Photogrammetry Statues #7

Work In Progress / 22 March 2020

Hey everyone, 

Another one down, although this one was a lot easier to retouch. I took the base scan in pretty much ideal conditions and it turned out pretty good straight out of RC. Still needed some retouching but it wasn't anything too crazy. 

Real Life Statue:

Now comes a part that which I've been dreading since starting this, which is Retopo. Been testing out a few automatic methods that don't involve a lot of manual work and some of them have been giving me pretty good results so hopefully it won't be as tedious as I'm expecting :)).

Have a great week everyone!


Cheers, 

Vlad


  

Photogrammetry Statues #6

Work In Progress / 08 March 2020

Hey guys, 

This is one that I took with my phone while on vacation in Florence. Turned out ok considering it's taken with a phone, but I really want to go back with a proper camera and scan as many statues as I can. Florence was a treasure trove in this regard. 

The retouch was quite straightforward when it comes to the girl, but I probably would have rebuilt the pedestal regardless of the quality, since it needs to end up as a game ready asset in the end and manually building it allows me more flexibility. This will probably be the last one of the phone scans unless I find another one that is worth retouching. 

And this is the actual reference of the statue:

Hope you like it and keep the great art coming!


Cheers, 

Vlad
  

Photogrammetry Statues #5

Work In Progress / 09 February 2020

Hey everyone, 

Managed to break myself away from Zelda BOTW enough to finish touching up this statue as well. The scan was taken with a proper camera this time and it has definitely helped a lot during the capture and with the final result of the scan. As you can see in the images below, the faces didn't really turn out quite right, so I went back and re-scanned them, which made retouching easier. 

Just a few more to go now and I can finally start working on the environment proper. Still a way to go since I need to get them game ready, but I'm looking forward to actually seeing them in Unreal with some proper lighting. 

As usual, hope you like them and hopefully I'll have another one done as soon as possible. 


Cheers, 

Vlad

Photogrammetry Statues #4

Work In Progress / 17 December 2019

This is by far the worst scan I've ever taken. 

So was it worth it to restore? No, for production I would just toss it. But I thought it would be a fun challenge since I missed doing something more organic, so I decided to try and salvage it. Even though there are still some areas where that I would like to work on more, I'm overall happy with the result.



I have one more taken with an iPhone that needs a bit more retouching, but the rest are taken with the new camera which saves quite a bit of time and I can get better results with more complex scenes, so some of those should need less retouching. After that, retopo and baking for all of them, that should be fun :)).


Until next time, 

Vlad


Photogrammetry Statues #3

Work In Progress / 28 August 2019

Hello again everyone, 

Another bust down, this is getting quite addictive. This particular bust was more of a challenge compared to the other one because of the poor lighting. Museum lights are great for statues, really bad for photogrammetry, so the result for this particular bust were not that great. The back side is basically non-existent and the face and cloth are extremely noisy. Still, managed to get it close to the original statue. I haven't perfectly captured the air of superiority and disdain in the original though, heh.

Real life statue:

That's all for this small update, started working on the next one which is the most challenging one yet in terms of retouching. Can't wait to get the DSLR and test it out soon. 


Have an awesome week peeps!

Photogrammetry Statues #2

Work In Progress / 20 August 2019

Hey everyone, 

Was meaning to update this blog more regularly, but clearly things have not gone as planned. Life has been busy in general and I also managed to break my ACL biking, so that put some things on hold. 

However, while recovering from the injury I couldn't just sit and watch Netflix all day. It's not nearly as entertaining as I thought it would be. So the little time I could actually sit at my PC, decided to go through some old photogrammetry photos I took while in Florence a while back. Was very pleasantly surprised to see that some of them turned out really well thanks to the overcast lighting.  So I decided to keep retouching multiple statues and create a collection for myself since I have several scenes in mind that could use them. This is the 2nd statue I've retouched and the process itself has improved quite a bit, although a good base scan is really the best thing to have.

This is what the statues looks like in real life. Again, the purpose of these is to have a hi poly base that can then be used in multiple ways and scenarios (broken, damaged, different types of materials, etc.). This is why there's no detailing (cracks, edge damage, etc.).

