Thats the one I (and Hanno Hagedorn) learned the basics of organic modelling from. Way back when our English was so rudimentary, we didnt know that timelapse means speed up.
So we tried to hit Irfans speed doing the same thing over and over and over again hot seating on Hannos PC, we got close, tho never this fast.
Took me a while to realize why we couldn't hit it, hehe
Johnathan Williams still has some public videos of his edge modeling and facial topology series from the late 2000's and early 2010's. This is fairly representative of how things were typically done prior to digital sculpting. At least for those using open source tools from the period. Tool sets do vary but a lot of the fundamentals are similar.
The site is long since defunct but back in the mid to late 2000's there was a forum dedicated to subdivision modeling where artists shared tutorials and offered critiques on work in progress. Unfortunately it seems like a lot of stuff wasn't archived and threads like "Super basic workflows" are now completely lost to time. https://web.archive.org/web/20070712091532/http://www.subdivisionmodeling.com/forums/
There were also a whole bunch of other sites, like CG society, that had similar discussion boards and those were a primary source for a lot of us who either couldn't afford or couldn't access professional level training but still wanted to learn subdivision modeling. Prior to video hosting platforms a lot of online learning content was either posted on forums, which have long since been shut down, or hosted somewhere as web pages and downloadable text documents.
Stuff like this was really all that was available for free and like
others have said, the actual modeling process is pretty rudimentary and
easy enough to replicate today using basic poly modeling operations. A lot of
stuff we either had to figure out on our own through trial and error or
find someone who knew what they were doing and was willing to share
their knowledge. Often with nothing more than a couple of paint overs or a scanned page from a book
and a few lines of text.
It is possible that someone has fragments of these popular workflow discussions or at least the media from them saved somewhere. Though, to be honest, if you want to see these older workflows in action then it's probably easier to try and find modeling time-lapses from 2006-2010 on YouTube. Back in the day it was a somewhat popular thing for organic and character artists to share modeling time-lapses. This trend continued for a while as more and more people shifted over to 3D sculpting applications like Mudbox and zBrush then slowly faded away.
When searching for videos on YouTube, try keyword phrases like: "edge modeling topology time-lapse before:2010" and the search results will be filtered to only included videos from before the selected date. This makes it a lot easier to find content from working professionals who were sharing their workflows back in the day.
Even stuff like basic box modeling time-lapses and other tutorials from over a decade ago are still out there.
When it comes to the question:"How did people do it?" Often the answer is: artists worked with whatever was available and iterated a lot until things looked correct. Especially when working from concept art or replicating obscure subjects.
Although it's not ideal, it wasn't all that uncommon for people (especially when learning) to create models based on a very limited pool of references that were nothing more than a front or side view and a handful of turnaround images. For characters it might even be as basic as a single concept sheet with just a three quarters view and some basic proportions from a front and side view.
As unsatisfying as that sort of answer may be, if there wasn't direct access to the subject or a lot of reference material that you or someone else gathered then it was often up to the individual artist to fill in the blanks as they were modeling. Usually with some input or guidance from an art director or concept artist.
In the how do you model this thread, Per used to have a few really
good posts with examples that demonstrated what older concept block-out
processes looked like for props and other stuff that was still being
done with subdivision modeling but that content is also lost to time.
Here's a link to the original JOA tutorial that was hosted on 3Dtotal: https://web.archive.org/web/20080621094924/http://www.3dtotal.com/ffa/tutorials/max/joanofarc/joanmenu.asp It takes some doing but it is possible to occasionally find this stuff. Whether or not it gets archived is kind of hit or miss. Knowing where it was hosted also helps and those lost connections sometimes make it difficult to find things people talk about seeing or learning from in the past.
Well, that's precisely why I was mentionning earlier that some of the confusion in this thread might be coming from where you are in your learning of polygon modeling in general. This (and the weird question about the ears) 100% confirms it. And again none of this relates to "old" or "new" models (beyond the fact that observing "old" models made you indirectly realize that there is a gap in your toolset).
Things like "edge modeling" or "box modeling" (or "strip modeling", or anything else really) are just casual terms that artists make up on the spot to summarize a given approach they like. You won't find books on these things. Perhaps some video recordings done by artists to document this or that trick, but nothing of substance - because it would be pointless. What you need first is to master the polygon modeling toolset of your software, in and out.
If you don't know where to start (for instance if the software you are using doesn't have any official Getting Started content - like this is unfortunately the case with Blender, as the official Blender help link basically states "we have no time for this, go look for stuff on youtube lol") then look up *quality* educational content from older sources. I mean specifically older in this case in order to avoid sculpting-centric material. You could basically sort through Gnomon DVDs and look for anything related to character modeling from 2010 or earlier. They had top-notch content.
https://www.youtube.com/watch?v=r8AxXG_6jyA Or, look up any version of the "Johan of Arc" tutorial PDF even if it isn't ported to your own software. You'll see all the steps people go through when building polygon models from scratch, you'll learn the generic names of the various tools and actions (extrude, merge, divide, delete ...), and you'll then be able to research how to perform these action in any software easily. Heck, you could even download help files from various software and research the tools that way.
At the end of the day you should be able to build any mesh topology that you can think of (or draw on paper). If you can't, then you're not familiar enough with your software. People learned all these things back then by studying the aforementionned ressources as well as studying casual video recordings of artists at work, shared in places like ... Polycount A lot of this naturally gets lost with time. In my experience there is also a lot of knowledge that gets exchanged in the workplace. And 15ish years ago was the peak of monthly CG magazines too.
Also the Vertex books by Ryan Hawkins are still as awesome as ever, and should be easy to find a DL for since they were a community project released for free. These are not just about polygon modeling though of course, they cover anything that the contributing artsts wanted to share.
More recently, I've seen people pick up polygon modeling from complete scratch as a direct consequence of their interest in modding games or contributing to a Steam Workshop. In such cases the information is mostly in the corresponding Discord servers.
At the end of the day, it all comes down to studying good models and being able to create any topology. Videos are cool and all and they can reveal neat tricks, but at the end of the day a polygon is a polygon.
Well, it wasn't PBR back then, but it was all photoshop. Those environment textures are photo sourced, edited to be tileable. In the case of iron man, its probably just raytraced reflections.
The Automaton race is baked and in engine. It's the last race I'll create for now, excluding a secret sixth race planned that'll be revealed in a DLC or expansion far in the future. Future plans are a bunch of minor improvements across all races, defect check all races for baking issues, and take these base meshes to the next level by actually painting them in Substance; at the moment they only have Zbrush polypaint vertex colour. Afterwards - basic locomotion animation tree in UE5 & FACS facial animation system.
I'm thinking it's hard to tell whether if you're working for microsoft or nvidia, and what percentage of the trillions invested into generative AI is allocated to marketing