SynthV Wiki
SynthV Wiki
Advertisement
Xingchen
SynthV Studio
VOCALOID4
|
Originals
|
Gallery

This article is about the Synthesizer V Studio software known as a voice database, dubbed Infinity. If you are looking for the Synthesizer V character then click here.

History[]

Xingchen was originally conceptualized as "Quadimensionko", a mascot for the Quadimension doujinshi album series and group.[1] She was later renamed to "Xingchen" upon announcement of her VOCALOID4 voicebank production, which was developed by Shanghai HENIAN Information Technology Co. Ltd. in collaboration with Beijing Photek S&T Development Co., Ltd. (then known as "Quadimension"). Throughout her development and time as a VOCALOID, Xingchen received a complete redesign and a character biography. She was released for VOCALOID4 on April 13, 2016 and was later noted to be a part of MEDIUM⁵.

2019[]

On March 28, 2019, Ddickky posted on Weibo that two Xingchen voice databases were being planned for Synthesizer V. These voice databases were labeled as "Xingchen/ZERO English" and "Xingchen/ZERO Chinese", as the character would be determined based on the differences between the new voice databases and Xingchen's original VOCALOID voicebank. If the differences were too great, the voice databases would be known as ZERO rather than Xingchen.[2] However, due to contractual issues with Chalili, Xingchen's voice provider, Xingchen could not be produced for another voice synthesizer engine at the time, greatly delaying Beijing Photek's Synthesizer V plans for her. Instead, another character was created with a similar but villainous image to Xingchen, which led to the release of Xingchen Minus, who was instead voiced by Sawako Sui Hua Zi.

2021[]

On June 19, 2021, Ddickky noted that he would like to create a "Xingchen Infinity" voice database that would be made for Synthesizer V AI, but could not due to well-known reasons. He was looking for solutions to the issue.[3] On August 13, Ddickky posted the VOCALOID contract for Xingchen to bilibili. At the time, Beijing Photek was discussing the matters with Shanghai HENIAN, working towards having the contract revoked.[4] This voice database was announced to be in production on August 25, with three conceptual designs revealed. The company announced they were looking for feedback and design proposals from fans.[5]

On November 27, Ddickky announced that the recording process for Xingchen AI would probably be completed at the end of the month or by early December. Afterward, things would speed up as Dreamtonics was noted to be very efficient. If nobody broke down the process, then she could be put out for sale before the Chinese winter holidays.[6]

On December 6, Quadimension's official bilibili account posted information about their upcoming music course plans that begin on January 18, 2022. Included in the voice database user benefits pack, Xingchen Infinity was listed. No other information was provided at the time.[7]

2022[]

On January 1, 2022, it was announced that Xingchen Infinity's recording period was nearing completion. Quadimension hoped to showcase her as soon as possible.[8] On January 25, Ddickky intended to have two demos released by that night:[9] a Mandarin Chinese demonstration (later published as a self-cover of "Zero-Sum") and a Japanese Cross-lingual Singing Synthesis demonstration (later published as a cover of "Zankyousanka").[10] Ddickky commented that improvements for Infinity's beta were plotted within the past few days. New training materials were submitted that morning, with more songs in the higher vocal ranges to perform better in power, as well as more Japanese and English songs. Pre-orders were planned to start the next day, along with the release of an English demo. Once the training was finished, the product could be released.[11]

On January 26, Ddickky released a statement regarding the production of Xingchen Infinity and the difference in sound compared to her VOCALOID4 version. He noted that the recording method was different and had its own requirements. For the sake of sound quality, it was required to record with a more solid or straightforward voice. Next, the recording material was put into the engine and came back out as a library. Because of the changes created by the engine's processing, this would result in its texture: the interventions and distortions produced by the engine when putting the library together from the samples.[12]

Later in the day, Xingchen Infinity was made available on Quadimension's Taobao shop for pre-ordering. Her normal price was confirmed to be ¥812, but during the pre-order period, users could get her with a discount of ¥100 off, pricing her at ¥712. However, for users who already had her VOCALOID4 voicebank, they could receive an additional ¥100 off, meaning they could purchase her for a total of ¥612. A release date was not yet available at the time. It was also revealed that her final illustration was by ATDan and used the MEDIUM⁵ 2050 design that was introduced on August 12, 2020 for Xingchen's birthday and was designed by xiao Ye You Xi.[13] On January 29, Quadimension announced that the product should be finalized to release in weeks.[14] On February 10, Ddickky posted that Xingchen Infinity needed further improvements and would be delayed because of it. This meant that the pre-order period would be extended.[15]

