Statement of Work (SOW)

Overview

The service provides 3D digital content production for digital content innovation. Professional delivery teams provide high-quality production services to enable businesses to go digital.

Service Content

Category	Sub-Category	Service Description
Solution	3D realistic virtual human	Provides consulting related to designing 3D realistic virtual humans, develops, renders, and composites a 3D realistic virtual human model based on the design. A 30-second CG video and a set of hairstyle and clothes are provided. Portrait: The 3D realistic virtual human portrait design is output. Select one from the three designs in the draft phase. The design to be submitted includes a colored front view and reverse perspective at 45° and an orthographic view draft. After the complete version is submitted, minor design modifications can be made a maximum of five times, but major modifications are not allowed. Model: Based on the 3D portrait of the realistic virtual human, a 3D realistic virtual human model is output. This model can be used for subsequent video production, holographic production, poster production, and livestreaming activities. Application scenarios include new media, production below the movie level, realistic faces, eyeball structures, and hairstyles (using XGen), skin with physically based rendering (PBR), and live UE projects. During modeling, minor design modifications can be made a maximum of five times, but major modifications are not allowed. The fees for scanning and hiring models are not included.
	3D beauty virtual human	Provides consulting related to designing 3D beauty virtual humans, develops, renders, and composites a 3D beauty virtual human model based on the design. A 30-second CG video and a set of hairstyle and clothes are provided. Portrait: The 3D beauty virtual human portrait design is output. Select one from the three designs in the draft phase. The design to be submitted includes a colored front view and reverse perspective at 45° and an orthographic view draft. After the complete version is submitted, minor design modifications can be made a maximum of five times, but major modifications are not allowed. Model: Based on the 3D portrait of the beauty virtual human, a 3D beauty virtual human model is output. This model can be used for subsequent video production, holographic production, poster production, and livestreaming activities. Application scenarios include production below and at the movie level, filtered head structure, skin with PBR (without realistic details), realistic eyeball structures, realistic hairstyles (using XGen), and live UE projects. During modeling, minor design modifications can be made a maximum of five times, but major modifications are not allowed.
	3D animated virtual human	Provides consulting related to designing 3D animated virtual humans, develops, renders, and composites a 3D animated virtual human model based on the design. A 30-second CG video and a set of hairstyle and clothes are provided. Portrait: The 3D animated virtual human portrait design is output. Select one from the three designs in the draft phase. The design to be submitted includes a colored front view and reverse perspective at 45° and an orthographic view draft. After the complete version is submitted, minor design modifications can be made a maximum of five times, but major modifications are not allowed. Model: Based on the 3D portrait of the animated virtual human, a 3D animated virtual human model is output. This model can be used for subsequent video production, holographic production, poster production, and livestreaming activities. Application scenarios include production below and at the movie level, animated skin and eyeball structure, model hair structure, and live UE projects. During modeling, minor design modifications can be made a maximum of five times, but major modifications are not allowed.
	3D stylized virtual human	Based on the modeling of a single photo, the system outputs a virtual human model similar to the photo, with clothes and hairstyles provided.
	CG-only poster production	Creates CG-only posters based on the virtual human model, virtual background, and other objects. The procedure includes sketch design, virtual human pose, lighting rendering and compositing, and retouching. After the complete version is submitted, minor design modifications can be made a maximum of five times, but major modifications are not allowed.
	Virtual-real poster production	Creates virtual-real posters based on the virtual human model, real background, and real person. The procedure includes sketch design, virtual human pose, lighting rendering, integration with the real background, and retouching. The fees for hiring models and location shooting are included. After the complete version is submitted, minor design modifications can be made a maximum of five times, but major modifications are not allowed. The fees for location rental, actors, and travel are not included.
	CG-only video production	Advertisement production based on the virtual human model, virtual background, and other objects. The scenes are complex and contain many special effects. CG-only videos are produced and high quality is required.
	Virtual-real video production	Advertisement production based on the virtual human model, real background, and real person. Virtual-real videos are produced and high quality is required.
	CG-only virtual livestreaming	Performs CG-only virtual livestreaming using the existing assets, virtual human, and virtual background. The livestreaming lasts for no more than 90 minutes. This service is benchmarked against simple Vlog livestreaming by virtual humans, including live UE project customization, motion capture device setup and commissioning, motion capture actor interview and hiring, rehearsal, and operations on livestreaming by virtual humans.
	Virtual-real livestreaming	Performs livestreaming using the existing assets, virtual human, real human, and virtual (or real) background. The livestreaming lasts for no more than 90 minutes. This service is benchmarked against medium-sized livestreaming by virtual humans and real persons, including live UE project customization, motion capture device setup and commissioning, motion capture actor interview and hiring, rehearsal, and operations on livestreaming by virtual humans. The fees for studio setup, lighting, video recording, and director team are included. The fees for location rental, actors, and travel are not included.
	CG-only video for news broadcast	News video production based on the virtual human model, virtual background, and other objects. The scenes are simple and contain few special effects. CG-only videos are produced without high quality requirement.
	Virtual-real video for news broadcast	News video production based on the virtual human model, real background, and other objects. The scenes are simple and contain few special effects. Virtual-real videos are produced without high quality requirement.
	CG-only video for 3D object display	Advertisement production based on the 3D object model and virtual background. The scenes are complex and contain many special effects. CG-only videos are produced and high quality is required.
