OK. I have an even better idea.
-
@Tamasg Oh my god. That's really damn trippy. I love it
@Bri lol it told me, there's like 3 or 4 examples of different tones mixed together with it, this should be interesting. If it got it right each would have higher pitch matching that of the binaural carrier as it mixes in. LOL based on the earlier glottal washing machine experiments, then the idea that just like the WinTalker Fred voice, we'll just mix in hex-dumped generations of the wave-files looping as part of the glottal source in that same way.
-
@Bri lol it told me, there's like 3 or 4 examples of different tones mixed together with it, this should be interesting. If it got it right each would have higher pitch matching that of the binaural carrier as it mixes in. LOL based on the earlier glottal washing machine experiments, then the idea that just like the WinTalker Fred voice, we'll just mix in hex-dumped generations of the wave-files looping as part of the glottal source in that same way.
@Tamasg yep, that's accurate.
-
@Tamasg yep, that's accurate.
@Bri darn it. Ok I listened to one of these. But it doesn't slow down or speed up depending on the vowels. So. This is not quite the true thing. ha. it would actually sound, well, way different. almost like a demonic binaural beat as the resonators chewon the beat and the vowels drive it underneath what it is. OMG. but it's building that now. Ha.
-
@Bri darn it. Ok I listened to one of these. But it doesn't slow down or speed up depending on the vowels. So. This is not quite the true thing. ha. it would actually sound, well, way different. almost like a demonic binaural beat as the resonators chewon the beat and the vowels drive it underneath what it is. OMG. but it's building that now. Ha.
@Tamasg I still thought it was trippy
-
@Tamasg I still thought it was trippy
@Bri ahahahaha! maybe all 4 in order especially ha. Someone could probably build that as a standalone app, though not part of any screen reader, almost like a book reading app that let you paste in text and just generated two wave files, changed out the glottal source each time and did all the mixing. What a worthless project but. April Fool's isn't far away. Many crazy things can happen especially on that day. I had this and a few other ideas cooked up. And as we get closer to it, well... A PR is likely to appear with something strange in it. Maybe not a full merge and update, I'm not quite that cruel.
-
OK. I have an even better idea. I'm not going to listen to this single file until after work ends in an hour, I want to surprise and shock myself. I just put all the wave files it created into Goldwave, it told me which order to put them in for you all into a single file to post, I did it, told it thank you.
Y'all can listen to it first, I have no desire to terrify myself. Astral_Adam.wav, something we could never do in reality because NVDA to my knowledge does not support stereo output for synthesis audio, and neither do the other screen readers.
But, somehow, Claude used the python tools I have in the repo to cook this up. No idea how good it is, how bad it is. Maybe it'll be clippy. So you've all been warned. A file I have not heard because I want to keep working and finishing a deadline by workday's end. So no. I would rather terrify the world before I do myself.@Tamasg Ummm... the hell? Forget the binaural aspect for a moment, what happened to the voice itself lmao.
-
@Tamasg Ummm... the hell? Forget the binaural aspect for a moment, what happened to the voice itself lmao.
@bscross32 ah yeah, it's like, a primitive copy of it. Ha. The python-only scripts in tools are there for testing, so like, no pitch model at all, it just loads the raw phonemes and feeds it through a Python-written simulator of speechbox
also why everything stays one pitch in that one. Great for testing a phoneme's raw sound, not so great for using as an actual synthesizer. -
OK. I have an even better idea. I'm not going to listen to this single file until after work ends in an hour, I want to surprise and shock myself. I just put all the wave files it created into Goldwave, it told me which order to put them in for you all into a single file to post, I did it, told it thank you.
Y'all can listen to it first, I have no desire to terrify myself. Astral_Adam.wav, something we could never do in reality because NVDA to my knowledge does not support stereo output for synthesis audio, and neither do the other screen readers.
But, somehow, Claude used the python tools I have in the repo to cook this up. No idea how good it is, how bad it is. Maybe it'll be clippy. So you've all been warned. A file I have not heard because I want to keep working and finishing a deadline by workday's end. So no. I would rather terrify the world before I do myself.@Tamasg Lol that was insane. Listened without headphones first but will do so again later and report back.
-
OK. I have an even better idea. I'm not going to listen to this single file until after work ends in an hour, I want to surprise and shock myself. I just put all the wave files it created into Goldwave, it told me which order to put them in for you all into a single file to post, I did it, told it thank you.
Y'all can listen to it first, I have no desire to terrify myself. Astral_Adam.wav, something we could never do in reality because NVDA to my knowledge does not support stereo output for synthesis audio, and neither do the other screen readers.
But, somehow, Claude used the python tools I have in the repo to cook this up. No idea how good it is, how bad it is. Maybe it'll be clippy. So you've all been warned. A file I have not heard because I want to keep working and finishing a deadline by workday's end. So no. I would rather terrify the world before I do myself.wellp, after work now. I told Claude, Hey Claude, why not bring back the old Glottal-tables-experiments branch! From January! And see what you can do if you re-did this entire thing, in CTypes Python, not with the Formant Trajectory tool. The result to me is actually less dystopian sounding than Claude's first try. But now, good times. While it did all this unsupervised nonsense, I could do my work, not focus on issues for once. And lay the idea of Speechbox doing Binaural beats to more rest for myself.
-
R relay@relay.infosec.exchange shared this topicR relay@relay.mycrowd.ca shared this topic
-
wellp, after work now. I told Claude, Hey Claude, why not bring back the old Glottal-tables-experiments branch! From January! And see what you can do if you re-did this entire thing, in CTypes Python, not with the Formant Trajectory tool. The result to me is actually less dystopian sounding than Claude's first try. But now, good times. While it did all this unsupervised nonsense, I could do my work, not focus on issues for once. And lay the idea of Speechbox doing Binaural beats to more rest for myself.
@Tamasg I could imagine myself having fun with that if it was in NVDA or the SAPI version lmfao
-
@Tamasg I could imagine myself having fun with that if it was in NVDA or the SAPI version lmfao
@alexchapman ha, I guess nobody thought of ever building the possibility of running two speech synthesis threads, one for each ear like that. I wish! Would be cool as heck, and then you could do like, individual tuning on each ear for all settings, and read anything you'd want for hours that way, would be so fun in some endless future of coding possibilities.
-
@alexchapman ha, I guess nobody thought of ever building the possibility of running two speech synthesis threads, one for each ear like that. I wish! Would be cool as heck, and then you could do like, individual tuning on each ear for all settings, and read anything you'd want for hours that way, would be so fun in some endless future of coding possibilities.
@Tamasg Hahaha yeah that'd be fun lol.
-
P pixelate@tweesecake.social shared this topic
