Це відео не доступне.
Перепрошуємо.
Synthesizer V AI: Enhanced Pitch Generation with RLHF
Вставка
- Опубліковано 16 лип 2023
- Introducing Reinforcement Learning with Human Feedback (RLHF), a new technology we are adapting to our Synthesizer V AI pitch model.
Through the years, we have been constantly evolving our singing synth. Now that we are so close to replicating the nuanced, natural human voice, this is the next step.
While RLHF has previously been used in NLP and image generation, we at Dreamtonics are the first to apply it on audio.
RLHF-enhanced pitch models will be part of our next update. More info soon.
Synthesizer V Studio
dreamtonics.co...
-
Follow us:
Facebook - / dreamtonics
Instagram - / dreamtonicsaudio
Twitter (JP) - / dreamtonics_jp
Twitter (EN) - / dreamtonics_en
Official Website - dreamtonics.co...
Please make all voicebanks available for purchase on your store 🙏✨
There’s a huge improvement with the English and Japanese vocals! I’m excited to try this out
I have been producing music for tv for 35 years and in my opinion this is the greatest breakthrough in music production to date. Bravo Dreamtonics!👏
2:31 This part sounds beautiful, I love his lower notes! And Feng Yi's voice is precious. This technology seems really promising!
Can't wait until more languages are implemented , like Spanish! 😊
This is fantastic news and it's amazing that, even after already reaching a very high level of accomplishment, you continue to work on improvements to get even closer to the human voice in terms of artificial singing vocals.
kevin mentioned
Although HDVM had its faults, I was attracted to the idea of it, but not the implementation. However, I am delighted that they have added such an implementation back into the program. I'm looking forward to the new opportunities it will give us in the future. I was hoping for a new implementation for expressiveness of voices and better support for notebending. I think it's great that they are listening more and more to the community.
They never removed HDVM from the engine, it's been there since the update? This is a refinement to pitch generation.
My human feedback: This is amazing! I'm super excited for when this becomes available! 🔊✨
The new versions sound better, so sounds like a great update!
I think it's awesome to get feedback so comfortably. I think you can do it even more, with more features, learn how users actually use your plugin. Sonible does that as well, they collect data from their users on how the plugin is used etc. In the settings user s can opt out, but I think that's just an awesome way of improving the product.
Please dreamtonics release English opera vocal (like Classical Soprano and Baritone) and tagalog male and female vocal 😢🙏
I like this approach, a great way to build upon this would be to link this learning ability with the dictionaries so you could use it to build different personalities or essentially unique individual vocalists...
that's quite a bit of improvement actually! i have a few vocals where notes tend to auto process flat/sharp or nonsensical for the context a LOT and i'm excited to see the difference this makes.
Looks great! I've been wondering what influence the current AI boom would have on SynthV. Also, thanks for an English language video. Keep up the good work!
WOW !! Thanks for all theses updates ! The result in this video is impressive.
please release a japanese rock female vocal already 😭
Cong Zheng XLS?
@@dannyundos8927love her but her voice is honestly just not rock enough & her tone can be bland
spill
Teto?
@@TaelTheFennecher and stardust are the closest we have I fear
Can't wait for this feature in the future
this is amazing, hopefully you guys will release this soon!
This is amazing. Great update. I'm really looking forward to it
more rock voices please and please add a screamo feature. ❤
holy moly huge time saver
Another thing I'd like to see is a notes section for individual voices which could be used to keep notes on the vocal range of a vocalist, best keys, preferred genre, etc... Just to help keep track of who's best for what or whatever, just someplace that we can scribble in our own thoughts to be used when we're choosing a vocalist...
this is so good i can't wait!!!
I cant hear much on Kevin's side but Ryo has been greatly enhanced with the new model!
Sounds more consistent and the higher notes dont sound as strained.
They didn't do anything about the timbre, they changed the pitch generation.
Cool!
Amazing work, keep going :)
Please fix the export bug which has been reported over a year ago, the rendering starts from behind the vocals on several daws
When will Brazilian Portuguese voices be available?
Great stuff!
RLHF needs new voicebanks? So the question: has Solaria RLHF?
Awesome!
Will the new update be free for customers? If not, I’ll wait for it. The results seem impressive.
they usually always are, there hasnt been a paid one since synthesizer v editor - which was more like the program just leaving beta
wondering if Synthesizer V will ever have something like the vocalo changer or tiktok’s voice filter (but with singing)
omg ya quiero escuchar a teto y a stardust con esta nueva tecnología
🙏侃门🙏
Please make possible to play notes with a keyboard and have random words just to compose the melody
I'm not sure what you are asking for exactly, you can input your midi tracks directly into synthV
iirc they do have midi keyboard support and a recording feature for it
Synth V is a impressive tool but sadly Ai technology is about to kill it. Why spend a long time typing lyrics in a piano roll, when i can use an AI voice cloner of many famous singer over some vocals and make it sound fantastic in just a matter of minutes. Noone likes to type.