	Virtual-real video for 3D object display	Advertisement production based on the 3D object model and real background. The scenes are complex and contain many special effects. Virtual-real videos are produced and high quality is required.
	Sara 3D virtual human video production	Advertisement production based on the Sara virtual human model and virtual/real background. The scenes are complex and contain many special effects. Advertisement videos are produced and high quality is required.
	Yunxiao 3D virtual human video production	Advertisement production based on the Yunxiao virtual human model and virtual/real background. The scenes are complex and contain many special effects. Advertisement videos are produced and high quality is required.
	3D stylized virtual human video production	Creates a stylized virtual human model based on the modeling of a single photo. The background is simple and there is no special effect. Advertisement videos are produced without high quality requirement.
	Virtual human video broadcast (basic)	Produces advertisement videos based on the simple virtual human model.
	Virtual human video broadcast (advanced)	Refines the virtual human image and produces advertisement videos based on the refined virtual human model.
Integration implementation	3D space integration implementation	Provides integration guidance for specific 3D application scenarios.
3D materials	3D realistic hairstyle modeling	Creates 3D realistic hairstyles for realistic virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D beauty hairstyle modeling	Creates 3D beauty hairstyles for beauty virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D animated hairstyle modeling	Creates 3D animated hairstyles for animated virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D realistic clothing modeling	Creates 3D realistic clothing for realistic virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D beauty clothing modeling	Creates 3D beauty clothing for beauty virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D animated clothing modeling	Creates 3D animated clothing for animated virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D realistic accessory modeling	Creates 3D realistic accessories for realistic virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D beauty accessory modeling	Creates 3D beauty accessories for beauty virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D animated accessory modeling	Creates 3D animated accessories for animated virtual humans. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	3D scene modeling	Creates 3D scenes. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	Dynamic special effects for 3D virtual humans	Creates dynamic special effects for 3D virtual humans, such as light and shadow effects on the body.
	Animated actions for 3D virtual humans	Creates animated actions for 3D virtual humans, such as punching and bowling.
	3D object modeling	Creates 3D models of products (such as electrical appliances, locomotives, and dolls) based on the real products/design drawings provided by the customer. Design modifications can be made a maximum of five times for the portrait and model, respectively.
	Scanning for 3D scene modeling	Scans large outdoor and indoor scenes for modeling.
	3D model polygon reduction	Reduces the number of polygons of a scene model.
Model rendering	3D scene model rendering	Renders a 3D scene model.
	3D virtual human model rendering	Renders a 3D virtual human model.
	Virtual simulation model rendering	Renders a virtual simulation model.
2D virtual human livestreaming	Virtual avatar video production (basic edition)	Based on customers' requirements, composite the existing 2D virtual avatar images and voices with the text or voice content provided by customers into a video. The video background can be a default image, or a customized image, video, or PPT. The virtual avatar actions in the video are generated based on the virtual avatar image and are not customized. The video does not include animation effects. Supports third-party preset timbres in languages other than English and Chinese. Maximum resolution: 1080p @ 30 FPS Customers are charged by video duration. The minimum billable duration is 5 minutes. If the duration is less than 5 minutes, billing is based on a 5-minute video. The value of the fee is rounded up. An output video can be modified for up to 3 times. Remote delivery is adopted.
	Virtual avatar video production (advanced edition)	Based on customers' requirements, composite the existing 2D virtual avatar images and voices with the text or voice content provided by customers into a video. The video background can be a default image, or a customized image, video, or PPT. The virtual avatar actions in the video are generated based on the virtual avatar image, including action orchestration. The video does not include animation effects. Supports third-party preset timbres in languages other than English and Chinese. Maximum resolution: 4K @ 30 FPS Customers are charged by video duration. The minimum billable duration is 5 minutes. If the duration is less than 5 minutes, billing is based on a 5-minute video. The value of the fee is rounded up. An output video can be modified for up to 3 times. Remote delivery is adopted.
	Virtual avatar image modeling (basic edition)	Collects video data (within 5 minutes) for a specified real-life image (any race) as required and reproduces the image (full body), including hairstyles, images, clothing, facial expressions, and actions, to generate a 1080p image model. The image model is used in scenarios such as video production. The image model moves naturally with good lip sync. This service adopts one-off charge.
	Virtual avatar image modeling (advanced edition)	Collects video data (within 10 minutes) for a specified real-life image (any race) as required and reproduces the image (full body), including hairstyles, images, clothing, facial expressions, and actions, to generate a 4K image model. Action orchestration is supported. The image model is used in scenarios such as video production. The image model moves naturally with good lip sync. This service adopts one-off charge.
	Virtual avatar voice modeling (Chinese/English)	Based on customer requirements and specified collection requirements, collect audio data (100 sentences) of a human voice (including Chinese and English) and reproduce the voice to create the voice of virtual avatars. The mean opinion score (MOS) is greater than 4. It can be used in scenarios such as video production. This service is charged by the number of voice models.
	Virtual avatar voice modeling (Non-Chinese/English)	Based on customer requirements and specified collection requirements, collect audio data (100 sentences) of a human voice (including French, Portuguese, Japanese, Korean, Spanish, Malay, Thai, Finnish, Turkish, and Arabic) and reproduce the voice to create the voice of virtual avatars. The mean opinion score (MOS) is greater than 4.2. It can be used in scenarios such as video production. This service is charged by the number of voice models.

