Search Results
5/16/2025, 8:36:43 PM
She sounds angry & sarcastic which is how I feel, but still unintended on my part.
https://pub-f3186dbecfd64ac085ddc742fc900f59.r2.dev/twilight_sparkle_neutral_1747418267794_variation0.wav
>>42195922
>Feel free to report any bugs or issues here, too
Yeah I see several bugs:
0. You're still not willing to jew out despite clearly needing the money and influence. Jew out or others will outjew you. Stop being a social recluse that's how all scientists die. Learn to sue everyone cause 11.AI clearly stole your technology you moron.
1. You're not open sourcing this to the community (which are of minimal help and lack money to pay for GPUs but they're willing to learn and are very loyal and creative despite me trashtalking them myself back in October)
2. I'm pretty sure ElevenLabs, Udio.AI, SUNO.Ai, etc. stole your technology and perfected it already since 90% of the singing & talking sounds like Tara Strong, Rebecca Shoichet & Ashleigh Ball. The AI can really sing too. To an audiophille it still sounds bad, but to a normie it sounds perfect. Get a fucking marketing team, both you and Tara Strong fucked each other up and should sue every single audio AI possible.
This is what Suno Ai can do right now with the paid model:
https://www.youtube.com/shorts/udOgG0M8pVI
3. Your options & UI is still limited. If I could search a reference line to use any emotion I want without typing in phonetics then that'd be useful for the average normie. You didn't understand what I just told you, did you? LET ME USE THE REFERENCE LINE TO QUICKLY & INSTINCTIVELY USE THE EMOTION I WANT. WE HAVE AN IMPECCABLE MEMORY OF THE SHOW'S DIALOGUE LINES.
Add a voice changer/voice to audio option. It would be so much more intuitive because the AI could hear what emotion I'm going for instantly.
Today's AI still lack a ton of UI options but are getting there at an insanely quick speed such as Suno's ability to grab an existing song and have either the same singer or a new singer sing the same notes with different lyrics.
Today's AI still sounds like an untrained voice actor slurring his lines on purpose and it still sucks compared to audiophille standards, but your current robot sounding AI is dreadful by normal standards. You still haven't learned how to remove the noise?
https://www.youtube.com/watch?v=qu5nnMOQ4VU&ab_channel=A
https://www.youtube.com/watch?v=I1Dy0Zfw6Qs&ab_channel=votums
3.5 You probably didn't notice cause you're not a voice director or you're autistic but ... S1 and S2-S9 's voice directing is completely different. 90% of the dialogue lines used in S2-S9 used only these emotions; depressed, angry, flirty, ANXIOUS, TIRED, reading-off-a-script-at-gunpoint. And that's the acting ... the voices?
In S2+ everyone sounds...
Twilight sounds much lighter in S2+
Applejack & Dash sound much deeper and not in a suave way.
Pinkie sounds way lighter & screechier.
Fluttershy always sounds anxious
Rarity & Spike kinda sound the same.
4. One more thing...
https://pub-f3186dbecfd64ac085ddc742fc900f59.r2.dev/twilight_sparkle_neutral_1747418267794_variation0.wav
>>42195922
>Feel free to report any bugs or issues here, too
Yeah I see several bugs:
0. You're still not willing to jew out despite clearly needing the money and influence. Jew out or others will outjew you. Stop being a social recluse that's how all scientists die. Learn to sue everyone cause 11.AI clearly stole your technology you moron.
1. You're not open sourcing this to the community (which are of minimal help and lack money to pay for GPUs but they're willing to learn and are very loyal and creative despite me trashtalking them myself back in October)
2. I'm pretty sure ElevenLabs, Udio.AI, SUNO.Ai, etc. stole your technology and perfected it already since 90% of the singing & talking sounds like Tara Strong, Rebecca Shoichet & Ashleigh Ball. The AI can really sing too. To an audiophille it still sounds bad, but to a normie it sounds perfect. Get a fucking marketing team, both you and Tara Strong fucked each other up and should sue every single audio AI possible.
This is what Suno Ai can do right now with the paid model:
https://www.youtube.com/shorts/udOgG0M8pVI
3. Your options & UI is still limited. If I could search a reference line to use any emotion I want without typing in phonetics then that'd be useful for the average normie. You didn't understand what I just told you, did you? LET ME USE THE REFERENCE LINE TO QUICKLY & INSTINCTIVELY USE THE EMOTION I WANT. WE HAVE AN IMPECCABLE MEMORY OF THE SHOW'S DIALOGUE LINES.
Add a voice changer/voice to audio option. It would be so much more intuitive because the AI could hear what emotion I'm going for instantly.
Today's AI still lack a ton of UI options but are getting there at an insanely quick speed such as Suno's ability to grab an existing song and have either the same singer or a new singer sing the same notes with different lyrics.
Today's AI still sounds like an untrained voice actor slurring his lines on purpose and it still sucks compared to audiophille standards, but your current robot sounding AI is dreadful by normal standards. You still haven't learned how to remove the noise?
https://www.youtube.com/watch?v=qu5nnMOQ4VU&ab_channel=A
https://www.youtube.com/watch?v=I1Dy0Zfw6Qs&ab_channel=votums
3.5 You probably didn't notice cause you're not a voice director or you're autistic but ... S1 and S2-S9 's voice directing is completely different. 90% of the dialogue lines used in S2-S9 used only these emotions; depressed, angry, flirty, ANXIOUS, TIRED, reading-off-a-script-at-gunpoint. And that's the acting ... the voices?
In S2+ everyone sounds...
Twilight sounds much lighter in S2+
Applejack & Dash sound much deeper and not in a suave way.
Pinkie sounds way lighter & screechier.
Fluttershy always sounds anxious
Rarity & Spike kinda sound the same.
4. One more thing...
Page 1