Simp Chat Hologana - Discussion Thread

Thanks a lot. I really appreciate this, but could you tell me your exact settings or give me exact instructions on how you achieved this? I would like to try it on my own.
 
Hi there, I don't know if it's ok to flood the discussion thread with a bunch of screenshots so for now I'm only going to show the Curve graph and then you can punch in the numbers in the following order.

The way I'm doing it with this new method is a "two-step preset layering process". Preset 1 is the Curve graph. In the Lights section there will be a "Curve" setting on the right side. Copy the curves and save it as Preset 1. Then when you paste it on a new raw photo, you will have the option to set intensity level. That is where you can choose between 50% to 70% if you don't want to change the original color that much (the image being used in the graph is set to 100% intensity because this is for you to make your own preset first).
Once you have Preset 1, Preset 2 focuses on adding Light, Effects, Details.
This is where you have to follow things in the right order.
Lights:
Contrast -25
Highlights -25
Shadows +25
Whites +25 ( hold down two fingers when sliding to see the bright point of the photo )
Blacks -25 ( same thing but for dark point )
Effects:
Texture +90
Clarity 0
Dehaze +3 ( too high will turn the skin orange even if you see more vein definition )
Vignette is more of a personal taste. I like to use -15 with feathers on +90
It blurs the background enough so our main object stands out more. Which is nice if you're like me who prefers to crop an image to match phone screen display. Seeing it fill up my entire phone screen is so hot.
Last one for this section is Grain, it helps add a bit of noise to simulate skin texture but too much can make it look gritty. So this is what I use:
Amount +3
Size +20
Roughness +45
Details:
Noise Reduction first,
Amount +25
Details +50 (default)
Contrast 0
Next we move to Sharpening,
Amount +25
Radius +0.90
Masking +90 (very important to hold down two fingers when sliding so that it only sharpens the right parts like eyes, mouth, hair and not the entire photo)
Detail +90
Now save that as Preset 2.
Pick a new photo and we can start putting it all together. Paste the preset in this order-
Preset 1: 50% intensity if you want the same result as the Cammy set. It's a balance between toning down the warm color but not washing out the original as much.
Confirm the check mark and paste
Preset 2 : 70% is usually the sweet spot to not over do the sharpening and texture enhancing.
The key to this layering process is in the Curve Graph. It changes depending on the skin types. This one is for fair skin tone and I have a separate one for Asian skin.
Forgot to mention that it's all done on my phone using Lightroom Mobile free version, no AI third party upscale.
 
Any way we can have all the sets unzipped on bunkr or Gofilem don't have enough space to download zips
 
Metadata consists of certain information contained within an image. For example, all images in Hologana sets include details about the camera used to take the photo and the software used to edit it. This data causes a 3MB image to end up between 10MB and 20MB in size.
With Metadata this Image has 20MB

without 2.91MB
the quality in the image is the same, but the size in HD or SSD is like 10% of the original...
 

Does anyone have Hologana Ciri set from september 2023?
It seems that in this set we were the closest to seeing her areolas!
 
Being honest, you aren't just stripping metadata there. You are also reducing (probably unknowingly) the quality. The important point in that is that it's a drop in quality that our eyes are unable to tell.
To be exact, you're doing the equivalent to jpegoptim --strip-all --max=85 --all-progressive .
Just saying this to clarify because I see some misunderstandings. Removing metadata will never land these results, but the optimisation you got is great. 85% quality for jpeg is a golden standard.
Code: 
❯ ls -l --no-user
.rw-rwxr--@ 16M 7 Feb 15:06  Burnice_25_original.jpg
.rw-r--r-- 3.2M 7 Feb 15:06  'firegirl (25)_HologanaSimp.jpg'
drwxr-xr-x - 4 Aug 14:23  stripped/
❯ jpegoptim --strip-all --max=85 --all-progressive Burnice_25_original.jpg --dest ~/Downloads/burnice/stripped/
Burnice_25_original.jpg 6240x4160 24bit N Exif,IPTC,XMP,Adobe [OK] 16381501 --> 3225137 bytes (80.31%), optimized.
❯ mv stripped/Burnice_25_original.jpg Burnice_25_stripped.jpg
❯ ls -l --no-user
.rw-rwxr--@ 16M 7 Feb 15:06  Burnice_25_original.jpg
.rw-rwxr-- 3.2M 4 Aug 14:28  Burnice_25_stripped.jpg
.rw-r--r-- 3.2M 7 Feb 15:06  'firegirl (25)_HologanaSimp.jpg'
drwxr-xr-x - 4 Aug 14:28  stripped/
 
it's easier to say no metadata but yes 85% to maintain the Quality
 
That's why original, source files are still the gold (and more importantly, preferred) standard...
It would be lovely if people would quit trying to improve upon source and let everyone else handle it in their preferred manner.
Some people want to minimize the file size as much as possible. Some want to keep everything in a lossless file format, such as PNG. Some want to play with the exposure, lighting and levels to adjust the intonation of the photo. Some are interested in photography and might actually want to see the metadata.
 
Though I agree, the same level of freedom should be extended to those who choose to share files however they like or however it's convenient for them.
In case anyone is wondering, I commented what I commented because I wouldn't like people to leave this forum thinking they can shrink their collections by a huge amount just by removing metadata.
 
I don't disagree with that in theory... however, it's just as easy to upload the source files to any of the many hosting sites available, such as GoonBox right here on the site to share with others. I'm not saying they need to host it on their personal cloud backups, using their data allowance. Just share source and let others mirror. Then convert and do whatever you want with the set for your personal collection.
I was glad you shared what you did - I started writing something similar and then decided against it... because, you're absolutely right and I agree that it's a little disingenuous to spread the notion that "metadata" is responsible for something akin to 80% of a photo's file size.
Metadata is just text and binary data, for the most part... it's generally less than 1% of a photo's file size. Even if there is an embedded thumbnail in EXIF from the RAW conversion, it's still only likely 5-6% of a photo's file size. Whatever is happening when HologanaSimp is "stripping" the files, it's involving some further form of compression, which already endured that compression in the export from the photographer's RAW file. JPGs are lossy file types, so conversion is going to result in degradation of photo quality, no way around that...
Which circles back to my original point of it being best practice to share source quality files.
 
Does anyone has a good ai edit of her? The fake/ai thread was taken down :(
 
You must log in or register to reply here.
Back
Top