02-12-2021, 09:26 PM
(This post was last modified: 02-12-2021, 09:43 PM by Firebolt391d.)
This user Amber made a voice without even really lowering the grid at all, but I'm assuming it works because of the nature if the sine instrument sounds, and how it sustains note length. This is also presumably just a direct MP3 to MIDI file conversion (rather than an MP3 to sequence file conversion, which uses a Fast Fourier Transform function).
You'll notice that this voice sequence does not have nearly as much intelligibility as the voices generated using Liam's converter. I've noticed that Liam's converter also appears to be able to adjust note volume based on the gain/amplitude of each frequency, which seems to make a great deal of difference.
If you want to make your own voice completely from scratch using just the website, be sure to watch what notes you are using. There is a pattern to how it works, as there is a specific set of harmonics and subharmonics associated with the human voice (as well as pretty much anything else that makes any kind of resonating sound). I might recommend analyzing some of the other users' voice sequences, or looking up things such as:
Harmonics,
Overtones,
or Overtone Singing.
(I mentioned Overtone Singing because there are good videos about Overtone Singing that let you hear all of the notes that can be heard when extrapolating the human voice, which is essentially what Overtone Signing is).
Note that it will probably be incredibly difficult to actually make something that even sounds similar to a voice if you are just trying to do it from scratch. You can give it a try, but it seems pretty hard to me.
You'll notice that this voice sequence does not have nearly as much intelligibility as the voices generated using Liam's converter. I've noticed that Liam's converter also appears to be able to adjust note volume based on the gain/amplitude of each frequency, which seems to make a great deal of difference.
If you want to make your own voice completely from scratch using just the website, be sure to watch what notes you are using. There is a pattern to how it works, as there is a specific set of harmonics and subharmonics associated with the human voice (as well as pretty much anything else that makes any kind of resonating sound). I might recommend analyzing some of the other users' voice sequences, or looking up things such as:
Harmonics,
Overtones,
or Overtone Singing.
(I mentioned Overtone Singing because there are good videos about Overtone Singing that let you hear all of the notes that can be heard when extrapolating the human voice, which is essentially what Overtone Signing is).
Note that it will probably be incredibly difficult to actually make something that even sounds similar to a voice if you are just trying to do it from scratch. You can give it a try, but it seems pretty hard to me.