Profile Picture

How to combine Headshot with SkinGen and photo editing for the best results?

Posted By midix 4 Years Ago
You don't have permission to rate!

How to combine Headshot with SkinGen and photo editing for the best...

Author
Message
midix
midix
Posted 4 Years Ago
View Quick Profile
Distinguished Member

Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)

Group: Forum Members
Last Active: Last Year
Posts: 59, Visits: 362
Now when we have photorealistic skins and tools, what would be the best and possibly cheapest pipeline to do the following?

- take a photo of a face as a reference (important: the photo itself is not of premium quality)
- use the photo to generate and adjust the general head shape and features - ok, there are many tutorials about this
- merge the person's characteristic features of the photo into the high-quality default facial skin textures provided by CC - that's what I'm missing

Are there any tutorials or planned webinars that go through that last step, which is crucial for achieving the highest resemblance possible and not lose face skin resolution because of the reference photo quality?  It is essential to transfer wrinkle normal maps and skin discolorations to the high-quality default skin templates, but not taking the ref textures directly because of their lower quality. I imagine wrinkles could be transformed into normal maps through some tool (xNormal maybe?) from the ref image directly because wrinkles are usually not that crisp and sharp anyway, so they would not suffer much from lower resolution. But skin pores and lip details and eyelashes and brows and beard stubble (hopefully, next CC update will improve in this regard) must be of the highest quality possible and should not be taken from the ref photo "as is".

Is it at all possible to achieve using just a photo editor and free or one-time-payment tools? Has anyone done that? 

I'm just a hobbyist, I don't have ZBrush nor Adobe subscription. I invested everything in CC+iClone with a hope that one day it will have all the necessary tools built-in (except photo editing) to create high-quality replicas of reference photos. However, the last step is still not clear. Most tutorials stop at the point "ok, now it looks somewhat like the photo, and we ignore that the facial skin is blurry because default skin texture got replaced by the facial ref photo".

I might get SubstancePainter perpetual license from Steam someday. But for now, I have only Blender and PaintShopPro and Gimp. Being a hobbyist who doesn't use 3D tools every day (sometimes a few months go by), it would be very inefficient to invest in Adobe's subscription.
planetstardragon
planetstardragon
Posted 4 Years Ago
View Quick Profile
Distinguished Member

Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)

Group: Forum Members
Last Active: Last Week
Posts: 11.5K, Visits: 45.9K
when I first started with iclone ...at around version 3,  we didn't have all the bells and whistles we had today,  so we had to fake the funk with various tricks ...mainly on images ....as time went by,  shadows became a bigger thing towards the path of realistic renders which was usually the biggest wish from the community -  so now came bump maps .....then eventually to enhance the bump maps - came ambient occlusion,  then after that came refining the shadows and their quality ....then was coloring the shadows and reflections with ibl lighting and global illumination ...long story short,  I have a good grip on how to get the look i like because I understand the basics and the history of how everything came to be today...

if you really want to understand how to render any style you want ...you want  to understand the fundamental ways Iclone handles images, lighting, reflections and shadows before using the advanced tools - then you can tweak your workflow for you how personally like to work - as there are many options and softwares to create a variety of workflows

to start you really just need an editor and I use the free nvidia normal map generator tool or substance designer b2m bump maker ....the other maps you can easily make with paintshop pro going grayscale ....

Familiarize yourself with the basics - all the advanced options are built on those basics -   lighting,  shadows ,  reflections  -  all the new tools / plug ins ...are additional options to tweak those basics.   Once you understand those basics,  then you'll understand how the new tools come into play in your workflow ...once you understand that ...you know what 3rd party tools you'll need to fill in the blanks.  Moral of the story,  the most important part of your workflow is your knowledge and understanding of whats already provided in the software,  starting with the basics.




Edited
4 Years Ago by planetstardragon
midix
midix
Posted 4 Years Ago
View Quick Profile
Distinguished Member

Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)Distinguished Member (1.0K reputation)

Group: Forum Members
Last Active: Last Year
Posts: 59, Visits: 362
Yes,experienced people with in-depth understanding of 3D design topics and their own toolset might have no problems to integrate iClone/CC into their workflow. However, for hobbyists who are starting from the other end it might be difficult.

I myself am a "mixed bag". I am a web programmer and 3D simulation development is my hobby (albeit I can devote too little time to it). I have learned some 3D basics long time before CC came out using 3DsMax education editions and Blender. However, I didn't have enough time to get enough practical experience, and I often find that when I open Unreal Engine and Blender I have to relearn some things (yeah, they are changing lots of stuff lately).

So I hoped that CC would become "one-stop-shop" and it has the potential.

I hope Reallusion will dedicate some time to produce tutorials for CC integration into full character creation process to achieve the highest quality possible. It's somewhat disappointing to see them stopping so close to the target. On the other hand, I understand how they might not be particularly interested in talking about integrating with other software because every time you have to switch it's like a "minus one point" for CC due to being unable to provide some function. But yeah, it would be too much to ask CC to integrate photo editing and 3D texture painting, and thus it would be nice to have more learning materials for getting it all working together.
planetstardragon
planetstardragon
Posted 4 Years Ago
View Quick Profile
Distinguished Member

Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)Distinguished Member (23.2K reputation)

Group: Forum Members
Last Active: Last Week
Posts: 11.5K, Visits: 45.9K
I'm actually sharing with you another hobbyists point of view.... all the extra tools like blender and substance painter actually complicate the process ( just more to learn ).  I do understand what you are asking,  and I'm trying to help simplify it for you so understanding the advanced tools need not be tedious or complicated. 

The simple answer would be use 4k images because they create sharper detail on the character,  and you just need paintshop pro to do any tweaks.

but in order to edit the images effectively to get the details you really want ...you want to take a moment to understand how CC3 reacts to the image detail.  For example,  the higher resolution you have on your base image, ie 4k  the sharper the lines ....the sharper the lines on your main photo image,  the better you will have a conversion to a normal map, further you will know how to edit the normal map in paintshop pro  cue shadows and how iclone plays with sharp images vs dull fuzzy images vs shadows, ( each of which has it's place in a high quality render ) Then understanding how the lights can completely highlight or conceal all the details you added via normal map.  - And if your lighting is wrong,  no amount of detail will make your image a higher quality, if your lighting is on point,  you don't need much detail to get the shot you want.

Iclone and cc3 are like onions,  layers and layers of tools built up on a foundation of  basic concepts.  If you understand those basic concepts, the advanced tools become intuitive and are surgical instruments,  without those basic concepts,  the advanced tools are simply a fantastic set of presets that you don't have to tweak much to make your work look great.

The art of illusion is in the shadows! 



Edited
4 Years Ago by planetstardragon



Reading This Topic