Author
|
Message
|
Kelleytoons
|
Kelleytoons
Posted 3 Years Ago
|
Group: Forum Members
Last Active: Last Year
Posts: 9.2K,
Visits: 22.1K
|
These voices all sound TERRIBLE - it just sounds like there are some filters being applied, which is nowhere near what an AI processed voice sounds like. I can't see any use at all for this - even $20 seems WAY too much money for this crap.
Alienware Aurora R16, Win 11, i9-149000KF, 3.20GHz CPU, 64GB RAM, RTX 4090 (24GB), Samsung 870 Pro 8TB, Gen3 MVNe M-2 SSD, 4TBx2, 39" Alienware Widescreen Monitor Mike "ex-genius" Kelley
|
|
|
Wilby
|
|
Group: Forum Members
Last Active: 2 Years Ago
Posts: 131,
Visits: 479
|
Kelleytoons (11/24/2022) These voices all sound TERRIBLE - it just sounds like there are some filters being applied, which is nowhere near what an AI processed voice sounds like. I can't see any use at all for this - even $20 seems WAY too much money for this crap.Yeah, I agree but to a certain extent based on my own basic hearing level. I have also checked two more other similar software and for me, they sounded worse and not worth sharing. Generated voices by MorphVox Pro 5 and MagicMic pales in comparison with those from Altered Studo and I guess it shows in their pricing differences. For indie devs and some hobbyists that have a small budget, maybe MorphoVox & MagicMic will do at the moment until there are better and more affordable alternatives. Available info online indicates these are all AI-powered. :D I hope NyVox will beat these software and can offer greater value. But time will tell, if they still have more of it... :P I'm sure there are more competitors coming and prices to be able to utilize these software should go down further in the future.
|
|
|
Kelleytoons
|
Kelleytoons
Posted 3 Years Ago
|
Group: Forum Members
Last Active: Last Year
Posts: 9.2K,
Visits: 22.1K
|
Those choices you mentioned don't, in any way, sound AI powered. They just sound filtered - MorphVox for sure is only a filter and my experience tells me that so is that Magic crap. AI powered is a completely different ballgame - and it requires processing power that a filter can't touch (which is why, as a general rule, it ain't real time - the exceptions use either a high end local GPU, or processing power in the cloud, which you end up paying $$$ for. I have no idea why NyVox is going to do in this regard but time will tell). Look - a filter is pretty easy to tell. AI voice sounds as if it's not your voice at all, as if there is another person there without any such crap as echo, or fuzz, or any distortion. If you hear even a tiny bit of distortion it ain't AI.
Alienware Aurora R16, Win 11, i9-149000KF, 3.20GHz CPU, 64GB RAM, RTX 4090 (24GB), Samsung 870 Pro 8TB, Gen3 MVNe M-2 SSD, 4TBx2, 39" Alienware Widescreen Monitor Mike "ex-genius" Kelley
|
|
|
sonic7
|
sonic7
Posted 3 Years Ago
|
Group: Forum Members
Last Active: 2 Years Ago
Posts: 1.7K,
Visits: 19.4K
|
I'd be happy with the "Altered Studio" quality (from the samples I've heard), - but then the price - ouch! It's a pitty they don't charge based on minutes of usage only - without the monthly clock ticking.
--------------------------------------------------------------------------------------------------------------------------------------------------------
Please be patient with me ..... I don't always 'get it' the first time 'round - not even the 2nd time! :( - yikes! ... ● MSI GT72VR Laptop, i7 7700HQ 4-Core 3.8 GHz 16GB RAM; Nvidia 1070, 8GB Vram ● iClone-7.93 ● 3DXChange Pipeline 7.81 ● CC-3 Pipeline 3.44 ● Live Face ● HeadShot ● Brekel Pro-Body ● Popcorn FX ● iRAY ● Kinect V2 ● DaVinci Resolve17 ● Mixcraft 8.1
|
|
|
Wilby
|
|
Group: Forum Members
Last Active: 2 Years Ago
Posts: 131,
Visits: 479
|
Kelleytoons (11/24/2022) Those choices you mentioned don't, in any way, sound AI powered. They just sound filtered - MorphVox for sure is only a filter and my experience tells me that so is that Magic crap.
AI powered is a completely different ballgame - and it requires processing power that a filter can't touch (which is why, as a general rule, it ain't real time - the exceptions use either a high end local GPU, or processing power in the cloud, which you end up paying $$$ for. I have no idea why NyVox is going to do in this regard but time will tell).
Look - a filter is pretty easy to tell. AI voice sounds as if it's not your voice at all, as if there is another person there without any such crap as echo, or fuzz, or any distortion. If you hear even a tiny bit of distortion it ain't AI.Since there was no major difference from the Stand-alone version in ScreamingBee.com (more expensive), I bought the Steam version with the Effects Rack plugin to test it further. Both just costed me about US$ 10. For that amount, it's worth the risk I think. To further clarify the technology used, I've reached out to the developer of MorphVox Pro 5. I asked whether the program was just using voice filtering and/or is actually AI-powered and if so, whether he utilized Machine Learning and/or Deep Learning... or something different. Will post an update here when and if I receive an answer. :D Furthermore, it looks like I was able to make the Effects Rack plugin work for me. Since the plugin was old, I tried using only VST files and not VST3 64-bit and then corrected the effects' directory path and assigned it to look one folder deeper. That made it recognize the external VSTs I downloaded which I got for free from Blue Cat Audio here. With support of more plugins, maybe it is possible to clean up some of the distortions and obvious filters applied to a voice. If possible, I just hope it wont take a lot of time to do. hehe :crazy:
|
|
|
AutoDidact
|
AutoDidact
Posted 3 Years Ago
|
Group: Forum Members
Last Active: 2 Weeks Ago
Posts: 2.1K,
Visits: 13.6K
|
I agree that morph Vox pro ,like most “filters”,is complete garbage. I used it in making “Galactus rising” ,years ago and it was a Nightmare.
