Thank you for bringing these issues to our attention. We'll take a look at both the Danny avatar and the stuff happening with the recording process when loading a video source.

Hello Wolfennar,

Regarding the recording issues the programmers are looking into it.



Regarding the transparency issue explanations is below:

the TL:DR version is: Please put Danny's default scarf back around his neck, (it is there for a good reason :) )


The long version :

Rendering good looking fur in real time in notoriously difficult. Our method, which is designed to work even on low spec machines and dx9 cards is quite optimized but comes with some technical drawbacks when you have different furred meshes from different drawing passes overlapping/intersecting, which is what you are seeing.
In your case the head/neck fur overlaps his chest fur or viceversa , and these are different meshes hence the "seams". So that's why his scarf was there, to address this problem.

Now there are other solutions aside adding a scarf over the intersecting area , some of them solve the problem entirely, other just make it less obvious ( not even the scarf is a perfect fix mind you) , though we do have to keep in mind the financial cost of those solutions, and currently placing the scarf over that was the solution that we could afford.

You could say : well just make all problematic furred areas one single mesh . However here we came across a "maximum bones per mesh limit in DX9 mode " imposed by the game engine that we are using and the number of constant buffers available for vertex shaders on Dx9 (yes we do GPU skinning, so all bone matrixes need to get in the vertex shader). Because we wanted FaceRig avatars to also work for people with weaker specs ( DX9 cards, lower specced CPU's) , and at the same time have a high number of bones for face animations and not force dx11 as a minimum requirement- > the affordable solution was to cut the mesh up in smaller bits, and draw those bits in separate render passes, though this was the price we had to pay .


You could also say : Why not create different versions of such problematic avatars designed and built differently for dx11 and dx9? Why not spend resources on investigating other technical solutions?

A: We want to be able to do this, but currently taking that avenue is a financial risk for us because of the costs involved.

If we have to pick between tasking an artist to build a Dx11 optimized Danny, that does not produce sorting side effects in some very particular conditions (& a programmer to develop a different fur rendering technique all together, specifically for the audience who was high powered dx11 cards) , or tasking that artist to create a brand new avatar that a larger public can enjoy playing with ( let's say the Engineer from TF2 that works just as well on dx9) , we have to go with creating the new avatar, because the extra attention added by that free avatar helps us sell FaceRig copies, and that revenue help us pay the artist. Kinda the same for programmers and what they are working on.

Please remember, that we are not rich we are an an independent studio: we have to sell enough copies to stay in business, and to balance monthly income and expenses. We have kept the price low for the program so everyone can afford this, and we are adding free avatars, instead of putting them in paid DLC (even though those new avatar cost money to make), we do not charge any type of subscription fee, but that approach means our revenue is lower than that of companies who nickle and dime the heck out of their audience.


Remember that from each sale we pay royalties to Steam and to external tech partners. Then we pay taxes, and in the end we are left with... not so much for salaries, office space rent, software licenses and everything :). That's why we keep telling everyone on social channels, that if they like our effort, please share it around to their friends. Every share and every sale counts for us and enables us to keep FaceRig growing with more content and more features , until , hopefully, its output will be indistinguishable from reality :D and the free avatar pool will be diverse enough to satisfy everyone :).


There are so many things that we wish we could afford now to make FaceRig better, faster, but unless we cave in on the indie dream and go the Venture Capital investment route (that usually comes with a lot of conditions and most likely a price hike on the program and the end of free avatar addons) we will have to be patient and get there with smaller steps. The speed at which it can evolve and still remain indie, is directly linked to the sales of the program on Steam :D.


We will be able to put resources in re-assessing those tech choices for furred avatars a bit later. Right now we have as primary mission to make sure we are financially sustainable so that we can finish implementing all the features that we have promised in the crowdfunding campaign. Top of that list: releasing the Studio beta version, extra sensor support, and then there are also new platforms: Mobile versions and Mac version, non Steam version for Desktops and so on :D.

So that's why for now, unless you are considering re- building the avatar yourself to get a dx11 specific version that corrects this short-coming (which is by the way totally doable if you know modelling and animation :P) , putting the scarf back around his neck is the only recommendation we can make :) . After we are done with mobiles, Mac, extra sensors, we will get to building dx11/dx12 -specific features for FaceRig, and by doing that we will most likely have a different solution for a bandana-less furry Danny :D.

Thanks for the very in depth explanation, while it wasn't really needed it's still good to know all that. I unfortunately don't know modelling and anim, but the transparency issue is only visible in some limited settings and there's always the scarf if necessary so it's not that bad. Let's hope the sales will be good and that the recording issue is relatively easy to fix.