You know, it occurs to me that I think "H" is a slang term for a certain illegal drug with the same first initial. Somewhere, I'm sure that'll probably get picked up on by Twitter, if it hasn't already.
OCR says:
「CeVIO AI IA TALK 」「CeVIO AI ONE TALK」のご購入を検討中の皆様へ
この度は、CeVIO Al 開発の遅延により、未だサンプルボイスをお聴かせすることかが出来ていない状況です。
ご予約、又は購入をご検討頂いている皆様には心よりお詫び申し上げます。
必ず発売日までにはサンプルボイスをお聴かせ出来るよう鋭意制作中でございますので
もうしばらくお待ち頂けますと幸いです。
尚、これに伴いまして、当初 3 月 12 日まで予定しておりました
予約割引・予約特典につきましては 3/31 午前 10 時までに延長させて頂きます。
発売は予定どおり 3 月 12 日に開始いたしますので、楽しみにお待ち頂いている皆様には
今しばらくお待ちいただけますようお願い申し上げます。
They appear to be apologising for a lack of demo so close to the pre-order period, and explaining that production has been delayed. Pre-order benefits have been extended.
「CeVIO AI IA TALK 」「CeVIO AI ONE TALK」のご購入を検討中の皆様へ
この度は、CeVIO Al 開発の遅延により、未だサンプルボイスをお聴かせすることかが出来ていない状況です。
ご予約、又は購入をご検討頂いている皆様には心よりお詫び申し上げます。
必ず発売日までにはサンプルボイスをお聴かせ出来るよう鋭意制作中でございますので
もうしばらくお待ち頂けますと幸いです。
尚、これに伴いまして、当初 3 月 12 日まで予定しておりました
予約割引・予約特典につきましては 3/31 午前 10 時までに延長させて頂きます。
発売は予定どおり 3 月 12 日に開始いたしますので、楽しみにお待ち頂いている皆様には
今しばらくお待ちいただけますようお願い申し上げます。
They appear to be apologising for a lack of demo so close to the pre-order period, and explaining that production has been delayed. Pre-order benefits have been extended.
Doesn't it seem strange that both IA, ONE, and KAFU are having mysterious delays? It's especially strange since IA and ONE already have Talk voices with older CeVIO.
They talked about production of CeVIO AI being delayed, it could just mean their specific product, but maybe the engine itself is undergoing some changes? Yukari came out wonky, so perhaps they're ironing out some bugs.
They explicitely talked about improving CeVio's security in the message on their site
So I think that's definitely causing at least part of the delays, probably along with engine changes/upgrades
So I think that's definitely causing at least part of the delays, probably along with engine changes/upgrades
Then again I'm not sure if it's such a big thing at all.
The release date is still the same, so it could also be that they were planning their development with some "buffer time", and have to use it up for making the VB and making a demo because something took longer than expected
It doesn't necessarily need to be more than that, I'd be far more concerned if they had postponed the release. To me, it sounds a bit like "things didn't go perfectly but are still mostly within margins"
The release date is still the same, so it could also be that they were planning their development with some "buffer time", and have to use it up for making the VB and making a demo because something took longer than expected
It doesn't necessarily need to be more than that, I'd be far more concerned if they had postponed the release. To me, it sounds a bit like "things didn't go perfectly but are still mostly within margins"
Still releasing on the 12th but the benefits and discounts from pre-ordering are gonna keep going until the end of the month. Guessing they'll start using them for ARIA STATION or whatever after that... I guess they've been focusing on making the Song voices since they have 3 of them to do at once.
Regarding Kiritan she seems to have far less glitches than Yukari, basically everyone I've talked to who has her says she works amazingly well. Yukari meanwhile is just a whole bundle of issues in one.
Regarding Kiritan she seems to have far less glitches than Yukari, basically everyone I've talked to who has her says she works amazingly well. Yukari meanwhile is just a whole bundle of issues in one.
A sample of IA's AI talk voice has been released:
She sounds like quite a cutsey voice as usual, it would have been nice to have a talking voice that's faithful to IA's original, more ethereal singing voice. She always sounds a lot younger to me, when compared to Lia's vocals.
She sounds like quite a cutsey voice as usual, it would have been nice to have a talking voice that's faithful to IA's original, more ethereal singing voice. She always sounds a lot younger to me, when compared to Lia's vocals.
It sounds better than the OG Talk bank, but it's still too cutesy and not like IA's singing bank at all, so I still don't like it overall. 
Personally, I feel like after all these years, IA is not supposed to sound exactly like Lia. I think she's more like a voice acted, younger character.
