Sign In

Eldritch Analog Photo | for Flux

28
206
41
7
Updated: Jun 2, 2025
styleanalog photo
Verified:
SafeTensor
Type
LoRA
Stats
104
29
107
Reviews
Published
Jun 1, 2025
Base Model
Flux.1 D
Usage Tips
Strength: 1
Trigger Words
analog photo
Hash
AutoV2
C60AF7EA8F
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Updated to a version 1.2 with better skin textures. Usually avoids that default Flux plasticky skin look.

My preference is to sandwich prompts between two parts:
analog photo of [your subject matter]. Scratched glossy photograph from a flatbed scan.

But you can eliminate or modify the second part to strengthen or minimize the distressed effect.

Most obviously, the model adds scratches and dust marks as though the image were an older scanned photo print.

There are a good number of analog film style models so why use this one? Three reasons:

  1. This model doesn't presume that analog film means garbage photography. With Eldritch Analog Photo, the color is balanced and contrast levels on point, the bokeh pleasing, and photography altogether looks great. It's just been scratched and dusted up before being scanned.

  2. This model doesn't presume you want to photograph only subjects from the 1980s, 1990s or earlier. This model remains flexible enough to generate historical, fantasy, futuristic sci-fi, or contemporary subject matter with the same aesthetic.

  3. This model doesn't rely on noise and film grain to convey the (quality) characteristics of film photography. Images are largely noiseless, as though shot on ISO 100 film by a photographer who knows what they're doing.

It was only trained on SFW subject material with the base Flux model, so if you're hoping to produce nudity or sexually explicit content, you either need to pair this with another LoRA or base model that can accomplish that. I can't vouch for how well they will work together.