Name: Eduard Khil (But only Trololo Sample) | AI RVC Model
Brand: Weights
Rating: 4.6 (21 reviews)

 Introducing the Eduard Khil (Trololo Sample) RVC V2 Model, an advanced AI voice model trained using Retrieval-Based Voice Conversion (RVC) technology. With over 1750 epochs of training, this powerful tool accurately mimics the unique vocal characteristics of renowned Soviet baritone singer Eduard Khil, famously known for his iconic “Trololo” performance.

Our cutting-edge RVC Model offers unparalleled precision when creating AI covers or text-to-speech applications, ensuring exceptional audio quality while preserving the distinct essence of Eduard Khil’s legendary vocals. Harness the capabilities of artificial intelligence and music to generate impressive renditions of classic hits or even transform your own texts into mesmerizing speech performances reminiscent of Eduard Khil.

The current results show promising potential; however, we at Weights acknowledge there is room for improvement, and our dedicated team is working tirelessly to refine and enhance the model’s accuracy for superior outputs. Don’t miss out – join us in exploring the exciting world of AI voices by trying our FREE AI tools for generating AI covers and text-to-speech today! Experience

Before Start - I find that for the moment the result is not yet perfect I'll try to improve it as soon as possible if I can optimize the dataset (even if I had already done a big improvement on Audition). - For the micro test my micro + voice is not perfect so there are some distortion of the result - Probably change the dataset to short them ? I think for this type of model (no really voice) 02 minutes is probably to long for the ia Last Update : 19/09/2023 - Model URL Version RVC V2.0 - Epochs - Steps 1.75k - 7.2k (for moment) - Dataset Time ~ 02:01 - Recommanded Usage All Usages - Search Feature Ratio 1

Eduard Khil (But only Trololo Sample)

Eduard Khil (But only Trololo Sample) - RVC V2 - 1750 Epochs

Rayzox57

box's junky

daunidebil

baldboicarti

Yeah I tried doing it with the part where Travis Scott was saying only fe!n but it didn’t really work out 😂

Now i'm working on one of the weirdest models here (I know there are a lot of models here) but I'm confident.  (it's litteraly just no sense)

i am making my models on local i have a cap on 1k

Last Update : <t:1707862174:R> Model URL : Version firev2: RVC V2.0 - Ov2 Super Pitch Extraction Algorithm : RMVPE Epochs - Steps : 750 - 3.750k ~ 00:00:33 [ S3RL - This Beat Has Got Me - Acapella ] Enhanced Audio Recommended Usage : Cover Search Feature Ratio : 0.75 Pitch : ( + 4 = Man ) - Depend Base Voice You can adjust if you find better version Previews : Preview_Cover.wav : Provide : Mustard - Ballin [ My DIY Acapella ] Contains External Effects : Yes Pitch : +4 Feature Ratio : 0.75

S3RL [ This Beat Has Got Me ]

Last Update : <t:1707655909:R> Model URL : Version firev2: RVC V2.0 - RIN E3 Pitch Extraction Algorithm : RMVPE Epochs - Steps : 650 - 29.90k ~ 00:10:01 [ PewDiePie's Tuber Simulator voices lines ] Enhanced Audio Recommended Usage : Speech Search Feature Ratio : 0.75 Pitch : ( + 0 = Man ) - Depend Base Voice You can adjust if you find better version Previews : Preview_TTS.wav : Provide : ElevenLabs Free Demo - Dave EN Contains External Effects : No Pitch : +0 Feature Ratio : 0.75 Preview_Cover_1.wav : Provide : Mustard - Ballin [ My DIY Acapella ] Contains External Effects : Yes Pitch : +0 Feature Ratio : 0.75 Preview_Cover_2.wav : Provide : Smash Mouth - All Star [ My DIY Acapella - Sample ] Contains External Effects : Yes Pitch : +0 Feature Ratio : 0.75

PewDiePie [ Felix Arvid Ulf Kjellberg ]

Last Update : <t:1707600965:R> - Model URL Version firev2: RVC V2.0 - Ov2 Super - Pitch Extraction Algorithm RMVPE - Epochs - Steps 1.50k - 3.10k - Dataset ~ 00:00:06 - Recommended Usage All - Search Feature Ratio 1.0 - Pitch ( +12 / + 6 = Man ) - Depend result Previews Preview_TTS.wav *Provide* : ElevenLabs Free Demo - Dave EN - *Contains External Effects* : No - *Pitch* : +6 - *Feature Ratio* : 1.0 - Preview_Cover.wav *Provide* : Mustard - Ballin [ My DIY Acapella ] - *Contains External Effects* : Yes - *Pitch* : +6 - *Feature Ratio* : 1.0