On February 18, it was found that some producers began to receive her voice database and activation codes early. Ddickky later followed up to note that he would not be sharing any uploads from other producers as this would create confusion among fans and would be troubling to explain to customers why they hadn't received their copy yet. He confirmed that her intended release was February 20.[16] According to her product page, the pre-order period was to end on February 21.[17]

At some point, Xingchen Infinity's Taobao link also allowed for a starter pack option with the Synthesizer V Studio Pro editor.[18]

Updates[]

MEDIUM5 Stardust AI v102 - March 25, 2022[19] 
A free update for MEDIUM5 Stardust (also known as Xingchen Infinity) was released following the Synthesizer V Studio 1.6.0 update. This enables her to use the new Vocal Mode options. Eight variations were included: "Airy", "Bright", "Cool", "Dark", "Emotional", "Power", "Solid", and "Sweet".
MEDIUM5 Stardust AI v106b1 - November 10, 2022[20] 
A free update for MEDIUM5 Stardust (Xingchen Infinity) was released following the Synthesizer V Studio 1.8.0b1 update. This enables her to use the new Diffusion Probabilistic Models (DPM) feature.
MEDIUM5 Stardust AI v106 - November 24, 2022[21][22] 
Final version update compatible for Synthesizer V Studio 1.8.0 which enables MEDIUM5 Stardust (Xingchen Infinity) to use the new Diffusion Probabilistic Models (DPM) feature if the beta version was not installed previously.
MEDIUM5 Stardust AI v107b1 - June 21, 2023[23][24] 
A free update for MEDIUM5 Stardust (Xingchen Infinity) was released following the Synthesizer V Studio 1.9.0 update. This implemented the new rap feature available for the English and Mandarin languages. Cantonese Chinese was added to the Cross-lingual Singing Synthesis feature as well as general improvements to the overall function, resulting in more accurate pronunciation that sounds closer to a native speaker when singing foreign words.
MEDIUM5 Stardust AI v107 - July 6, 2023[25] 
Final version update compatible for Synthesizer V Studio 1.9.0 which enables MEDIUM5 Stardust (Xingchen Infinity) to use the rap feature (English and Mandarin Chinese only) and allows for Cantonese Chinese Cross-lingual Singing Synthesis if the beta versions were not installed previously.
MEDIUM5 Stardust AI v108b1 - August 2, 2023[26][27][28] 
A free update for MEDIUM5 Stardust (Xingchen Infinity) was released following the Synthesizer V Studio 1.10.0b1 update. This enhanced pitch generation by utilizing Reinforcement Learning from Human Feedback (RLHF), translating to fewer off-key errors, making better use of vibrato, and creating more expressive transitions between notes. This also allowed for the use of the new "Enhancement" slider in the AI Retakes panel, which controlled the level of changes introduced by RLHF. For Synthesizer V Studio Pro users, feedback submission was enabled, allowing users to consensually share their preferred takes with only Dreamtonics for research and development purposes, where models trained with this feedback would be deployed in future voice database updates. Users may also opt out of this feature if undesired.
MEDIUM5 Stardust AI v108b2 - September 28, 2023[29][30] 
A free update for MEDIUM5 Stardust (Xingchen Infinity) was released following the Synthesizer V Studio 1.10.0b2 update, which was the second beta version following version 1.10.0b1 that introduced enhancements to voice databases and integrated the latest user feedback via Reinforcement Learning from Human Feedback (RLHF). Regarding the editor, numerous reported issues were addressed for a more fluid experience. A new generation of pitch models were trained using the extensive feedback from users over the past month, and the new models offered greater accuracy in voice timbre reproduction, as well as a reduction in mispronunciations. Dreamtonics also focused on issues from the previous beta release that made Vocal Modes seem less distinct on certain vocals, ensuring clearer differentiation in the 1.10.0b2 update. The latest update also introduced a new feature in the Note Properties Panel that allowed for even division of a note into multiple syllables, proving to be a useful tool for songs with faster tempos.
MEDIUM5 Stardust AI v108 - October 19, 2023[31] 
Final version update compatible for Synthesizer V Studio 1.10.0 which enables MEDIUM5 Stardust (Xingchen Infinity) to use Reinforcement Learning from Human Feedback (RLHF) as well as other enhancements (as listed in v108b1 and v108b2's notes) if the beta version was not installed previously.
MEDIUM5 Stardust AI v109 - October 26, 2023 
A free update for MEDIUM5 Stardust (Xingchen Infinity) was released that added support for the Japanese language to the rap feature.
MEDIUM5 Stardust AI v110b1 - December 18, 2023[32] 
A free update for MEDIUM5 Stardust (Xingchen Infinity) was released following the Synthesizer V Studio 1.11.0b2 update. This implemented support for the Spanish language in the Cross-lingual Singing Synthesis feature.
MEDIUM5 Stardust AI v110 - April 4, 2024[33] 
Final version update compatible for Synthesizer V Studio 1.11.0 which enables MEDIUM5 Stardust (Xingchen Infinity) to sing in Spanish through Cross-lingual Singing Synthesis if the beta version was not installed previously.