While these scans turned out pretty good, I can feel the lack of resolution and image quality in the scans. Many facial features are not as sharp as I'd like them to be and constantly refining them can become tedious. So I'm going to upgrade to a DSLR camera, which also means more vegetation and also photography (whoop whoop).

And that's it for this small update. I'm going to post the statues as I finish the hi polys and then just do a big Retopo pass on all of them so I can get them inside Unreal. Really excited to finally see them come together, even though it's still a ways off.


Have an awesome week!

Guest on The Student Art Podcast

General / 03 July 2019

Hey everyone, 

I recently had the pleasure of talking to Ross Baxter on his Student Art Podcast about a variety of topics concerning Environment Art and art in general. We covered a lot of topics including:

  • Pros and Cons of going to University versus being self-taught
  • Best resources to learn Substance Designer
  • Large scale environments
  • How to tell when it's good to struggle, when you should take it easy and much more

You can listen to the discussion here:

Also, make sure to check out Ross's page here on ArtStation: https://www.artstation.com/rossbaxter . He was recently nominated for a Rookies Award so make sure to support him if you like what you see!

I really appreciate the invitation and I hope this helps you on you art journey!

Photogrammetry Tests

Work In Progress / 12 April 2019

Hey guys, 

While playing through the massive Ghost Recon Wildlands, I've also been playing around with photogrammetry lately and testing out a more "guerrilla" style way of doing photogrammetry. Basically, what I'm trying to test out is how far I can take photogrammetry of certain things (statues in this case) in sub-optimal conditions. I'll go through each point in a bit more detail, as this is something I've been thinking about trying out for a while now.

1. Why sculptures?

Multiple reasons, but the most important are these:

  • They take a lot of time to make and are difficult to get right, even for character artists
  • In production, they take up very little space visually, but can add a lot, depending on how they're used. Unfortunately, they also take a long time to create and that time usually comes out of the character art team, which is not the best use of their time
  • They can be used in multiple scenarios, but because of their high production cost, they rarely are. These are just a few examples of their use:

Some of the above examples would be prohibitively expensive to create during production, which is a shame, since they add so much to the quality of an environment or a prop.

2. Challenges

Normally, good photogrammetry requires good equipment and good weather conditions (overcast), which is not always possible, especially since a lot of these sculptures are inside museums, where the lighting conditions can vary quite a bit. Because of this, I've decided to simply go the other extreme and work my way up from there if needed. Basically, all I'm going to use is my phone and take the pictures inside a museum and see if the result is usable. 

One thing to note though, I am only interested in the 3D geometry, not the base color. That offers me the most flexibility when it comes to how I use these complex sculptures. They can be made out of marble, bronze, damaged or anything else you can think of to add even more variation. The Hi Poly needs to be the same as something a character artist would produce. For this purpose, I set my target quality to the gorgeous sculptures done by Elliot Mallon for Assassin's Creed Unity:

3. Scanning

The statue that I want to capture is here in Stockholm in the National Museum. You can see the lighting in the pictures below. As you can tell, it is definitely less than optimal for photogrammetry - a lot of hard shadows, different colored lights and a lot of the areas are overexposed to light while others are almost not lit at all.

4. Results

Ok, having gotten everything out of the way, this is the result I got after about 280 pictures with an iPhone 5 SE (definitely not the best camera):

As you can see, the capture is not usable at all straight after processing. The scan overall is extremely noisy. The cloth, hair and base turned out alright, but the face, legs and arms are all extremely poor quality, noisy and there's even some holes in the mesh.

However, these are the results after about a day's worth of retouching the mesh in ZBrush:

This is still WIP, as I still need to do the hair, base and refine it overall, but the results are very encouraging. The retouched Hi Poly is definitely usable for baking and all the details from the real life counterpart are still there. Better still, all of this took about a day's worth of work, which is definitely usable for production.

Next steps are to finish the retouching and do a test bake to see how it all works out.


Have a great weekend and let me know what you think in the comments!