Відео

I believe NVIDIA Broadcast only works with specific graphic cards. You can find the System Requirements here: www.nvidia.com/en-us/geforce/broadcasting/broadcast-app/

Yes! Nvidia Broadcast is free.

Interesting. In this video, I’m using the Razer Kiyo Pro webcam.

Hi Kingsley! Last year, I worked on a much more elaborate course for Faceware in collaboration with Unreal's Online Learning platform. It went live in November, but I presented it just today during a webinar. The VOD should go live with a week or 2, but in the meantime you can check out the course itself: www.unrealengine.com/en-US/blog/online-course-metahuman-facial-animation-with-faceware-analyzer-and-retargeter And in case you missed it, I also hosted a previous Faceware webinar to present the facial animation pipeline that I helped develop for "Marvel's Guardians of The Galaxy": ua-cam.com/video/_jZe3KgO61Q/v-deo.html

@@echoperformances wooo thanks

My pleasure! I hope it helps you with your own projects!

Hey! In my first reply I mentioned my most recent webinar. In case you wanted to check it out, here’s the recording: ua-cam.com/video/4x15o_ik0Qg/v-deo.html

Hello, thanks for your message. The first time you introduce Retargeter to a new character rig, you can either create poses using the Character Setup’s Expression Set (50 default/generic poses), or you can create custom poses specific to your actor’s performance. You can start with 5-10 poses per face group, retarget to fill your timeline, evaluate your animation, then add corrective poses where needed, and repeat the process. Don’t forget to hit the “Update” button to register your controller values in Retargeter before hitting the “Retarget” button. For your initial poses, I recommend adding poses on the same frames that you had created Training Frames in Analyzer. It’s not required, but generally you’ll find most of your extreme or most defined expressions.

@@echoperformances Thank you so much

My pleasure! Good luck and have fun!

Hi William! Last year, I worked on a much more elaborate course for Faceware in collaboration with Unreal's Online Learning platform. It went live in November, but I'll be hosting a webinar on February 4th to talk about it (see links below). Make sure to tune in if you'd like to post your questions live during the Q&A portion. In case you missed it, I had hosted a previous Faceware webinar to present the facial animation pipeline I helped develop on "Marvel's Guardians of The Galaxy": ua-cam.com/video/_jZe3KgO61Q/v-deo.html Unreal Online Learning - Intro to Faceware: www.unrealengine.com/en-US/blog/online-course-metahuman-facial-animation-with-faceware-analyzer-and-retargeter Faceware Webinar announcement Tweet: twitter.com/FacewareTech/status/1613975237665619994

It's a bit early to make a final statement on which combination is better. I believe using Portal will be beneficial for a large volume of videos to track, but that Analyzer will give you more accurate tracking whenever a particular performance justifies the manual adjustments. If you'd like to learn more about the Retargeter workflow, you can check out the following course that I published on the Unreal Online Learning platform: dev.epicgames.com/community/learning/courses/lw1/unreal-engine-faceware-retargeter

@@echoperformances wow, thank you, i dont know about its course, its exactly what i need)

It seems it went a bit under the radar without much promotion. It’s actually the second of a 2-part course. The first part covers Analyzer. If you find it beneficial, feel free to share it with others any way you can.

@@echoperformances you need upload this on youtube, because many people dont know about stuff on site unreal, or cant find. Its gold knowledge you create.

Unfortunately, the course was built with Unreal’s online learning team, so it doesn’t belong to me. While it’s great their courses are free, it’s up to them to promote them.

Correct! The video is tracked using the new Faceware Portal, then animated using Faceware Retargeter in Maya.

To get prices for Analyzer and Retargeter, I recommend you browse the options here and contact a Sales Rep to obtain a quote: facewaretech.com/pricing/

@@echoperformances i will do that and see if its affordable as a freelancer. Just wanted to make sure that analyser + retargeter are all i need to follow this exact workflow!?

Correct! This tutorial covers Analyzer and Retargeter. Faceware has recently introduced their neural-net cloud-based tracker called Portal. This removes the need to manually track your footage, however it’s still relevant to understand how Analyzer works. Other than that, you’ll need a license of Autodesk Maya. You can check out their Indie pricing here: makeanything.autodesk.com/maya-indie

Hi David! I’ve always used a control rig that drives my blendshapes and/or joints, but in theory, Retargeter’s character setup can drive any animatable parameters (transforms or otherwise). In the character setup window, you can try selecting the mesh that has the blendshapes and hitting the “Update” button to display its blendshape parameters.