Prerequisites

The customer shall submit a service request at least 10 working days in advance so Huawei can evaluate customer requirements and assign experts in virtual human models.
The customer and Huawei Cloud must reach an agreement on the requirements and objectives of the service and sign a contract about them.

Service Scope

Services Covered
Modeling of virtual humans, accessories, and virtual scenes, and applications such as virtual human videos, posters, and livestreaming. Videos and livestreaming support English, Mandarin, French, Portuguese, Japanese, Korean, Spanish, Malay, Thai, Finnish, Turkish, and Arabic.
Services Not Covered
Huawei engineers will only deliver the services specified in this statement of work (SOW). If the customer demands the delivery of services not specified in this SOW, additional fees may be charged. The services that are not covered include but are not limited to the following:
1. Traveling to other cities for work. Additional expenses for travel, accommodation, and communications will be billed.
2. Working beyond normal working days and working hours as required by the customer (Note: Fees include overtime pay and subsidies.)
3. Purchasing products from other companies or individuals.

Service Process

Service request submission
The customer submits a service request.
Communication and confirmation on requirements
- Huawei organizes experts to learn about and evaluate the customer's requirements.
- Huawei confirms the service scope and delivery objectives with the customer.
Service order confirmation
- Huawei proposes a solution based on the customer's requirements.
- Huawei confirms the proposal with the customer and then proposes an official service delivery plan.
Consulting & implementation
- The project is officially initiated.
- Huawei prepares project acceptance materials and deliverables.
Service acceptance
- The customer signs the Cloud Video Delivery Service Acceptance Report or clicks the acceptance link on the official website.
- The project has been accepted.

