/3/ - 3DCG

>>569212
I'm in the same boat. Itd actually what prompted me to try daz for the first time but they were too detailed and too difficult and time consuming to set up.

There really is a lack. Especially of a variety of ethnicities too. I work in Asia and we have this client who almost got upset there were no pretty Asian girls shown around in our real-time render.

When necessary to show big areas with civilian / pedestrian and traffic flow we use twinmotion, and it's got a good variety of people. I haven't tried to rip the models from there but that's doable, you can set up paths for them to walk too

>>569224
>I'm in the same boat.
Yes. We're all looking for female packs but they seem to be rare and expensive these days.

>>569224
OP here, we're actually making the switch from Twinmotion to UE4, because we are doing huge scenes and Twinmotion just can't keep up and there's no way to optimize anything..

We did a 50k square meter retail store and it's just no fun to present that at 10 FPS.

But populating areas is amazing in TM. We're looking to create similar tools in UE4 but with the lack of females it's a shitload of work. I'm really excited about Twinmotion 2018 but I doubt it can handle our large scenes much better desu.

What do you think it would cost to hire a studio to make a nice female pack?

>>569232
Yes the performance is a big issue with TM. My company mostly works on large urban developments, hotels, resorts etc and although it's great how easy it is to get projects in and flying around them looking 'good', the performance can just be too shitty most of the time.

I want to get into unreal more but the UV map channel stuff is annoying especially when some models are coming from sketchup. Have to put everything in 3ds max first I just can't find an easier way to deal with that.

you have no choice but to commission a real character.

you can flip assets all you want, but characters are not something you can afford to fuck around with.
and archviz people have supbar assets to begin with so i don't expect them to have good characters.
no offense, but having a cinematic level character quality is the only way you can get realism in your project

>>569236
the biggest problem with TM imo is that you simply can't use lights other than the sun without FPS crashing to <10 in anything larger than a small house. I'm guessing TM18 will do better with that but I can't imagine it handling large scenes because they will probably not include any optimization options.

We're just doing our first project in UE4 so I can't really say anything about UVs right now. But we're working on a Grasshopper script to sort of make it easier a bit.

It will be a huge challenge especially in the future, when most 3D models are going to come from something like Archicad or Revit, that have nothing to do with CG at all.

>>569237
it depends on the project

When you're doing a large retail store or urban scene, you can get away with using generic characters because you're not going to see them up close. But in general I agree with you.

Thats why I dont understand why there is nothing like this on the marketplace. A pack with decent male and female characters, a couple of children and animals would sell like hotcakes imo. Give them editable clothing, animations and props and I'd probably pay 2k for this.

>>569237
Sorry but you are way off. When doing large areas like a resort, you think the client cares what the girls earings look like or the colour of her eyes even? No. Just a bit of variance is fine as long as you have the variety of people fleshing out the project.

For closeup promo shots sure, you want the detail, but if I could get Twinmotion quality models walking around easily in UE4 I would pay for it.

I'm in archviz, not character modeling. People in scenes help portray the vision to the client but are not essential or important, thats why even top firms still use photos of people in their still renders.

>>569239
2k easily we would pay for that, especially if it was as simple as just dropping them into scene.

>>569239
because there is no demand.

most of the 3D market is based on fantasy bullshit so you won't see people selling realistic human characters other than base meshes.
people want fantasy crap so they can put together their asset flip quickly and sell it on android market.
if you commission 2 realistic characters (male and female) with optional heads and clothing then it might work. but its also gonna cost a fuckload of money

>>569242
Honestly how much would a commision cost for what we are looking for here?
>>569241
>>569239


For enough variety, if they were designed well enough for literally just drop in with idle and walking anmiations baked in I could probably convince my company to go for up to 5k for it. Maybe more but I don't know what the going rate would be.

>>569243
>Honestly how much would a commission cost for what we are looking for here?

i can't really say because i only commissioned hard surface to other people.
but lets say: a base mesh is couple hundred dollars (or a hundred without the head).
clothing: business/casual + variations will maybe add another hundred or two if its fully textured, could be more depending on how complex the outfit is

heads might be the most expensive since they include hair and good surface detail. a head might cost a hundred or so too.
its important that everything is done by 1 artist.
also a good artist will usually give you instructions how to use the shader or how to tweak the shader based on the cloth or the skin the character has. he might provide a unity/ue4 shader if you are not just planning on using the standard shader
each character should be about 20-40k in polygons with 4k texture maps if we want to get very upront in detail

>>569245
and for the rig + animation you should get an animator unless its just idle animations. you don't want your characters to move like uncanny puppets

>>569242
I think there is demand but it's not been realised yet. Of course the market is smaller than fantasy bullshit, but it's there and it's growing.

There is no comparable software to create realtime archviz, and certainly none to create interactive VR scenes and all that jazz.

>>569243
I'm close to buying the personnel pack I posted in the OP, and considering this one to go with it:
unrealengine.com/marketplace/gr-female-heroine

I guess you could play around with this for a few days and create a pretty diverse cast of females. Throw in some other animations (there are nice packs on the marketplace for civilian casual animations) and it would be ok I guess.

But it's work I don't really want to do as an Archviz studio.

There's this company here in germany called CADMAN, they're doing videos for real estate firms. Idk if they're using UE4, probably not, but in their videos they're using real people that have been filmed. But idk if you could implement this in real time somehow? Or use something like Golaem?

There's also firms like AXYZ who are selling 3D people, but they're not rigged for UE4 afaik.

Alembic you stupid fucks.

>>569250
What is it

>>569250
are you suggesting to create animated people and set up their paths and animations in another software and then import that into UE4?

>>569285
I am not suggesting anything, i am telling you that UE4 reads alembic files.
You have to use your brain and figure out the rest.

>>569288
so you're a useless cunt then

>>569289
I gave you the missing piece of the puzzle, if you can't finish the job for yourself you are the useless cunt because your brain isn't able to solve easy problems. Maybe you should look for a different job.

>>569342
Or maybe alembic isn't the solution to our requirements.

>>569212
If you are willing to put in a couple of days into it, you can learn to use Epic Games' "ART" or "Animation and Rigging Toolset" for Maya. it's very easy to use and any model you rig will have the same hierarchy as the SK_Mannequin making for easy retargeting.