Hi POKATILO, apologies for the late reply. The step you're describing is the Character Setup. This is the process in which we register our facial rig's controllers and assign them to their corresponding face groups. I've been working on a course that goes into a lot more detail about the entire Faceware workflow, including the Character Setup. I expect it should go live in a few weeks, but in the meantime, you can check out this video that should help you get started: ua-cam.com/video/_VrxWvIb0Rc/v-deo.html

@@echoperformances Thank you! Now I have studied FaceWare more and I can say that I asked the question incorrectly. I am more interested in how you set up the Metahuman poses in Retargeter to better match the video reference, do I understand correctly that you moved the controls manually from the Meta-human in Maya or did you use ready-made meta-human poses from Unreal?

Very useful insight. Thanks! To be honest, the reasons you've provided are why I've hesitated to try out FG thus far. I've been a long-time user of Retargeter and I suppose I've been fortunate to not encounter crashes that often, granted I'm still working in Maya 2020. For what it's worth, I generally work on facial animations in a standalone scene with only 1 character, no body animation, and no environments to keep things as light as possible. When adding, copying, or updating poses in Retargeter, best to avoid undoing the previous operation with CTRL-Z. I would say that's probably my most consistent crash. I hope that helps!

Hi Toby! Thanks for the comment. It can definitely be frustrating to encounter bugs and crashes along the way. My best advice would be to save your scenes often and regularly hit the "Update" button in Retargeter to store your changes independently of your scene.

Those are great suggestions. Facial animation is really my expertise, but I’ve started exploring different options for body animations. I’m hoping to incorporate some of them in my next projects rather than keyframing.

Hello Daniel, if you visit our channel, under the "About" section, you’ll find an email address where you can send your inquiries.

Hi Belal, thanks for stopping by. I work primarily with creating facial animations, including the pipelines and tools that support it, so I would need to explore Mixamo before I could create a tutorial for it. Have you found any particularly interesting videos that you would recommend?

@@echoperformances I will be happy if you make a detailed video to use facware because I'm having some problems with it..and about animating the body, you can watch these videos, some of them are related to the topic ua-cam.com/play/PLlJ0LuZkvf-XueBlSDGPhpMmX7pV3SczH.html

@@B.BAKDASH Thanks for the playlist! I'm currently working on a more detailed and more "official" Faceware course that should be available in March. Keep an eye on my Twitter page for more details as the date approaches.

To be honest, I haven't met any Faceware users working on a mac. Perhaps their support team could recommend you an alternate solution to get their products running on a mac. You could enquire by contacting them through this page: facewaretech.com/contact-us/ Thank you for your comment! I hope that helps.

@@echoperformances Thanks a lot for your recommendation ! Gives me the next steps ! :)

Precisely. In fact, with Retargeter you can continue to add as many in-between poses to improve your transitions and iterate as many times as you want to fill in the blanks. After you feel your animation is satisfactory, you can edit the curves directly to iron out some jitters or fine tune some subtle details. You could also add an animation layer on top for your changes, just keep in mind that Retargeter only evaluates keys on the base layer.

@@echoperformances That's awesome. I am working on a pipeline/solution that allows me to get base first pass animation from our Facemotion3d ios app to metahuman rig inside maya. After the firsr pass I want to apply more details on visimes and other suttle nuances and tongue anination via layers then merge it before expirtibg. Your Faceware animation is one if the best I have seen. :) Would love to show you my progress and also integrate Faceware pipeline into my workflow. Are you subscribed to the pro-version of analyzer and retargetter? I would really not want to not use the storable training frames function even if I have to pay close to four times the amount compared to the base version!

@@facemotion3d971 Thanks for the praise! What you describe in your first lines is "spot on". Regardless of the solver (Faceware, ARKit, ...), that approach will give you a solid foundation, but if you're looking for high quality facial animations, any automated method won't get you 100% of the way there. A polish pass is always recommended, which includes the tongue if it's not part of your solve. To answer your question, I am using the pro-version which offers batch support for Analyzer and Retargeter. This includes the ability to export Tracking Models from Analyzer and Shared Poses for Retargeter. If you're looking to experiment with these features, I would recommend you look into Faceware's PLE license (Personal Learning Edition).

@@echoperformances The PLE faceware studio which I already tested is not same as analyzer n retargetter is it? Didnt like the quality of the studio. I think there is a trial version of analyzer n try them. :)

@@facemotion3d971 You're correct. Faceware Studio is their real-time solution, whereas Analyzer and Retargeter offers a lot more customization. The batch feature I mentioned in my previous reply is mostly beneficial for large volumes or for repeat use of the same actor/character. If your project justifies it, you might want to consider the cost of the higher license with batch support versus the cost of an animator setting up each job "from scratch".