AI assisted or even complete AI generated will be the way going forward.
I recently bought a one year subscription ($50) to Voicemaker.ai ( text to speech)to voice my CTA video narrator host and I am very pleased with the results
|
|
|
Wilby
|
|
Group: Forum Members
Last Active: 2 Years Ago
Posts: 131,
Visits: 479
|
AutoDidact (11/24/2022) I recently bought a one year subscription ($50) to Voicemaker.ai ( text to speech)to voice my CTA video narrator host and I am very pleased with the resultsI've also tried using the free version of Voicemaker.ai in the past and the outputted quality is excellent even in my native language. I've played around importing the audio file generated from Voicemaker.ai into MorphVox and the produced quality is acceptable for me especially for use of non-human characters. I haven't tried cleaning up the audio though. Maybe I have to play around with the software a bit to recognize more if it's weaknesses. So far, what I paid for it and what I got is ok. :) EDIT: Oops, sorry for the confusion. I meant Voicemaker.in and NOT Voicemaker.ai. :pinch:
|
|
|
Kelleytoons
|
Kelleytoons
Posted 3 Years Ago
|
Group: Forum Members
Last Active: Last Year
Posts: 9.2K,
Visits: 22.1K
|
Sigh - I wish folks in this thread would stop pushing AI GENERATED VOICES - this thread was all about using AI to alter existing voices. Robotic voices (and, folks, that is what they are and what they sound like to native language speakers - they could no more replace actors than my Roomba could replace a maid) are fine for folks who can't speak English (or whatever the language they are generated are). But they AREN'T the tech that folks started talking here and it's about as silly to bring them up here as it would be for me to mention using video and filtering it so it looked like animation to replace rendering in iClone. Let's try and keep this on track - we want something that takes a voice (a real, HUMAN voice) and alters it with AI processing so it sounds like someone else. I can promise you no matter what that dev says the Magic thingee does not do that, it's just a filter (and a bad one at that). And there are lots of things that generate AI voices that sound passable - but can't act. Again, it would be like me expecting to tell Roomba "Now please go make the bed."
Alienware Aurora R16, Win 11, i9-149000KF, 3.20GHz CPU, 64GB RAM, RTX 4090 (24GB), Samsung 870 Pro 8TB, Gen3 MVNe M-2 SSD, 4TBx2, 39" Alienware Widescreen Monitor Mike "ex-genius" Kelley
|
|
|
sonic7
|
sonic7
Posted 3 Years Ago
|
Group: Forum Members
Last Active: 2 Years Ago
Posts: 1.7K,
Visits: 19.4K
|
rr Nicely summed up Auto. I feel that ultimately "technology" leaves no one safe, since the mighty $ will, (and always has), empowered the minority to leverage returns to their own advantage.Then there's a largish number of 'wanna-be's' trying to play the same game and become king pin. Competition is a two edged sword with many casualties, yet we all (mostly) play the same game to a larger or lesser extent I think.But the only asset that's truly yours (that no-one can take or supplant with tech and Mike will agree with this) is 'story'. Your very own unique story. Changes in technology, I think, just means different ways of presenting/displaying your story, so story will always survive over time. A bit off topic - but just my 2c
--------------------------------------------------------------------------------------------------------------------------------------------------------
Please be patient with me ..... I don't always 'get it' the first time 'round - not even the 2nd time! :( - yikes! ... ● MSI GT72VR Laptop, i7 7700HQ 4-Core 3.8 GHz 16GB RAM; Nvidia 1070, 8GB Vram ● iClone-7.93 ● 3DXChange Pipeline 7.81 ● CC-3 Pipeline 3.44 ● Live Face ● HeadShot ● Brekel Pro-Body ● Popcorn FX ● iRAY ● Kinect V2 ● DaVinci Resolve17 ● Mixcraft 8.1
|
|
|
Wilby
|
|
Group: Forum Members
Last Active: 2 Years Ago
Posts: 131,
Visits: 479
|
Wilby (11/24/2022)
AutoDidact (11/24/2022) I recently bought a one year subscription ($50) to Voicemaker.ai ( text to speech)to voice my CTA video narrator host and I am very pleased with the resultsI've also tried using the free version of Voicemaker.ai in the past and the outputted quality is excellent even in my native language. I've played around importing the audio file generated from Voicemaker.ai into MorphVox and the produced quality is acceptable for me especially for use of non-human characters. I haven't tried cleaning up the audio though. Maybe I have to play around with the software a bit to recognize more if it's weaknesses. So far, what I paid for it and what I got is ok. :) Oops, sorry for the confusion. I meant Voicemaker.in and NOT Voicemaker.ai. :pinch: :crying::doze: I've tried using it's TTS feature to generate an audio voice file I used as source voice for some of the so-called "AI-powered" software to see how it would sound like. :P
|
|
|