Website Update and some new plans

#1 Nice to see you're active again with this @Ron.

#2 Thanks Nils, I have some questions concerning your work with the sunflare/bokeh postFx filter from Felix....

You mentioned you were able to scale it and, honestly I did not understand your piece of code at all. Care to explain that in more depth? I have added this into my 'cinema' build and I do not quite get what you did with it. (great resource from Felix by the way).

Sorry to pick your brain... and thanks... glad to BE active again.

Ron

#3 Sure @Ron, a bit off topic though :-D :-D :-D

In lightFlareData.cpp:

if ( mElementScale[i] < 16.0f )
{
size *= mElementScale[i] * mScale * lightSourceIntensityScale;
}
else
{
size *= mElementScale[i] * mScale;
}

means that if you define in the datablock for example:

// Halo with a fixed size
elementRect[0] = "0 256 256 256";
elementDist[0] = "0";
elementScale[0] = "16.2";
elementTint[0] = "0.54 0.54 0.54 1";
elementRotate[0] = "0";
elementUseLightColor[0] = true;

This element won't be scaled by lightSourceIntensityScale so it will have a fixed size around the light source. I used the value 16 because that's just below the size of my halo element. Of course you could turn this into a boolean instead; but this solution is backwards compatible since the structure of the datablock isn't changed.

While it's fixed and fading in/out with the light source; could this also be used for other effects then halo. The only limitation here is the size of your sunflare sheet so to speak.

Good you brought this up; apparently we didn't migrate this from the old DM to the new. Now I can push this tiny hack right away hehe ;-)

#4 Just went and read the blog post you put up on photogrammetry on your site - excellent stuff! The cemetery aesthetic is pretty perfect, as you discovered. And it's nice when people don't just assume you're up to no good... I imagine if you were 20 and wearing a hoodie you'd have had a different response ;).

Funny thing - last year I worked at a place right across the road from the largest cemetery in the southern hemisphere. I'd get off the bus at 8am and the cemetery would look absolutely beautiful in the horizontal morning sun - I always wanted to try to capture that in a game level. So I'm now waiting in anticipation...

#5 Website looks nice! Great work!

2 things, there seems to be an issue with labels on the images on the frontpage. For some reason they start blinking on and off, but I have not been able to find a stable reproduction case and it doesn't really matter :P

2nd, this blog page: http://3tdstudios.com/blog.html says "article 2" where is the first one? :P

#6 Good to see you back in action Ron.

#7 yooohooo Ron thanks for the non-update :D and glad to hear everything is going good!!

#8 Thanks all,

@ Nils, bit off topic but, not really. Thanks for the reply. I will dig into it this weekend. I like the effect. After that I want to apply it to the other light types as well. I think I have some ideas for that.

@ Lukas, you are correct. I will put the first article up again once I figure out a good way to navigate back and forth. Also, I will check the labels issues as well (It a simple javascript, I probably messed with it at some point and did not reset it.)

Ron

#9 I like the new website Ron, well done.

#10 Thanks Ron, everything sounds great. The Blender tutorials will help a lot.

#11 Always interested to see some tutorials for the new Blender (still using 2.44 myself, was thinking of updating to the latest version though).

#12 Steve, Highly recommend it. I FINALLY jumped in for real at 2.7 and have not looked back. First tutorial should be up next week. (Blender to Game Environment Basics - High Polygon Model to Game Ready Asset) It will cover a ton of stuff... mesh refinement (using Sculpting tools), retopology, texture baking (using cages).... a ton of stuff to take a high poly model (think a zbrush sculpt or photoscanned model) game ready with texture maps and all.

Hell, here is the list of 'in progress' tutorials so far;

High poly to Game ready
Light mapping interior spaces (using Blender Cycles)(better way to do interiors without the silly 'shadow map' corner gap problems)
Trees Rocks and Plants (oh my)
Custom Character Models
Basic Blender Animation to Game Ready Asset
Blender Sculpting to T3D Terrain tutorial
and more

Figured that would be a good way to start :-)

Ron

#13 I'm kind of excited for those tutorials as well! Just wish I had time to get my art on :(.

#14 Thinking about this, in theory one could produce assets for a large FPS. There is an asset on the Unity store that allows one to make textures out of any google street view image and turn it into a panorama, then using something like mega grab and SSAO pro, one could get fairly high res textures out of Unity. Then put these panorama views through Photoscan and end up with a realistic city (in chunks) cheaper and quicker than normal. Of course, google street view images cannot be used except to imbed in apps (so I believe, though never checked) but no doubt a large company could license the rights to use google street views in this way. Theoretically feasible but way beyond the means of an indie developer. Another one of my numerous 'if only' ideas.