Absolutely! What you're describing is called a "Tracking Model" in Analyzer. Storing your training frames for each of the face groups means you can benefit from your previous poses in future tracking jobs.

@@echoperformances Thanks for response. Faceware announced price decrease for the Faceware studio live for indies. However the studio version has lower overall quality compared to the offline version. what do you think?

Both options cater to different teams' needs, depending on the scope of the production and its animation pipeline. Faceware Studio comes at a lower cost for Indie developers and can provide very quick results with no tracking or pose training required. Using Unreal Engine's Take Recorder, these animations can serve as a decent baseline for blocking and timing, then they can be edited/polished afterwards. On the other hand, the "offline" approach of Analyzer and Retargeter requires an upfront cost to set up your actors and characters, but offer a lot more fidelity in the performance. They can also be batch processes to handle large volumes of animations for bigger productions.

@@echoperformances Thanks for quick response. The offline option costs ten times more than the studio version but it provides much better results seeing from your examples. :) The studio live version is lesser quality than the iphone arkit tech I think, but it does have asymetrical brows which arkit cannot do. I am also looking into Facegood's Avatary and their infrared HMC which along with the software provides very high quality that rivals or even surpasses FW offline results at a much lower cost. They are also coming out with a ML live version that is very high quality as well. Do you have any more recent FW works youve done? I really like your work

@@facemotion3d971 Once you start digging into the multiple options available in this field, you'll find that a lot of them have similarities. It boils down to each user's personal comfort level, project scope, pipeline scalability, and output quality expectations. On my end, I'm working on a new short film, an online course, and a conference presentation. Juggling a lot, but I hope to post new content to the channel in the next few weeks.

In Over 40, I think putting a microphone in front of her face and giving her headphones like in the video would have made the performance more believable. In your video it looks like she's in a live interview setting and her eyes don't seem to be looking at the interviewer. As a result, the eye movements just seem random. What do you think?

@@pointandshootvideo Thanks for the feedback! I really appreciate your notes. It's funny how either character seems to appeal more or work better for different people. I can't put my finger on it either. Maybe it's a matter of relatability or familiarity of the model? I'm currently working on another short film with the same 2 actors. I'll probably give them new models to better fit the story. The biggest thing I learned from working on "Old Scores" is that animating the skeleton directly in FK was a bad idea. A proper character rig will go a long way. There are a few autorig scripts built for MetaHumans available for pretty cheap. I'll also try to get actual body animators to lend a hand this time around. As for the mic idea and the wandering eyes in "Turning 40", I completely agree. Trying to replicate the performer's original environment will definitely help to sell the final result.

Your team is definitely on my radar. It would be fun to experiment with it soon.

Oh. But then I get stuck on the character setup file (which I found in the comments). Can you go over how to create the character setup file? Thanks!

Thanks for the feedback! I’m glad you found the tutorial helpful. And thanks for the suggestion to make a tutorial specific to the Character Setup. I’ll try to find some time to record one.

That’s great to hear! Your feedback is really appreciated.

Thanks for the comment! Feel free to share this video and the channel with others. Perhaps I’ll create a more in-depth tutorial specific to create a Character Setup file. For now, let me know if this one works for you: bit.ly/MH_FWR_CharSetup_01

@@echoperformances Woww. thanks so much bro for this I really do appreciate this. I will try it out. Please if you don't mind how may I reach out to you, I would love to discuss somethings with you. Please I would really love for us to connect I wanna have a discussion with you. thanks and God bless Bro

@@matthewisikhuemen8907 In the "About" section of our channel, you can find an e-mail address to contact us with your questions.

@@echoperformances Oh thanks so much, I will send an email. thanks once again, I really appreciate

Thanks for inquiring. Here you go: bit.ly/MH_FWR_CharSetup_01 Let me know if that works for you.

@@echoperformances Thank you very much . You've been really helpful.

Glad to help! Let me know if you encounter other blockers along the way.

I’ve never tried iClone or Daz, but I could add those to my list of future projects. Thanks for the suggestion! As for publishing your work with MetaHumans outside of Unreal, I believe Epic offers the options to license MetaHumans to be used outside of Unreal. As long as it’s used within Unreal, it comes included for free.

Ah, good to know! Thanks for the reply!

Hi Arvind. That’s an interesting proposition. I haven’t thought of experimenting since I work primarily in Maya. If you do get around to testing it, let me know.

Thanks for the feedback! I really appreciate it.