Service Deliverables

Category	Sub-Category	Service	Deliverable
Digital content delivery & implementation service	Solution	3D realistic virtual human	3D realistic virtual human model
		3D beauty virtual human	3D beauty virtual human model
		3D animated virtual human	3D animated virtual human model
		3D stylized virtual human	3D stylized virtual human
		CG-only poster production	CG-only poster
		Virtual-real poster production	Virtual-real poster
		CG-only video production	CG-only video
		Virtual-real video production	Virtual-real video
		CG-only virtual livestreaming	CG-only virtual livestreaming
		Virtual-real livestreaming	Virtual-real livestreaming
		CG-only video for news broadcast	CG-only video for news broadcast
		Virtual-real video for news broadcast	Virtual-real video for news broadcast
		CG-only video for 3D object display	CG-only video for 3D object display
		Virtual-real video for 3D object display	Virtual-real video for 3D object display
		Sara 3D virtual human video production	Sara 3D virtual human video production
		Yunxiao 3D virtual human video production	Yunxiao 3D virtual human video production
		3D stylized virtual human video production	3D stylized virtual human video production
		Virtual human video broadcast (basic)	Virtual human video broadcast (basic)
		Virtual human video broadcast (advanced)	Virtual human video broadcast (advanced)
	Integration implementation	3D space integration implementation	Technical support
	3D materials	3D realistic hairstyle modeling	3D realistic hairstyle model
		3D beauty hairstyle modeling	3D beauty hairstyle model
		3D animated hairstyle modeling	3D animated hairstyle model
		3D realistic clothing modeling	3D realistic clothing model
		3D beauty clothing modeling	3D beauty clothing model
		3D animated clothing modeling	3D animated clothing model
		3D realistic accessory modeling	3D realistic accessory model
		3D beauty accessory modeling	3D beauty accessory model
		3D animated accessory modeling	3D animated accessory model
		3D scene modeling	3D scene model
		Dynamic special effects for 3D virtual humans	Dynamic special effects for 3D virtual humans
		Animated actions for 3D virtual humans	Animated actions for 3D virtual humans
		3D object modeling	3D object modeling
		Scanning for 3D scene modeling	Scanning for 3D scene modeling
		3D model polygon reduction	3D model polygon reduction
	Model rendering	3D scene model rendering	3D scene model rendering
		3D virtual human model rendering	3D virtual human model rendering
		Virtual simulation model rendering	Virtual simulation model rendering
	2D virtual human livestreaming	Virtual avatar video production (basic edition)	1080p virtual avatar video
		Virtual avatar video production (advanced edition)	4K virtual avatar video
		Virtual avatar image modeling (basic edition)	Virtual avatar model
		Virtual avatar image modeling (advanced edition)	Virtual avatar model (advanced edition)
		Virtual avatar voice modeling (Chinese/English)	Virtual avatar voice model (Chinese/English)
		Virtual avatar voice modeling (Non-Chinese/English)	Virtual avatar voice model (non-Chinese/English)

When planning the project, Huawei must negotiate with the customer on the final deliverables, which should be project-specific.

Responsibilities

Shared responsibilities
- Negotiating and confirming the virtual human service scope and objectives
- Signing a contract
Customer responsibilities
- Providing accurate information about requirements and scenarios
- Registering with Huawei Cloud and topping up the account (with at least CNY5,000) for purchasing cloud products and resources used for virtual human models
- Reviewing and confirming the guidance plan and deliverables provided by Huawei
Huawei responsibilities
- Before modeling: submitting a development plan and a BOQ to the customer for confirmation
- During modeling: developing a model for the specified customer as scheduled, and confirming the modeling results phase by phase
- After modeling: formulating service items based on the selected model and providing the deliverable list
Responsibility matrix

The following responsibility matrix is for reference only and can be modified based on project requirements.

R: Responsible; S: Support

Delivery Item	Phase	Activity	Huawei	Customer
Pre-sales support	Commercial involvement	Business negotiation	R	S
	Pre-sales survey	Business issue identification	S	R
	Pre-sales survey	Business understanding	S	R
Model definition	Application scenarios and characteristics of the virtual human model	Business understanding and model building	S	R
Modeling	Modeling based on information provided by the customer	Modeling	R	S
Acceptance	Model delivery	Deployment summary and documentation handover	R	S
Acceptance	Confirmation	Deployment acceptance by the customer and handover	S	R

Acceptance Criteria

The deliverables of each sub-service must be submitted in compliance with the following criteria. If the customer accepts the deliverables, they sign or seal the deliverables or click the acceptance link on the official website.

Parent topic: Cloud Video Delivery Service

Previous topic: Cloud Video Delivery Service

Next topic: FAQs