Thread inspired by this tweet:
AITalk:
The engine behind things like VOICEROID and GynoidTalk, it was announced in September that it is getting a deep-learning update, which will be available in April 2020. Its features include a more human/natural pronunciation and the ability to switch between emotions like joy, sadness, and anger.
This month, VOICEROID announced they are working on English VOICEROID. They revealed Tohoku Zunko is getting an English VOICEROID and there will be a male Japanese VOICEROID who has a soft and cool voice.
flower is getting a talk voice, which was demoed in October.
Verdict: This popular software will get more popular and even higher quality while expanding its reach to English-speaking fans.
Alter/Ego:
ALYS (a French vocal) has a counterpart named LEORA who has been in development since 2017. Their creators had a livestream that revealed LEORA's voicebank has a release date goal of before Christmas.
Verdict: Alter/Ego is deader than dead, but LEORA finally releasing will be a nice end note for the software which was previously riddled with controversy and give fans closure.
CeVIO:
Deep-learning update coming possibly next year (they had a working demo running on a laptop in October). Will for sure feature a more emotive Satou Sasara.
Verdict: Probably going from pretty dead/unpopular to catching a lot of attention.
DeepVocal:
Note: I am not an expert on what's happening with DeepVocal, so my summary is a bit vague.
This is a deep-learning-based successor to Sharpkey that was released this year. Many UTAUs were brought over to this new engine. Fans complain that it is confusing to get started due to lack of guides.
Verdict: It's free, it's new, it can only improve.
Emvoice:
Emvoice released Emvoice One and Lucy in November, but the release was met with criticism. Fans weren't pleased with the marketing, the lack of range, the trial being limited to 7 notes, lack of MIDI support, and Emvoice being cloud-based. It is unknown when the demands of fans will be met or how long it will take for another vocal to be developed.
Verdict: Met with cold reactions, the future does not look bright.
M. Morise's mystery synthesizer:
Masanori Morise (creator of STRAIGHT and WORLD (real-time vocal synthesizers, with WORLD powering UTAU and current CeVIO)) released Tohoku Kiritan's a cappella singing data this month for the general public to use to further vocal synthesis research and it's related to a new singing synthesis engine. Little else is known about the project.
Verdict: Benefitial to the community and something to keep an eye on.
Next Generation Piapro:
Prices of the Appends will be anounced this year and then sold some time during January ~ June (all Cryptonloids are recieving tweaked voices based on their V2/V3s). The new Piapro will feature plug-ins such as VocalDrive (for growls) and Cherry Pie (a vocoder so you can convert your voice into a Cryptonloid's). Not much has been shown in terms of demos (only 1 official demo from March)/screenshots, so it is very mysterious software. The relationship between Yamaha and Crypton is even more mysterious, with fans wondering if non-Crypton Vocaloids will work in the new Piapro.
Verdict: It's Crypton, so it's going to shake up the industry.
Synthesizer V:
Synth V has been working on a version 2 of the software, which includes a whisper ability and updated interface. They have been looking for Japanese and English-speaking business partners, but it seems they have found a "big" Japanese company to work with this month (we're not sure if it's an existing company in the vocal synth scene or an unrelated newcomer). Fans have been criticizing Synth V for its free vocals being stuck in development underworld (Genbu, who had his new voicebank demoed in April but still not released, and Yamine Renri, have been marked as demo versions since August 2018). It has also been difficult for non-Chinese fans to purchase voice banks from the website, so an improve store is in dire need.
The Qaudimension girls are almost complete, with either Stardust or ZERO English and a mysterious male vocal remaining.
Verdict: Slower/rockier start than some fans would like, but this young software (only a year old) seems to be on track to improving in its vocal selection and overseas availability.
UTAU:
Note: I am not an expert on what's happening with UTAU, so my summary is a bit vague.
Fans were shocked when UTAU creator, Ameya, appeared to be active again, stirring up hopes for a modernized UTAU (which is getting harder to run on modern computers). But it appears to be just a false alarm.
Verdict: No fancy modern UTAU update appears to be on the horizon.
VOCALOID:
VOCALOID5 came out in July, 2018 which included starter Japanese/English Vocaloids and Haruno Sora. In March 2019, Meika Hime and Mikoto were released and suprised fans. Vocaloids in development underworld: Zing, Yao Luniang, UNI Append/English, Cocorobo. The only Vocaloids that might actually be released: YANHE update. VOCALOID: AI also came out, but is probably not for normal people to use. People working for Yamaha have stated they downsized the Vocaloid department, further indicating the end of a reign.
Verdict: Pretty dead. Don't have much hope for new voicebanks.
Virtual Voice Incarnation (Also known as Tensei Koe Uranai):
Developed by GREE VR Studi Lab, this is a vocoder that uses M. Morise's WORLD. It currently offers 13 voices that are being developed by the community, users are asked to read dialogue to help teach the engine how to pronounce better. I'm not sure if it will end up being a software people can use for VTuber activities in the future, but their Twitter doesn't really have a lot of info.
Verdict: Competition in the vocoder world is a good thing! Hopefully it will continue to develop and bring us more news.
Voidol:
Voidol is a deep-learning-based vocoder that had been around for iPhone for a while, but this year a Mac version released in April. It suffered repeated delays for a Windows version, which finally released in August. The add-on voices weren't released until this month. Fan reaction was generally poor, with Voidol receiving 2/5 stars (almost 50% 1 stars) on Amazon Japan, due to the software not running at all or having poor, crackly quality worse than a normal voice changer (for changing pitch) and definitely useless for serious singing/narrating projects.
Verdict: A general disappointment, though novel.
Let me know if I missed anything and feel free to share your humble opinion on the current projects/controversies!
