Digital Human Product Evaluation: D-ID/HeyGen

I recently played with digital human generation tools: D-ID, HeyGen and SadTalker (among which SadTalker is SD open source, so I won’t make a special introduction).

There are many tutorials about these digital human tools on the Internet. You can search and learn by yourself~ This article is mainly to chat about user product experience design, and it is also AGI Product Experience Design Miscellaneous Talk 001-Digital Human Generation.


D-ID

Access link: https://www.d-id.com/

2d40cfe411079549730a4445eeeeb998.png

(The picture material comes from D-ID official website)

One of the product functions of D-ID, chat.D-ID, face-to-face chat dialogue. For example, in sales, customer service, training and other scenarios, use digital human models to communicate. It also supports API interface and model integration.

47ac05dc72a033c1522624ca1929d00f.gif

The second product function of D-ID is to create a digital human. Currently there are 2 creation methods, one is prompt generation and the other is image generation. Voice supports GPT3 input, text input and audio input.

59470459ca327b880937529ae96cb205.gif

Tried both ways. The keywords have also been adjusted many times, and the numbers generated by the prompt are a bit eye-catching. If you don't want to use what comes with it, it is recommended to upload your own pictures.

Upload image generatione74df7376076a9d8667243fe9355dd29.png

Prompt generation

2a99f99f38483114c161b25577ba60fa.png

The interaction process is relatively simple. After registration, directly enter the main interface to generate a digital human. Mainly mouse click and input, no voice input.

Interface layout structure design, more engineering-oriented. The information modularization is relatively clear, but the human-computer interaction framework is relatively rigid. When the user's main operation is to generate digital portraits, the interface information is redundant, and the functions that do not need to be operated are fixed in the entire interface frame, resulting in limited effective use of space.

303d4cb99385dcb76bd82e4a9d13f79b.png

The voice can be selected, the language of the country and region (even Wu Nong soft language dialect), gender (including the sound quality of different age groups), tone (negative and positive emotions). The effect after the digital human is generated basically meets the needs of a simple digital human, but it may be the reason for the free version, the mouth shape matching and image quality are not very good (see the video at the beginning of this article for details).

8be783fd9c06fcd5d72b5494abcf6350.png

71b89cdb804c7d4a94551da63ab54e80.png

c04084320b5bcc202eaec06d524648b3.png

The product price positioning is as follows. The D-ID digital person is limited to 20 minutes of free experience.

8c0cb3318d750f12ef180e4c93cdce87.png

Hey Gen

Click on the invite link: https://app.heygen.com/guest/templates?cid=a01967c5

8ca407ef99917046c14c31694ade9574.png

(The picture material comes from HeyGen official website)

c27891e4f6b87bad2201e659dcc5b16e.png

Compared with D-ID, which introduces the main product functions, HeyGen's official website has a lot of marketing content. For example, the four reasons for choosing HeyGen, as well as the application cases in five targeted scenarios, fully allow people to understand the use of the product.

b4ecd13c2dabbc5bbdff28dbf64adba2.png

Upload image generation

a8605d3b43c576af4b33f66e37fd60cf.png

Comes with Prompt generation

9f54e489bc06ae2c99deec7cb73270d2.png

It is still recommended to upload suitable digital human photos and build your own digital human model.

In the use process, HeyGen disassembled the user steps. When the user selects a horizontal or vertical interface, enter the edit generated content, the interface is as follows. HeyGen has a wealth of video templates for selection, and you can also edit and generate PPT by yourself.

4ec4094d9486b074d4368d8946d8d6c9.png

Voice input, in addition to regular text and upload audio, HeyGen provides 5 minutes of voice input,

f84fd083e12ad376c490a89699a972f3.png

The sound option has a lot of content, HeyGen is a separate layer to choose from, and the filter item is set, which is relatively clear.

c72624f8c678c930593c37de7f65bba8.png

The product price positioning is as follows. HeyGen digital human 1 minute free experience every day.

df035a6815b922967ba3b1459dbe7e0a.png

Digital Human Generation Tool

In addition to D-ID, HeyGen... There are many AGI tools for digital humans, such as synthesia, pictory... etc. There are also video face-changing technologies such as fakeface, online chat version Her such as kupid, Creative digital humans like Chriper... The user experience of digital human generation tools and the experience of digital human in actual application scenarios are two separate but related topics. This time we only talked about the experience of the digital human generation tool, and you are welcome to leave a message and communicate more~

8d64ca60a284ff506b1015d58c4f6bb8.jpeg

community entrance

cd6e0e5b5e69d427bb09d450f5b28df6.jpeg

Guess you like

Origin blog.csdn.net/shadowcz007/article/details/131618243