>>9223
I've been putting it off for too long; I think I need to write a tutorial on my morphing process like I did for SD. Unfortunately I didn't save the original PSD from when I did that morph at
>>8108, plus I'm not entirely satisfied with the result. So, I decided to do it again from scratch, and this time document my steps. Consider the following an extremely rough draft for a tutorial that doesn't go too deep on detail.
In the first image (CompositeCompare), you'll see the source image followed by the first round of selections (I darkened the background just to highlight the individual pieces). I've broken the model up into Face, Chest/top, Arms, and Trunk (lower abdomen and legs), and that is also the layer stack order from top to bottom. In the third image, you'll see I've added in a new belly (this one was sourced from Mandy Majestic). I know that the fourth image is a bit of a 'Draw the rest of the fucking owl,' but you can see that I've morphed each of the different body parts, added in shadows, color corrected the belly, loosely painted in new/extra cleavage, and fattened the face. Before SD, this would have been a much more painstaking process where I would have worked much harder at color matching the new belly, making the shadows have the right luminance, making the painted cleavage look more natural. and compensating the 'pixel stretch' that's most visible on the hemline of the top and the outer edges of the belly. But thankfully I can leverage SD to automate this, and in some specific areas wind up with much more realistic looking skin and details.
As an aside, I've also highlighted the process I use for plumping up a face and adding a double (or triple in some cases) chin. The first image shows a duplicate of the original face selection with parts of the nose and mouth filled in. I just use the photoshop Fill feature, nothing fancy. The second step shows the general shape I strech the face into, taking care to only nudge the bottom of the eyes while leaving everything north of the browline untouched. (lifting the lower part of the eye can enhance the 'fullness' of the cheekbone.) The third image shows the original face selection brought forward, and the last image is the final result after strategically masking parts of the original face to blend in with the new layer of fat. From there I just paint in some more shadows where necessary, and voila, that's a chubby face alright.
And finally we get to the rendering. The first image is the final stage of the edit: I've make most of the adjustments I want to see, and flattened it down into a single layer. I select our subject (as the top 20% of the image is just flat ceiling, no reason to waste cycles reprocessing that) and place her into the img2img section of SD. (If you'd like to learn more about my approach to SD itself, I've linked my writeup at the bottom of this post). So, moving to the second image, we see the result of processing. Not too bad, but we've lost the model's face and hair, making her completely unrecognizable. The third image is the result of morphing the SD result again (mainly tweaking belly and chin/waddle shape), color correcting the face and chest, and masking the edited face back in. I flatten this and send it back into SD one more time after tweaking some img2img settings, drop the new generation back into photoshop, and mask the face and hair back in, resulting in the final image.
And there you have it! I know this is pretty brisk and glosses over a lot of details, but you should be able to understand the basic process. Also, this image is still pretty slapdash. For most of the images I post to my DA, I usually spend a bit more time ironing out smaller imperfections and tweaking composition a bit more. For instance, in this image it would probably make more sense for her right forearm and hand to be visible (perhaps resting on her thigh) or for it to exaggeratedly mirror the position of her left arm to make it more clear that she has her hands on her hips like in the original image. But that would take more work than I'm willing to commit to this tonight, so I hope you'll forgive me.
I'll eventually do a much more thorough, in depth write up on this, or perhaps even record a video some day. I hope this will suffice for now, but feel free to ask any questions you may have.
Link to my SD process writeup (you don't need an account to view it): https://www.deviantart.com/theguywhodidathing/art/How-did-you-DO-that-989204336
Note that it's technically about generating images from scratch, but once you get to the img2img section the process remains largely the same.