Parakeet

Parakeet [ Nintendo 3DS Camera - Ov2 Super ] - RVC V2 - 1500 Epochs

Before Start : I wanted to test the RIN E3 pre-training model again but this time with a very short but very good quality dataset (these are audios that were on an old external hard drive I can't find the source anymore but it's not from a game). This model is therefore more "experimental" I no longer have access to ElevenLabs "pro" so I had to refer to the "free" trial version which gives a more robotic TTS (less practical for testing) and don't count on my lousy English to test directly with my voice :kekw:. Last Update : <t:1707531572:R> Model URL : Version firev2: RVC V2.0 - RIN E3 Pitch Extraction Algorithm : RMVPE Epochs - Steps : 800 - 15.2k ~ 00:03:48 [ Duke Nukem Studio Voice - ??? ] Yes the dataset is very short Recommended Usage : Speech Search Feature Ratio : 0.75 Pitch : ( -12 10 = Man ) - Depend Base Voice Voice is Deep You can adjust if you find better version Previews : Preview_TTS.wav : Provide : ElevenLabs Free Demo - Joseph EN Contains External Effects : No Pitch 12 Feature Ratio : 0.75 Info : As mentioned above the tts is less expressive so the voice is just as expressive when applied with Applio. Preview_Cover.wav : Provide : Mustard - Ballin [ My DIY Acapella ] Contains External Effects : Yes Pitch 12 Feature Ratio : 0.75 Info : For Fun Important (Need Feedback) : I would need some feedback (not about the quality of the model itself) but about its fidelity because I know (having done some tests) that RIN E3 often provides a very good hearing quality but each time I have this strange feeling that the voice is different (it may be just an impression) but i need to be sure before continue to use this pretrain ! :nerddog:

Duke Nukem [ RIN E3 ]

Before Start : This is the French Version Last Update : <t:1707405159:R> Model URL : Version firev2: RVC V2.0 - Ov2 Super Pitch Extraction Algorithm : RMVPE Epochs - Steps : 300 - 52.8k ~ 00:59:01 [ Adibou 2 - 1996 - RAW Files ] Decompilation + Manual sorting Recommended Usage : Speech Search Feature Ratio : 0.75 Pitch : ( +10 / +8 = Woman ) - Depend Base Voice Voice is very difficult to adjust Work better on base Woman voice You can adjust if you find better version Previews : Preview_TTS_FR.wav : Provide : ElevenLabs V1 - Dorothy FR Contains External Effects : No Pitch : +10 Feature Ratio : 0.75 Preview_TTS_EN.wav : Provide : ElevenLabs V1 - Dorothy EN Contains External Effects : No Pitch : +10 Feature Ratio : 0.75 Preview_Cover.wav : Provide : Mustard - Ballin [ My DIY Acapella ] Contains External Effects : Yes Pitch : +10 (idk why work with this pitch) Feature Ratio : 0.75 Info : For Fun

Eduard Khil (But only Trololo Sample)

Description

Add a comment

Morshu (The Legend of Zelda)

Eduard Khil (1976 performace)

Saiba Momoi (Blue Archive)

Villager (Minecraft)

Sas (Counter-strike global offensive)

FIXED HATSUNE MIKU V3 - VOCALOID [SCROLL DOWN FOR LINK]

Fang [Brawl Stars]

Ariana Grande AI

JP Furina / Focalors CV: Inori Minase (Genshin Impact)

Mortis [Brawl stars]

Jungkook (BTS)

ENHYPEN Heeseung

Miss Circle (Fundamental Paper Education)

CaseOh

Takanashi Hoshino (from Blue Archive)

Surge (brawl Stars)

Saiba Momoi (Blue Archive) (VA: Tokui Sora)

SpongeBob SquarePants (Talking And Singing)

Melodie - (Brawl Stars) - (English & Korean Voices)

Sonic The Hedgehog [Roger Craig Smith]

Jay (ENHYPEN)

Satoru Gojo (JJK) [VA Yuichi Nakamura]