HOME | DD

Description

Inspired by "Nightbloom" by Masayoshi Soken from Final Fantasy XIV.

The Nightbloom is a theme for Yotsuyu's character in the game. She is one of the best villains (with very tragic backstory and development) in Final Fantasy XIV. I really wish to cosplay that character one day but until then, I am happy with a artwork inspired by all of that.

The model was a result of AI generation for my previous image. I fell in love with the pose but it was way too dark for that light image I had in mind at that time. But I kept the image and started to have this idea around it. I knew it will be girl sitting by the pond on the moonlight and I knew I want red Japanese maples (if you have not noticed yet, I love red Japanese maples - I actually even have a tree in my own little Japanese garden I started few years ago). So I started with the tree and girl "floating" by the pond. Then I spent ridiculous amount of time cutting spider lilies (symbolizing loss, death, separation and abandonment, sometimes even called "death flowers" in Japanese folklore) - considering the tragic backstory of Yotsuyu, these were perfect on symbolic level but I just could not get select color feature to cleanly separate flowers from background. And then after I had sorted out the flowers, I felt lost with the night sky. I tried tens of combinations of stars and crescent moon (Midjourney is horrible when generating crescent moon almost as bad as with hands). I think I even gave up one day without any real progress. And then I just managed to create almost what I had in mind - not all days are same progress wise.

For this image, I updated my photoshop (not sure why, but my photoshop had not had update for a while and then I was missing one windows update that did not allow me to update to latest version - sorted it all out in the end) and got to try PS version of AI tool too. It was different experience than using Midjourney. So far I have been using Midjourney to generate myself "stock images" from where I cut out the object, adjust colors etc, paint over etc.. then PS AI tool works like "you select the are and we modify that for you" - most likely I could also generate full image out of nothing but I mostly used it as a blending tool - I had hard time blending the clouds and garden silhouette, so PS generated some mountains between them for nice blend. Same with the red leaves at the top. Also, it was good for adding some additional details with correct angle (like small koi fish in the pond - tried it for so long to get Midjourney to generate my koi with right kind of angle.. hopeless).

Also, the funny story how I tried adding tome-ishi (or stopstone - basically a stone with rope around it. An element from Japanese tea gardens). I know I have photos of these somewhere but ofcourse you can never find it when you want to use it. I tried telling Midjourney what I want - first using the term in English and Japanese, then describing what it is. No luck. Also tired this generative fill - I got stone with sign "stop" on it when using Japanese term. That was funny but not what I was after In the end, I gave up on that detail.

So hopefully with next image, I will learn even more cool tricks.

Cloud brush , AI generated stock + some personal images and painting

Description

Inspired by "Nightbloom" by Masayoshi Soken from Final Fantasy XIV.

The Nightbloom is a theme for Yotsuyu's character in the game. She is one of the best villains (with very tragic backstory and development) in Final Fantasy XIV. I really wish to cosplay that character one day but until then, I am happy with a artwork inspired by all of that.

The model was a result of AI generation for my previous image. I fell in love with the pose but it was way too dark for that light image I had in mind at that time. But I kept the image and started to have this idea around it. I knew it will be girl sitting by the pond on the moonlight and I knew I want red Japanese maples (if you have not noticed yet, I love red Japanese maples - I actually even have a tree in my own little Japanese garden I started few years ago). So I started with the tree and girl "floating" by the pond. Then I spent ridiculous amount of time cutting spider lilies (symbolizing loss, death, separation and abandonment, sometimes even called "death flowers" in Japanese folklore) - considering the tragic backstory of Yotsuyu, these were perfect on symbolic level but I just could not get select color feature to cleanly separate flowers from background. And then after I had sorted out the flowers, I felt lost with the night sky. I tried tens of combinations of stars and crescent moon (Midjourney is horrible when generating crescent moon almost as bad as with hands). I think I even gave up one day without any real progress. And then I just managed to create almost what I had in mind - not all days are same progress wise.

For this image, I updated my photoshop (not sure why, but my photoshop had not had update for a while and then I was missing one windows update that did not allow me to update to latest version - sorted it all out in the end) and got to try PS version of AI tool too. It was different experience than using Midjourney. So far I have been using Midjourney to generate myself "stock images" from where I cut out the object, adjust colors etc, paint over etc.. then PS AI tool works like "you select the are and we modify that for you" - most likely I could also generate full image out of nothing but I mostly used it as a blending tool - I had hard time blending the clouds and garden silhouette, so PS generated some mountains between them for nice blend. Same with the red leaves at the top. Also, it was good for adding some additional details with correct angle (like small koi fish in the pond - tried it for so long to get Midjourney to generate my koi with right kind of angle.. hopeless).

Also, the funny story how I tried adding tome-ishi (or stopstone - basically a stone with rope around it. An element from Japanese tea gardens). I know I have photos of these somewhere but ofcourse you can never find it when you want to use it. I tried telling Midjourney what I want - first using the term in English and Japanese, then describing what it is. No luck. Also tired this generative fill - I got stone with sign "stop" on it when using Japanese term. That was funny but not what I was after In the end, I gave up on that detail.

So hopefully with next image, I will learn even more cool tricks.

Cloud brush , AI generated stock + some personal images and painting