Voice Database Information[]

Demonstrations[]

Synthesizer V AI[]

Demonstrations

零和Zero-Sum (Líng Hé Zero-Sum) ft. Minus Favicon-bbbilibili
残響散歌 (Zankyousanka) ft. Cangqiong & Muxin - Japanese Cross-lingual Singing Synthesis Favicon-ytYouTube Favicon-bbbilibili
淋雨一直走 (Lín Yǔ Yīzhí Zǒu) Favicon-bbbilibili
淋雨一直走 (Lín Yǔ Yīzhí Zǒu) ft. Xingchen V4 - Short Comparison Favicon-bbbilibili Audio
7rings - English Cross-lingual Singing Synthesis Favicon-bbbilibili
年至愿今宵 (Nián Zhì Yuàn Jīnxiāo) ft. Cangqiong & Chiyu Favicon-bbbilibili
最终祈愿 (Zuìzhōng Qíyuàn) - Short Version; English Cross-lingual Singing Synthesis Favicon-bbbilibili
最终祈愿 (Zuìzhōng Qíyuàn) - Full Version; English Cross-lingual Singing Synthesis Favicon-bbbilibili
New Era of Virtual Singers Music Library[34]

涟漪 (Liányī) Favicon-bbbilibili
绮云泽 (Qǐ Yún Zé) ft. Haiyi Favicon-bbbilibili
人偶之梦 (Rén'ǒu zhī Mèng) Favicon-bbbilibili
D!s!llus!on Favicon-ytYouTube Favicon-bbbilibili
Virtual Singer Tuning Class Textbook[35]

知否知否 (Zhī Fǒu Zhī Fǒu) ft. Muxin Favicon-bbbilibili
aLIEz - Japanese Cross-lingual Singing Synthesis Favicon-bbbilibili
I Really Want to Stay at Your House - English Cross-lingual Singing Synthesis Favicon-bbbilibili
Kick Back - Japanese Cross-lingual Singing Synthesis Favicon-bbbilibili
零和Zero-Sum (Líng Hé Zero-Sum) ft. Yongye Minus Favicon-ytYouTube Favicon-bbbilibili

Voice Databases[]

Product Information
Versions
  Lite Version/FLT Available?: No  Starter Pack Available?: Yes  Web Version Available?: No

Package details as noted:
Xingchen Infinity is a feminine Mandarin Chinese AI voice database based on the singing style of Chalili. As expected of an AI voice database, this enables her to sing with a more human-like and realistic voice while sounding closer to her voice provider.

Vocal traits as noted:
  • Due to the difference in the recording method and engine processing, Xingchen Infinity results in a different sound and texture compared to her VOCALOID4 voicebank.
  • As per the Synthesizer V Studio 1.7.0 update, through High Dynamics Voice Models (HDVM), which later migrated to Diffusion Probabilistic Models (DPM) per the 1.8.0b1 update on November 10,[36] most AI voice databases may use AI Retakes to create multiple "takes" of a sung section with variations in pitch, dynamics, and timbre. This replicates the randomness of human vocalists, adds a more realistic feel, and allows users to find an ideal vocal without the need for "laborious tuning." Xingchen Infinity, however, was one of the AI voice databases lacking full HDVM support, and as such, the feature, could only be used for multiple pitch retakes.[37][38] As of the 1.8.0b1 update, Xingchen Infinity was made compatible with DPM via her v106b1 update.
    • This feature is only available in the Synthesizer V Studio Pro Editor.
  • According to the Dreamtonics page for Synthesizer V Studio, Xingchen Infinity was described to have a firm and confident voice.[39]

Phonetic notes as noted:
  • Phoneme format: X-SAMPA
  • Through Cross-lingual Singing Synthesis, Xingchen Infinity can sing in not only Mandarin Chinese but also in English and Japanese. Since the release of her v107b1 update on June 21, 2023, she can also sing in Cantonese Chinese with this feature; since her v110b1 update on December 18, she can also sing in Spanish with this feature. Cross-lingual Singing Synthesis is only available in the Synthesizer V Studio Pro Editor.
    • It should be noted that Xingchen Infinity's default language is set to Mandarin Chinese. She was confirmed to also be trained with English and Japanese songs in order to further optimize her Cross-lingual Singing Synthesis capabilities.
  • Ddickky confirmed that Xingchen Infinity was trained with songs in the higher vocal ranges to perform better in power.
  • According to Rui'an Ryan, one of Quadimension's associated tuners, to have Xingchen Infinity sound more like Xingchen V4, he suggested the following parameters:[40]
    • Gender -0.11
    • Enable aspiration output
    • On the unaspirated audio file, EQ it as depicted in this image
    • Combine it with the aspirated audio file.

