use each audio + video as a stepping stone to improve something, with time that will slow down, but I see uncharted ground that will bring up the quality and reduce the workload for me at the same time. I am sure you enjoyed my ramblings. I have not planned the next video yet. I am working on improving, well, everything. When I get the inspiration for the next part, I'll jump into it.


Meanwhile if you have suggestions where the story should go, feel free to ping me in private chat, I am open to ideas, doesn't mean I'll do it, but it will be interesting to discuss, if you're into shaping the experiences of other's that's one way to do it.

P.S. Here's a last blowjob with a little ball squeeze for reading all this.


Visual enhancements, rambling and goals Visual enhancements, rambling and goals Visual enhancements, rambling and goals Visual enhancements, rambling and goals Visual enhancements, rambling and goals

Comments

Not that you asked, but it might be interesting for someone - I was using TensorArt until now, as it seems good quality for price, but it's so clunky, I had to write my own scraper to download the images I generate, for example the batch download button requires you to manually click "save" for each image (disaster), and I had to wait for queues, copy and paste prompts 1 by 1 and click generate literally a thousand times per video, which takes like literary 2 days x 6-7 hours. This new art is now running locally with ComfyUI (+ loras, as you said) but I am renting GPU because my machine will just breath out it's last breath if I run this, so yeah, it becomes much faster, better quality, and infinitely more customizable. Now the bottleneck becomes good prompts, which is the next challenge. I'll figure the rest out with time, what's useful for me as a creator is when you share what you liked, so feel free and even empowered to go out of your way and say "this part of that video", or "this script", or "that trigger",

Sect

Interesting to see how far AI art has come lately. Do you use a service for generating or your own GPU? If you run Stable diffusion you can fine tune and change your style to anything with a lora. Although I haven't watched every video IMO the art looks really good so I wouldn't worry too much about it. AI has advanced enough that with enough time/research you should be able to get any art style/content you want without being a computer genius. I am not that experienced in general hypnosis content but my favourite joi creator of all time is BlackTantra and I see similarities between your videos, which is what lead me here in the first place (but I must say your voiceover is much better). Except while BlackTantra has a following from their niche (clench/hfo/breathplay joi) you have your own unique content. You clearly care about the quality of your work, otherwise you would have made 20 generic joi's with ai art and voice by now (aka joi database). So I am looking forward to your vision of your content.

Amogoos


More Creators