I think that the expressiveness has improved noticeably with her new talk banks, but she still sounds "like IA".
Also, I see a lot of disappointed English-speaking fans that wanted her to talk just like Lia, but in all seriousness, only a small fraction of Western synth fans can actually use talk banks or enjoy content made with them due to the language barrier. So I think Westerners will quickly lose interest in this product, to be blunt.
I think that the expressiveness has improved noticeably with her new talk banks, but she still sounds "like IA".
Also, I see a lot of disappointed English-speaking fans that wanted her to talk just like Lia, but in all seriousness, only a small fraction of Western synth fans can actually use talk banks or enjoy content made with them due to the language barrier. So I think Westerners will quickly lose interest in this product, to be blunt.
Talkloids are actually pretty popular in the English speaking fandom, there's multiple uploaded a day with Japanese-only banks being surprisingly some of the most used. I think so long as there are fans interested seeing the characters interact, they'll be excitement for speech-synths. I've even seen cover artists complaining that speech skits get more views than their musical works.
(I am not sure how to properly word this, so my reply will be awkward.)
In my mind, there are two types of Talkloid fandoms.
For English speakers: They make skits by using singing synths (ex: VOCALOIDs) and have Japanese-only voice banks speak in Engrish. They do not seem to touch actual TTS voice banks (ex: VOICEROID) due to the language barrier (interface only in Japanese) and it being less easy to edit phonemes into Engrish compared to in a singing synth where you can chop notes to be very tiny. (In TTS, you can basically only work with katakana). When an English-only speaking person buys a character who ONLY is a VOICEROID, they usually want to make them sing with Kotonosync and don't use it for anything else.
For Japanese speakers: There are users who make skits using singing synths, but a huge amount of them make videos using actual TTS. Videos are usually much longer than Western fans' videos (ex: Let's Plays, tutorials, cooking videos, multi-video skits). The Western fandom barely interacts with this content due to lack of subtitles, and barely knows it exists or how popular TTS characters are, let alone being able to remember the character names.
So I feel like if a person who ONLY speaks English is going to make a talkloid video with IA, they would use her singing software (ex: VOCALOID). A Japanese user would use her TTS. I would be very surprised if someone who can't speak Japanese bothered buying an expensive TTS IA voicebank for a few very short joke talkloid videos due to being unable to fully utilize the bank. If someone who can't speak Japanese buys her singing voice, they can also use it for covers, which is more bang for their buck because you can cover songs without being able to speak the language. In addition, CeVIO is not compatible with Kotonosync, so there is less appeal of jumping through hoops to make TTS IA sing because she already HAS singing banks.
(This is just my humble opinion.)
There's definitely a difference in how talkloids are made between the international and Japanese fandoms. I hardly ever see people using Japanese talk banks for them simply because it's a pain to get them to express things in English properly.
Regarding IA Talk having a cute voice; Talk works with the Alpha parameter too if I remember correctly, so you don't need to worry about her always sounding cute, plus it'll probably respond better than before thanks to AI technology. IA was never meant to be a 1:1 replica of Lia's voice anyway, just based on her singing. If she was meant to have been a direct recreation then she would have been recorded where Lia's natural registers lie on VOCALOID. I think her Song banks will be fine regardless since IA English C wasn't nearly as cute sounding as IA Talk was on CS, but we'll have to wait for demos still. Hope that we can hear ONE soon!
Regarding IA Talk having a cute voice; Talk works with the Alpha parameter too if I remember correctly, so you don't need to worry about her always sounding cute, plus it'll probably respond better than before thanks to AI technology. IA was never meant to be a 1:1 replica of Lia's voice anyway, just based on her singing. If she was meant to have been a direct recreation then she would have been recorded where Lia's natural registers lie on VOCALOID. I think her Song banks will be fine regardless since IA English C wasn't nearly as cute sounding as IA Talk was on CS, but we'll have to wait for demos still. Hope that we can hear ONE soon!
I also think her overall quality has improved so much. The choppiness of the non-Ai version is gone, and she has become a lot more expressive with a lot more variation between emotions.
I think that‘s far more important than whether she sounds „too cute“ on the default settings or not. You can change them and turn her both into a cute and a mature vocal, depending on what you like. There‘s audiences for both, and I think 1st Place hit a good balance between them
I think that‘s far more important than whether she sounds „too cute“ on the default settings or not. You can change them and turn her both into a cute and a mature vocal, depending on what you like. There‘s audiences for both, and I think 1st Place hit a good balance between them