Analyze facial expression and movement from video
Data can be used in most CG, VFX, or AAA production pipelines
Easy-to-learn process with free training resources
Create highly automated workflows for large volume demands
General Workflow for Analyzer 3.0
A new "job" is the term for a tracking project in Analyzer. An Analyzer job consists of two separate things:
- An .fwt file (FaceWare Tracking) that serves as the main project file for your job. This is the file you'll select when you wish to open an existing job.
- A folder that shares the same name as the job with "_fwt" added to the end. This folder contains many of the working files for your tracking job, including the image sequence that you are working from.
The way one gives the software information about the actor’s performance is to “markup” frames that have extreme and distinct expressions, creating training frames. These frames are then processed by the software to allow it to hit all of the in-between expressions. The user input is moving points (called “landmarks” in Analyzer) around the face in a specific way.
The main purpose of Faceware Analyzer is to track facial expressions quickly and accurately so those expressions can be applied to facial animation using Faceware Retargeter. Tracking lets the user create the shapes for a particular actor, eventually ending up with a model that reflects the actor's expressions. This increases the accuracy of the track and does a better job at picking up the kinds of small movements that really help to sell a facial performance.
Our export process is called “Parameterization.” This is the process through which tracking data in Analyzer is converted into a form that can be used to produce animation with Retargeter( a .fwr file). Parameterizing is the final step in the tracking process which combines the four processes of Parameterization.
Once parameterization is complete, Analyzer will have created a .fwr file in the same location as the .fwt file and project folder. If your video was 30 seconds long, your corresponding .fwr file will contain 30 seconds of motion data.
This .fwr is the performance file that will be loaded into Retargeter.
Features of Analyzer 3.0
Expression and texture tracking for complete facial motion analysis
Analyzer is a high-quality production software that tracks facial movement from video using machine learning and deep learning. It uses a markerless technology to track everything the face can do at an extremely high quality.
Robust toolset with intuitive workflow
Looking for the the premiere facial tracking software used by professionals around the world? This is it. Faceware products have been used on thousands of ground-breaking projects and we’re constantly making them better. Analyzer is simply the most powerful tool available for professional facial tracking.
Eyes are the key to emotion and performance
Eyes bring out the emotion and believability of any performance. Our advanced pixel-tracking technology provides extremely accurate eye and gaze tracking, further reducing the barrier between your performance actor and their digital counterpart.
Patented & best in class technology
Our technology is built from Image Metrics consumer-grade technology meaning Analyzer can work on nearly any face in nearly any lighting condition. Enjoy exclusive features built from millions of facial performance images. Our patent portfolio is entirely homegrown from nearly two decades of practical experience.
FREE easy-to-learn resources to get up to speed quickly
Our support team is second to none. Check out our huge gallery of video tutorials and online knowledgebase to get you started and teach you best practices for tracking faces in Analyzer. You can also work directly with our Support Team to get the best results.
Distinguishing Features of Analyzer 3.0
Parameterization is a process that extracts additional facial movement from our feature tracker by sampling every pixel for every frame of video that is analyzed. Without any additional user input, Parameterization allows hard to capture areas such as the upper cheeks, lower eyelids, and jaw position to be tracked with extreme accuracy.
Have a large amount of videos to process? Analyzer Studio Plus has features to export the statistical tracking data you create on your first video to to help track any additional videos with the same actor. This allows for significantly large-scale projects to benefit from our machine learning technology; particularly if there are a large number of shots with the same actor.
Strapped for resources? Analyzer contains an entire library of batch API commands and arguments to access nearly any part of the Analyzer workflow. This library allows users to create semi and fully automated workflows. Our largest users process tens of thousands of seconds of facial animation without any user input on a shot-by-shot basis.
Visit the Knowledgebase to learn more about analyzerbatch commands and arguments.
We know a lot about tracking faces and are constantly looking at ways to improve our workflow. Creating tracking data is a breeze with a feature called Intelligent Drag. Once one landmark tracking point is correctly placed on the face, all other tracking points can be quickly placed on the face. Combined with a smart approach and Analyzer's clean, simple workflow, you don’t need to be a scientist to get professional facial tracking.
Getting started with Analyzer
Explore the many free resources to help you learn Analyzer.
FAQS
Commonly asked questions about the Analyzer workflow and best practices for use in productions.
What resolution of video do I need?
We recommend a minimum 720p image for high-quality results, however it's worth noting we do not recommend 4K image files to achieve photoreal results as they add unnecessary processing time without adding to the emotion analysis.
Analyzer does support all standard video resolutions and aspect ratios. Non-standard video resolutions (e.g. 720x496 vs. 720x486) may result in odd behavior in the application. If you experience trouble using a non-standard resolution we recommend re-exporting your video to a standard aspect ratio and size.
Will applying tracking markers to the face help Analyzer track more accurately?
The answer to this question is slightly complicated. Analyzer, without AutoTrack, relies on the accuracy of the training frames created to perform its tracking functions as smoothly and accurately as possible. Adding ink markers to the face does not inherently increase tracking accuracy, but markers can be added in such a way as to aid the user in creating the training frames. Placing the markers in the same positions as Analyzer landmarks are laid out while tracking in the software can help the Analyzer user place the landmarks in the correct spot on each and every training frame they create, leading to faster training frame creation and more accurate tracking overall. When creating the training frames, the user can see exactly where each landmark should go and can use the ink markers to place the landmarks quickly while still maintaining the consistency needed for accurate tracking.
What video codecs are supported by Analyzer?
Almost every video codec on the market works with Faceware Analyzer, so to make things easier we’ve put together a comprehensive list of preferred, production-proven codecs that have a strong balance between quality, file-size, and speed. These work well with both our ProHD camera as well as any standard video camera.
Whenever possible, it is recommended to use the .mov wrapper for your movie files.
- H.264
- Apple ProRes
- Photo-JPEG (75% [High] Quality)
- Motion JPEG B
- DV
- Avid DNxHD
For legal reasons, there are some proprietary video codecs that we are unable to support:
- GoPro Cineform
- G2M4 (GoToMeeting)
Do I need to be connected to the internet to use Analyzer?
For the Studio Plus version of Analyzer, no internet connection is required. For Studio, Trial, and PLE versions, an active internet connection is only required during the initial credential check whenever the software is launched. During the actual tracking process, no internet connection is required.
How long does it take to learn Analyzer?
We have found from our direct experience and feedback from our Analyzer service providers that users can start to track a shot in a few days, but plan on about a week to get new artists completely up to production speed.