AITalk:
The engine behind things like VOICEROID and GynoidTalk, it was announced in September that it is getting a deep-learning update, which will be available in April 2020. Its features include a more human/natural pronunciation and the ability to switch between emotions like joy, sadness, and anger.
This month, VOICEROID announced they are working on English VOICEROID. They revealed Tohoku Zunko is getting an English VOICEROID and there will be a male Japanese VOICEROID who has a soft and cool voice.
flower is getting a talk voice, which was demoed in October.
Verdict: This popular software will get more popular and even higher quality while expanding its reach to English-speaking fans.Alter/Ego:
ALYS (a French vocal) has a counterpart named LEORA who has been in development since 2017. Their creators had a livestream that revealed LEORA's voicebank has a release date goal of before Christmas.
Verdict: Alter/Ego is deader than dead, but LEORA finally releasing will be a nice end note for the software which was previously riddled with controversy and give fans closure.CeVIO:
Deep-learning update coming possibly next year (they had a working demo running on a laptop in October). Will for sure feature a more emotive Satou Sasara.
Verdict: Probably going from pretty dead/unpopular to catching a lot of attention.DeepVocal:
Note: I am not an expert on what's happening with DeepVocal, so my summary is a bit vague.
This is a deep-learning-based successor to Sharpkey that was released this year. Many UTAUs were brought over to this new engine. Fans complain that it is confusing to get started due to lack of guides.
Verdict: It's free, it's new, it can only improve.Emvoice:
Emvoice released Emvoice One and Lucy in November, but the release was met with criticism. Fans weren't pleased with the marketing, the lack of range, the trial being limited to 7 notes, lack of MIDI support, and Emvoice being cloud-based. It is unknown when the demands of fans will be met or how long it will take for another vocal to be developed.
Verdict: Met with cold reactions, the future does not look bright.M. Morise's mystery synthesizer:
Masanori Morise (creator of STRAIGHT and WORLD (real-time vocal synthesizers, with WORLD powering UTAU and current CeVIO)) released Tohoku Kiritan's a cappella singing data this month for the general public to use to further vocal synthesis research and it's related to a new singing synthesis engine. Little else is known about the project.
Verdict: Benefitial to the community and something to keep an eye on.Next Generation Piapro:
Prices of the Appends will be anounced this year and then sold some time during January ~ June (all Cryptonloids are recieving tweaked voices based on their V2/V3s). The new Piapro will feature plug-ins such as VocalDrive (for growls) and Cherry Pie (a vocoder so you can convert your voice into a Cryptonloid's). Not much has been shown in terms of demos (only 1 official demo from March)/screenshots, so it is very mysterious software. The relationship between Yamaha and Crypton is even more mysterious, with fans wondering if non-Crypton Vocaloids will work in the new Piapro.
Verdict: It's Crypton, so it's going to shake up the industry.Synthesizer V:
Synth V has been working on a version 2 of the software, which includes a whisper ability and updated interface. They have been looking for Japanese and English-speaking business partners, but it seems they have found a "big" Japanese company to work with this month (we're not sure if it's an existing company in the vocal synth scene or an unrelated newcomer). Fans have been criticizing Synth V for its free vocals being stuck in development underworld (Genbu, who had his new voicebank demoed in April but still not released, and Yamine Renri, have been marked as demo versions since August 2018). It has also been difficult for non-Chinese fans to purchase voice banks from the website, so an improve store is in dire need.
The Qaudimension girls are almost complete, with either Stardust or ZERO English and a mysterious male vocal remaining.
Verdict: Slower/rockier start than some fans would like, but this young software (only a year old) seems to be on track to improving in its vocal selection and overseas availability.UTAU:
Note: I am not an expert on what's happening with UTAU, so my summary is a bit vague.
Fans were shocked when UTAU creator, Ameya, appeared to be active again, stirring up hopes for a modernized UTAU (which is getting harder to run on modern computers). But it appears to be just a false alarm.
Verdict: No fancy modern UTAU update appears to be on the horizon.VOCALOID:
VOCALOID5 came out in July, 2018 which included starter Japanese/English Vocaloids and Haruno Sora. In March 2019, Meika Hime and Mikoto were released and suprised fans. Vocaloids in development underworld: Zing, Yao Luniang, UNI Append/English, Cocorobo. The only Vocaloids that might actually be released: YANHE update. VOCALOID: AI also came out, but is probably not for normal people to use. People working for Yamaha have stated they downsized the Vocaloid department, further indicating the end of a reign.
Verdict: Pretty dead. Don't have much hope for new voicebanks.Virtual Voice Incarnation (Also known as Tensei Koe Uranai):
Developed by GREE VR Studi Lab, this is a vocoder that uses M. Morise's WORLD. It currently offers 13 voices that are being developed by the community, users are asked to read dialogue to help teach the engine how to pronounce better. I'm not sure if it will end up being a software people can use for VTuber activities in the future, but their Twitter doesn't really have a lot of info.
Verdict: Competition in the vocoder world is a good thing! Hopefully it will continue to develop and bring us more news.Voidol:
Voidol is a deep-learning-based vocoder that had been around for iPhone for a while, but this year a Mac version released in April. It suffered repeated delays for a Windows version, which finally released in August. The add-on voices weren't released until this month. Fan reaction was generally poor, with Voidol receiving 2/5 stars (almost 50% 1 stars) on Amazon Japan, due to the software not running at all or having poor, crackly quality worse than a normal voice changer (for changing pitch) and definitely useless for serious singing/narrating projects.
Verdict: A general disappointment, though novel.Let me know if I missed anything and feel free to share your humble opinion on the current projects/controversies!
Last edited:
