Adventure!
I’ve been on an adventure these past couple days. And not the fun kind. I’ve been on the hunt for the elusive PERFECT workflow. Basically, I’m not happy with the current workflow I use to make my Daz Renders look real. I know I can get better. It’s 2026. So I’ve been studying and testing and trying to find it. I think I’ve finally come across the best I’m going to get for now.
nothing will be absolutely realistic and retain perfect consistency. And consistency is more important to me than hyper real humans.
Some places won’t even allow hyper realistic human porn games.
Text on shirts, tattoos, background objects and time of day lighting matter more than realism because if you’re clicking through images and the light is constantly changing on you, it’s going to be obnoxious.
So, I’ve come to a middle ground. Better than what I had before by a long shot. No photoshop cleanup required which will save me days of work each month. They look more realistic and they match the character face better.
The trick is to run PULID, then a character specific LORA I made, then a Face Detailer all through a mask on the specific character in the image. This gives the best results possible for consistency.
Sure I could make each character look hyper real but:
That would cost money per render. Tens of thousands of renders
Eye direction and expressions would be random.
Lighting would be different each image
I’d have to write a prompt for every image
Object consistency would be a mess.
Comments (4)
Have to agree with Bobo, For me in order of importance kinks, plot, and then graphics.
Yup. I understand. But to me, all of those are important. :) It bothers me to see drifting faces. Or faces in the background that don't look great. Example... I just improved the whole games UI for the same reason. I didn't need to. But when the game goes 1.0 live, it needs to be perfect. Not just another Visual Novel in a sea of visual novels. Anyone can make a quick visual novel that looks okay and has an okay story and okay animations. And there are a million of them. I don't want to be in that group.
I like your creative process, but in my opinion, you're too hard on yourself. As for the plot, I really like it, and I was also thrilled with the original graphics. For me, and I think for many others, we prefer your ideas to perfect graphics that might improve over time and perhaps aren't as important.
The thing is. I have to run my AI workflow regardless. If I can get that workflow updated, you get better images, and I don't have any extra work. EXCEPT for CREATING the workflow. Getting that right is a pain, and takes time. But I'm getting there.