#15 Actually, I have done it with some of Nasa's large scale panoramas from the Mars Rover Missions. The results were not bad at all. (Though mars is a tad flat and boring.) I will post some shots eventually. So, yeah I can confirm that it does indeed work well for that. The only 'drawback' is exactly what I stated in my article, If the camera can't see it....then the computer can't produce a mesh. SO, panorama sources are ok but not ideal. Although that leads me to another fine use for this technique... terrain.

Ron

#16 I don't think it will be that easy, there is already technology that can scan things pretty good, but it is all still far away from being usable for games.
For games you need full geometry with full interiors, things you can interact with and not static scenes.
Another issue is you can only reproduce existing things, so for fantasy worlds you still have to build by hand.
Scanning has it's uses, but it is not the solution for everything.

#17 @Duion - There are some models made with this technique already on Turbosquid that are just the front and side of the buildings, useful for background props for certain types of game, they cost $10 and saved a lot of time in the 2D point and click game I am currently finishing off When Jester Interactive made a racing game about the TT Race, the team drove the entire length of the course filming and photographing every inch and then modelled it by hand. What they didn't do though was drive into the houses to model the inside as well given that doing so in the actual race is an automatic disqualification.

Obviously photogrammetry is not a panacea for all the ills of the game making process. However, the vast majority of games are not 3D but even being able to scan a building front would save time in a lot of instances. It's a moot point though as no indie developer has the resources to license the use of street view images from Google. In fact, there is a standalone software called World Composer which makes use of Satellite imagery but the developer explicitly points out the copyright to any images used belong to the image creator.

@Ron - it would be interesting to see those pictures. Even from a non-game asset perspective though, I'm hoping to be able to use this technique because from a preservation POV it's brilliant, less than a hour drive away are 3 Celtic burial mounds 2500 years old it would be great to recreate them as a model.

On the other hand, if I was creating a realistic looking click and point 2D adventure being able to scan an interior like the room I'm sitting in which has over 5000 books on shelves, piled 4 deep with boxes full of crap, would save untold hours of modelling time to render a realistic looking scene, so even if one did not have the full rig to capture large exteriors it has an enormous amount of uses for small games.

#18 @Duion, Yes you are correct. It's NOT easy and it is NOT a 'perfect/complete' 3D solution. It is a NEW tool in the game artist arsenal.
With a bit of practice artists can use this as a SOLID way to make game ready models. Vanishing of Ethan Carter and Get Even have proven this already. Even the CD Projext Red (Cyberpunk 2077) via Platige Image) used it in a 'teaser' trailer. Granted this was a full-rez, CG style movie, still.... it's an 'emergent' tech. Try, learn, adapt or fall behind.

I don't JUST put this stuff out for T3D. In fact, unless T3D does some major 'F'ing stuff, it's going to go the way of the dinosoaur. I put this info out for Indie developers, in an attempt to show what CAN be done (even on a budget). The tutorials I plan on releasing (this week) ALL use photoscanned objects as a starting point. I show HOW to make them game ready. Yeah, and IF you bothered to look at my site you would already know I have a pipeline in place to do this.

You have some great skills brother, just do not get 'bogged' down because you read somewhere or saw something on the freakin internet that said you 'can't' do something. Artists, (me and you) always need to think, ok 'F' you.... I want to do that, now how can I given my restrictions? I really think that is one of the base aspects to art in general.

Lastly, the best Sci-Fi (In my Opinion) is about taking what we know now and 'updating it'. Think Blade Runner, or even the original Alien movie. Heck, even Star Wars has a number of good examples. Good sci-fi (to me) is about taking the familiar and turning it on its head a bit. The assumption is that our basic morpholology is never going to change so, we will still need things like handles, buttons, etc etc.

This technique allows artists to take something common and (with a bit of skill and training) create something new from it. I like to think of the first model as a lump of clay. I start with photo scans, then drop it into Blender... From there.... I can do almost anything.

Ron

#19 @Duion - this has been knocked about with by a number of indie devs who need to look for a way reduce asset creation costs. There's an article from Ludum Dare by a couple of German developers that pretty much come to the same conclusion and the game they used the models in is quite interesting to see what can be done with it, even if not thinking about realistic looking props: http://ludumdare.com/compo/2014/05/04/360358/

#20 The most awesome stuff i've seen is from euclideon, they also scan everything and they also have some demos of complete scanned scenes: www.youtube.com/watch?v=DbMpqqCCrFQ
This almost looks like real, you can only distinguish, if you look closely.

I'm not against this technology, I just say it creates different problems, like getting material to use, because of copyright issues, since you cannot just walk into other peoples buildings etc and scan them.
You can also achieve very similar results to 3D scanning, if you just project photos on a shape, since most shapes are very simple.

But on the other hand I see the physically based rendering systems coming up, they are also very good and close to reality, no idea what technology will win.