I think you need to make to be more easily to use the software for people who wear glasses！（I mean I wear glasses，and the facerig is hard to catch my face！！）

Add support for VRM models if possible. Or develop a converter that makes the VRM format into Facerig's one.

desperately needs more active support for extra features. I personally had to switch over to Crazytalk v8 pro, even though I really prefer the facerig avatars

Oh. I have FaceRig! I wonder if they fixed that thing where it couldn't detect my mouth properly because It's naturally frowny.

Nice survey, doesn't give any box to comment about anything lol.
I can't use this since I moved to Linux(~4 years ago), doesn't work with Wine/Proton, Linux support or at very least Proton support would be great.
Also having better API would be great, so we can process our own video buffering and detect hands without needing to buy a $400 camera

The windowed mode acts funky. So hard to adjust. Was never like this. Noticed this a couple months ago.

Thank you for opening up to feedback. I like the idea of FaceRig, and the technology behind it seems to work well.

I'm a professional animator, so I deal with the 3D character pipeline every day. Creating custom avatars using the import tool is impossibly difficult. It gives almost no feedback for errors. In many attempts over a number of years, I've gotten a total of one custom avatar working. But trying to reimport that avatar using the same set of files only caused errors. The tool is so prone to errors that I can't properly test and use a normal development loop to create an avatar, only guess and hope it somehow works. Anything that does work is too fragile to try adjusting, and the same thing might not even work if tried a second time. This is all after a very specific and tedious modeling, rigging, and animating process to meet the minimum requirements for the tool to even attempt to import an avatar.

It's beyond frustrating and ultimately not worth using. I know some people have figured it out, but it should not be this difficult. I want to create avatars as a hobby, but the difficulty curve is much more than that of a hobby.

I believe Unity support was in development a couple years ago, and that seems ideal for setting up custom avatars. The interface is clean and visual unlike the existing FaceRig import tool, and error handling and feedback is already there. Rigs don't break entirely if one animation is missing or broken. There is a lot of room for custom animation controllers so creators can pick and choose how much detail to provide in terms of clips - maybe they just flap their mouths open and closed, or maybe their character needs every visime and subtle facial expression. VR Chat already uses this kind of Unity integration for their custom player avatars.

I would suggest prioritizing Unity development if you want to encourage custom avatars and wider adoption of FaceRig. I think the market for avatars is there in YouTubers and streamers, but the difficulty makes a FaceRig avatar unavailable to them (even through a qualified third-party artist).

At the very least, the import tool documentation needs to be greatly expanded. It can not just tell us what is required, but you must go through error scenarios in order to train us on what not to do. Tell us what will cause errors, how to prevent them, and how to fix them. As far as I understand, I've followed the manual to the letter, still gotten errors, and have no way of knowing what could have possibly caused them. If you stick with the current import tool and customization process, please have professionals create educational videos on avatar creation.

I hope that is helpful even though you didn't solicit that particular feedback.

Avatar creation needs to be streamlined or more lienient during importing process.

Support for shapekeys/blendshapes instead of bones would be really really great

Avatar creation for the tool is too confusing...

There's a thing I've been using and maybe you guys could look into it:
https://vrm.dev/en/

I've been using it for my own solution (uses Unity or Unreal and OpenCV/Dlib) and VRChat for character distribution.

Survey summary: Give us personal info to possibly sell to spammers, lol?
:/
That's it. Literally it.
:/

But to ACTUALLY give feedback... ditto on the glasses problem. I can't use this whatsoever due to problems it has with my glasses. That's completely understandable, but there's GOT to be a workaround.

My idea: Maybe there can be buttons on the keyboard assigned to certain eye expressions, with the option to either hold or toggle. S for Sad, A for Angry, H for Happy, W for "Wow!", etc. Or other assignable keys according to whatever the user feels best using.

That might require a hotkey window in the foreground so your computer doesn't accidentally type somewhere. Or maybe it can be bound to game controller buttons for those who have one.

Or there can be hover-buttons over the camera interface so one can move the mouse over the eye expressions they want.