Vocal Modes as noted (Synthesizer V Studio Pro Editor only):
  • As of March 25, 2022, Xingchen Infinity became compatible with the Vocal Mode function introduced in Synthesizer V Studio 1.6.0. Her v102 update included eight variations:
    • "Airy"
    • "Bright"
    • "Cool"
    • "Dark"
    • "Emotional"
    • "Power"
    • "Solid"
    • "Sweet"

Voice database sample

References[]

  1. Quadimensionko
  2. https://weibo.com/2182360953/Hn3f83q9L
  3. https://t.bilibili.com/537903039890499513
  4. https://t.bilibili.com/558425733660905885?tab=2
  5. https://www.weibo.com/6802075123/Kv4ram7mv
  6. https://t.bilibili.com/597601998680530256
  7. https://t.bilibili.com/601051909687946100
  8. https://t.bilibili.com/610673293554229556
  9. https://t.bilibili.com/619559602369571331
  10. https://t.bilibili.com/619632926054112178
  11. https://t.bilibili.com/619866988878952464
  12. https://www.bilibili.com/read/cv15002320
  13. https://forum.synthesizerv.com/t/topic/4913
  14. https://t.bilibili.com/621111769193096025
  15. https://t.bilibili.com/625549776072583574
  16. https://t.bilibili.com/628423959597161086
  17. https://cdn.discordapp.com/attachments/695380056943493280/944082015039418368/unknown.png - Screenshot as of February 17, 2022.
  18. https://cdn.discordapp.com/attachments/695381508139122748/1182708737488719902/image.png - Screenshot as of December 8, 2023
  19. https://dreamtonics.com/en/synthesizer-v-studio-1-6-0-update/
  20. https://dreamtonics.com/en/synthesizer-v-studio-1-8-0b1-update/
  21. https://dreamtonics.com/en/synthesizer-v-studio-1-8-0-final-update/
  22. https://twitter.com/ahsoft/status/1595674717200916481/
  23. https://dreamtonics.com/synthesizer-v-studio-1-9-0-final-update/
  24. https://twitter.com/dreamtonics_en/status/1671427830629154816
  25. https://www.bilibili.com/read/cv24826703/
  26. https://twitter.com/dreamtonics_en/status/1686648078571655168
  27. https://t.bilibili.com/825177687051993121
  28. https://dreamtonics.com/synthesizer-v-studio-1-10-0b1-update-enhancing-pitch-generation-with-user-feedback/
  29. https://twitter.com/dreamtonics_en/status/1707304150059639108
  30. https://dreamtonics.com/synthesizer-v-studio-1-10-0b2-update/
  31. https://dreamtonics.com/synthesizer-v-studio-1-10-0-final-update/
  32. https://dreamtonics.com/synthesizer-v-studio-1-11-0b2-update/
  33. https://dreamtonics.com/synthesizer-v-studio-1-11-0-update/
  34. https://space.bilibili.com/15817819/channel/collectiondetail?sid=304207 - A series uploaded by Quadimension that mainly demonstrates tuning. May be related to the tuning courses offered by the company. Many of these examples may be uploaded post-release of the voice database(s) and are thus separated from pre-release examples, which may also be part of the same series as is the case for Xingchen Infinity.
  35. https://space.bilibili.com/277698869/channel/collectiondetail?sid=124634&ctype=0 - A series uploaded by Quadimension that mainly demonstrates tuning. May be related to the tuning courses offered by the company. Many of these examples may be uploaded post-release of the voice database(s) and are thus separated from pre-release examples, which may also be part of the same series as is the case for Xingchen Infinity.
  36. https://dreamtonics.com/en/synthesizer-v-studio-1-8-0-final-update/
  37. latest?cb=20220722180019
  38. latest?cb=20220722175905
  39. https://dreamtonics.com/en/synthesizerv/
  40. https://t.bilibili.com/629226341093853800

Navigation[]

Advertisement