- What is MotionPortrait?
- MotionPortrait is a technology that can create a variety of facial expression animations from a single image. Please check "MP top page" for a demo and more information.
- Wow! But what can I use it for?
- The possibilities are limitless. Please check "Showcase" page.
- How many pictures are needed?
- A single photo is all you need.
- Do I have to use a special camera?
- No, you can use an ordinary digital camera. You can even use the camera on your smartphone.
- This is a pre-rendered video clip, right?
- No. This is real-time rendered computer graphics.
- How long does it take to create the animation after taking a picture?
- It normally takes only a few seconds on a personal computer.
- Can I create an animation that makes a model turn to the side or face the other way?
- It is possible to make the model turn its face to some degree. The image looks natural within ±20 degrees. However, since we have no image of the back of the head, we cannot provide full rotational freedom without it seeming unnatural.
- Can I automatically make a model from a non-human?
- Our technology is tuned to work with human faces, and is not readily capable of handling non-human subjects. In such cases, models may be created manually, and the animation process handles the rest automatically.
- I want to use this technology right away...!
- You can try our face modeling, animations, and several other features at our Top Page.
If you are a game developer who interested in our technology, please contact us through Silicon Studio.
- Was MotionPortrait originally a Sony technology?
- Yes. MotionPortrait, Inc. was founded by members who originally developed the technology at the former Sony Kihara Research Center.
- What are the minimum requirements for my PC?
- A computer with a 1-GHz CPU can create animations in real time. MotionPortrait technology should still work reasonably well with computers that do not meet these specifications.
- If I take a picture from the side and use it, can I make a perfect 3D head model?
- Theoretically, yes. However, MotionPortrait uses only one picture because of the following reasons:
a: Taking multiple pictures needs more time and effort.
b: When using models made by stitching together multiple images, the quality of animation worsens.
- Is automatic lip-sync possible?
- Yes, this is certainly possible. You can match lip movements with voice in real time using the microphone.
- I want to use it for game production...
- We offer middleware packages for developers. Please contact Silicon Studio Corporation through the link below:
Silicon Studio, MotionPortrait solution.
- Does the automatic 3D face modeling ever fail?
- Yes, but this is very rare. If this ever happens, you can easily correct the mistake.
- How smooth can it animate on a mobile phone?
- We have confirmed that it runs at approximately 30 fps with graphics acceleration and 15 fps without it.
Questions About Licensing
- Can I have a quotation for licensing?
- Please contact us with a brief description on the project, application, hardware and the usage period, and we will get back to you as soon as we can.
- How long does it take to deliver the technology?
- We can provide the developer environment and sample code within 5 business days of signing the contract.
- Which platforms are supported?
- We support multiple platforms including iOS, Android, Windows, Mac, Linux, Flash and HTML5/WebGL.
- Are servers needed?
- Both server and embedded licenses are available.
- Is embedding (into apps, car navigation, signage etc.) possible?
- Yes. Please contact us for details.
- What will MotionPortrait provide?
- Synthesizer software for transforming photos into 3D models,
and software to view and control the generated models.
We ask our customers to prepare the network, server, UI system, and design by themselves.
- Do you undertake software development?
- Yes, we do. However, generally we do not undertake such projects, and instead work on a licensing model.
- Can you introduce companies who do development using your technology?
- Yes, we can do that.
- If using a server, what kind of a server is needed?
- A Linux (CentOS or Red Hat is recommended) server with 2GHz or faster CPU, and 2GB or more memory is recommended. For video creation, a graphics card is necessary.
- Do we need a dedicated server?
- A dedicated server is preferred. While it is possible to run the MP server and web servers and the like on the same system, this is not recommended due to load balancing problems.
- What should we prepare beforehand?
MP Movie Original video for face compositing, server with graphic card, UI/System MP Narikiri Original photos for compositing (512x512 pixels or higher resolution), server, UI/System MotionPortrait Images of items, server, UI/System MP Face Analyzer Sample photo/s for analysis, UI/System
For likeness-generation solutions, designs for the various face parts
MP Avatar Photos and drawings for creating avatars
- Do you have a service for making avatars from our photos or drawings?
- Yes, we do. Please send us the photos.
- Can this technology be used in TV programs?
- Yes. Our technology has already been used on quite a few TV programs.
- Do you provide development support?
- We provide reliable support and experties. We can prepare sample code, reference materials, and even developer environments.
- Can you provide customized software?
- Feel free to contact us and let's discuss the details.