Maybe this can even be expanded to mouth, or full-combo expressions too, as some folks have differently shaped faces (malformations, RBF, perma-frown, etc). People can balance what they can control fine with their own faces, and what they might need additional help with.

Maybe we can have settings to manipulate the full expression for each emotion preset, which will then be activated with the keys/hover selections/whatever activates them.

Anyway, that's what I would have suggested if that had been an actual survey. I hope something along these lines can be considered, to make this application more accessible to everyone. I'm sad that I spent the money on this but still can't use it even after all this time waiting for improvement.
:(

Honestly, my two biggest complaints about FaceRig during its lifetime - and the two primary reasons I lost interest in it so quickly - is the user-hostile, borderline threatening license agreement for it and its relative lack of modability. Sure, there's a couple of tutorials here and there, but no real resources or collections to choose from.

Ditto on the requests to make importing custom models easier. As others have said, this is the main reason I stopped using the software.

Not sure if it matters, but there is no way to set up a heavier set avatar. I've had better luck using free online apps (except you can't use them in youtube videos). This is a great start, but considering it's been around for over 5 years, it's pretty sad. I'm glad I only paid 6 bucks for it.

Originally posted by Bees of War:

Thank you for opening up to feedback. I like the idea of FaceRig, and the technology behind it seems to work well.

I'm a professional animator, so I deal with the 3D character pipeline every day. Creating custom avatars using the import tool is impossibly difficult. It gives almost no feedback for errors. In many attempts over a number of years, I've gotten a total of one custom avatar working. But trying to reimport that avatar using the same set of files only caused errors. The tool is so prone to errors that I can't properly test and use a normal development loop to create an avatar, only guess and hope it somehow works. Anything that does work is too fragile to try adjusting, and the same thing might not even work if tried a second time. This is all after a very specific and tedious modeling, rigging, and animating process to meet the minimum requirements for the tool to even attempt to import an avatar.

It's beyond frustrating and ultimately not worth using. I know some people have figured it out, but it should not be this difficult. I want to create avatars as a hobby, but the difficulty curve is much more than that of a hobby.

I believe Unity support was in development a couple years ago, and that seems ideal for setting up custom avatars. The interface is clean and visual unlike the existing FaceRig import tool, and error handling and feedback is already there. Rigs don't break entirely if one animation is missing or broken. There is a lot of room for custom animation controllers so creators can pick and choose how much detail to provide in terms of clips - maybe they just flap their mouths open and closed, or maybe their character needs every visime and subtle facial expression. VR Chat already uses this kind of Unity integration for their custom player avatars.

I would suggest prioritizing Unity development if you want to encourage custom avatars and wider adoption of FaceRig. I think the market for avatars is there in YouTubers and streamers, but the difficulty makes a FaceRig avatar unavailable to them (even through a qualified third-party artist).

At the very least, the import tool documentation needs to be greatly expanded. It can not just tell us what is required, but you must go through error scenarios in order to train us on what not to do. Tell us what will cause errors, how to prevent them, and how to fix them. As far as I understand, I've followed the manual to the letter, still gotten errors, and have no way of knowing what could have possibly caused them. If you stick with the current import tool and customization process, please have professionals create educational videos on avatar creation.

I hope that is helpful even though you didn't solicit that particular feedback.

I *absolutely* support this idea --standardizing on Unity's pipeline.

I use it regularly in my workflow for building VRChat avatars using Daz 3D, Blender 2.8 + CATS Blender plugin, and Unity + the VRChat SDK (my technique is a modification of the techniques taught by Fruitpex on Youtube: https://www.youtube.com/watch?v=UXGX3-a6OGc )

I like the ideas behind Facerig and really want to use it more, but the hurdles I face when trying to import custom characters in FBX standardized Unity format is a serious problem as far as I'm concerned. Facerig needs a Unity plugin or SDK module that will allow a VRChat-ready model to be directly ported to Facerig with no changes necessary to be really successful as far as I see it --this is actually something that would make the app take off like a rocket.

not work cameras C920 :/ since many week /: bad