SynthV Synthesizer V Discussion Thread

[mary34]

Here is a full translation of ST Media's statement if anyone is curious:

ST Media's Statement (ENG by Infoholic+Bf)

Original copy by ST Media Translated by Infoholic (& boyfriend (hewwo :3 — gah this guy is rambly even in writing)) 안녕하세요. 에스티미디어 대표 윤민수입니다. Hello, this is ST Media's CEO, Yoon Min-soo. 먼저 가처분 신청과 관련하여 여러분께 심려를 끼쳐드린 점 진심으로 사과드립니다. 더불어 여러분 사이에 여러 의견들이 분분한 줄 압니다. 하지만 “판매금지가처분”이라는 것은 누구의 잘잘못을 가리는...

docs.google.com

the voice provider also made a statement on her youtube community page but i cannot find the link, only a screenshot with machine translation, and since i do not trust mtl as far as I can throw it I have choosen to not share that since a bad translation of a screenshot would just muddy the waters

edit: found the community post:

Post from [이정은]미루홍TV

안녕하세요 이정은입니다. 20년 남짓 보컬리스트로 활동하며 소송이라는 단어조차 생각해 본 적 없던 저에게, 이번 상황은 참으로 유감스러운 일이 아닐 수 없습니다. 그럼에도 불구하고 여기까지 올 수 있었던 것은, 절실하게 지켜온 제 삶 그 자체였던 목소리가 예상하지 못했던 방식으로 ...

www.youtube.com

 

[lIlI]

Heads up: There seem to be serious issues with tonal changes in Eclipsed Sounds voicebanks in the new update - it's getting frustrating how often SynthV updates feel like downgrades. It's a roulette for who gets improved and who gets nerfed

 

[mary34]

Heads up: There seem to be serious issues with tonal changes in Eclipsed Sounds voicebanks in the new update - it's getting frustrating how often SynthV updates feel like downgrades. It's a roulette for who gets improved and who gets nerfed

Tbh I think the issue here is there weren't any beta's for the ES voices (actually none of the third parties that updated had that)
given DT said 3rd parties get their own rounds of beta's first and many of the others haven't updated (None of my AHS or YnK SV2 voices have updates, only Mai) I do wonder if it might be some third parties just rushed out an update without consideration.

DT voices had extensive betas with many revisions so that is the only explanation I can think of. I don't wanna blame the 3rd parties too hard but there is nothing suggesting they were forced to release the updates so the way I see it it's more on them (as mean as that might sound).

 

[sunnyp4rk]

It's a niche instrument that no one really seems to see or be able to use as a direct replacement for voice actors or the specific VPs themselves. But on that note, I'm talking about known VPs, and no one even knew who Uni's VP was at this point, so it feels like it doubly wouldn't have replaced her?

If they had marketed UNI as her VP, and not as a character, then I'd completely agree that the voice replication COULD be seen as replacement...but, it wasn't. It was marketed as a character, one that already had an existing Vocaloid vb that people were a big fan of.

Most likely case is that people will buy any vsynth voicebank based on the character, rather than the VP. I doubt many people buy SV AI vbs with the intention of stealing the VP's job (not that you could, really).

We won't really know the VP's side of it aside from what's been revealed publicly, so I feel like speculation isn't good. But I do wonder about this point. Because like you said, other singer VP's have had SV vbs without any problems in their outside career (it probably helps that despite it being called "AI" you still have to USE the product lmao)

 

[Alphonse]

I just don't get the argument of being worried a voicebank will replace you. Did you think all those thousands of amateurs were gonna hire you? It's almost like a piracy argument: the people who buy a voicebank were not going to hire you to sing their song in the first place, and the people who were going to hire you are not going to settle for a voicebank.

 

[Leon]

Now that I'm home and at my PC, I have thrown the VP's statement from the Youtube post into DeepL, which isn't perfect but tends to be pretty good vs like Google Translate.

Hello, this is Lee Jung-eun.
Having worked as a vocalist for about 20 years, I had never even considered the word “lawsuit,” so this situation is truly unfortunate for me.
Nevertheless, I was able to come this far because I could not just stand by and watch my voice, which I had desperately protected as my very life, being used indiscriminately in an unexpected way.
The reason I have remained silent about my position until now is that I felt it was important to be cautious with my words until a decision was made,
and I believed that the most clear answer I could give would be shown through the outcome.

On July 21, a preliminary injunction was issued, granting my request and prohibiting the other party from selling the product. I would like to express my deep gratitude and respect to the court for issuing this decision.
I would also like to express my sincere gratitude to the many people who supported and encouraged me throughout this process, and I will do my best to repay their kindness.

My voice is a precious asset that I have carefully cultivated and nurtured over a long period of time.
I also believe that the development of AI technology is inevitable and should be supported, but it should not infringe on anyone's rights.
I sincerely hope that this decision will serve as a meaningful precedent for a society that protects and respects the rights of artists,
and that it will give courage to those in similar situations to continue fighting.

Thank you.

Translated with DeepL.com (free version)

And also it is correct, as in the translated document from Info that I referenced (and thanks for linking that here, Mary, I realized belatedly that it was linked in the Uni thread but not here), the screenshot of the court ruling that she has included shows point 1, where Uni is to be taken down from sale, but not points 2 or 3, which disregard her other requests.

The comments, at least according to machine translation, are not terribly supportive. A lot of them are pointing out other successful VPs who have publicly voiced banks and who are still doing well and getting work. Others are pointing out that she seems to misunderstand AI in this case, and that users still have to put the work in unlike generative AI that just does the thing when you prompt it to. I still don't know. I wish everyone well. Maybe she and the company can still come to an agreement, or maybe Uni is just taken down and that's that. Guess we'll see.

 

[lIlI]

In my view, whether or not vocal replication poses a financial risk to singers is moot; the voice is a part of the body, and if a voice provider realises it is being used in this way makes them uncomfortable, that should be respected without any need for justification. I think arguing whether or not her conjecture is correct might be missing the forest for the trees. After all, financial impact is often used in court to add support to a case that is primarily motivated by emotional impact.

That said, I don't think we should be too quick to assume that a realistic replica of someone's voice being available more cheaply than a session with them is an entirely unrealistic thing to be concerned about. We're on the very cusp of mainstream vocal synth adoption, and we haven't had time to see what its full ramifications will be for singers. I've been reflecting on this lately, because we've seen a few usages of SynthV from large corporations. These companies have the money to hire singers, and would have done so in the past. They chose to use SynthV over a real singer because it was cheaper. AI is having a provable effect on the jobs of workers in many areas, and it will change the landscape of the music industry in ways we cannot predict.

I don't believe AI vocal synths are inherently dangerous to artists; their creative benefits are why I'm a fan. However, I think it would be good for companies to be motivated to continue developing them as ethically as possible. If a singer loses a case where she objected to a product after its release, this gives companies reason to obscure the nature of their work when hiring talent, as they won't face consequences if the VP is misled. It opens the door to companies being able to pressure singers into giving up the rights to their voice for opportunities within the industry.

However, if a voice provider can revoke access to their voice at any time, this motivates companies to offer generous compensation and to only hire singers who give true consent with genuine enthusiasm. It would prevent companies from using the wording of old contracts to justify using the voice of someone uninformed who gave up their recordings before AI existed. As sad as it would be to lose Uni, I believe her voice provider winning this case would be healthier for the development of the industry overall, and for the protection of all ordinary people against technological exploitation.