syng3 beta team

Official support for: xoxos
RELATED
PRODUCTS

Post

seedpress - had a thought that may help you - since you're not interested in my five part entheogen course, now at 20% off ;)

spend time listening to birds. observe the fundamental, perceived fundamental is fine. observe with your pristine engineering ears.

if you have the experience of travelling to a distant point on the globe where human dialect pitches discernibly differently, you will observe more informedly that pitch phrasing is indeed contiguous between the two species.

work with just pitch for a while :) before apprectiating phrase duration and other parameters. they be talking. eventually you'll figure the part about enculturated people hilariously disenfranchised from responding to anything that does not bear authority to inform them. go and measure it with your real ear actually there sense, g/l :)
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.

Post

well, i guess we are here. i'm finally releasing this. when did i say i made most of it? 2011?

did the gui damascus -
Image

no write ups or manual so next week i guess. it's imperfect but tired of sitting on it, there may be updates if people chew at me for stuff. but it's down to stuff like slight adjustment of volume of phonemes.

(also being lazy about portamento.. se's default midi only reads the value at noteon, whereas continuous update would be better, but i don't feel like wrangling another module now and you can retrigger the note if it's important to do. i'm lazy, i suck. try and buy my stuff if i care.)

for good use, you will need to put time in, tucking consonants and massaging transitions. the good news is betterment comes with use, it's kind of like a paintbrush, you can paint with it, instead of it painting you. kind of. maybe just if you think like me. it's good! well pretty good, i'd use it.

it is what it is.. 7/8 band source/filter model with a single set of phonemes.. transitions between phonemes aren't always natural but i like my results and the potential i see ymmv.

what i did about the "hey, that's my voice!" problem is knocking out 0.08 octave either side of zero. i know that may be a frustrating feature in a tool, but i don't think it will reasonably impact the user experience unless someone automates a formant bend through this point.

but i could have made male and female two separate vst, here, you have a slight block in the shift range.

the row of knobs at the bottom can be assigned as mod sources if your host provides smoother than midi modulation for gui parameters.

the rightmost portion of the gui is an useful graphic when sequencing on channel 2.. it offers no interactive functionality, but the colours are nice and flat, and i always like how that brings out the natural reflectivity of the monitor and if you're stoned or visually attuned you like trip out woah flat pixels are special.

i know right now, you are looking, thinknig you can hardly wait to own this monstro-erm, nice thingy!
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.

Post

holy shit it sounds great! a while back i was all hopped up on taking apart a ten yr old voice synth i found for reaktor, and jumpy for any news about this, so i almost posted a comment about the nifty mechanism for shifting formants i found inside, because it seemed to be an improvement over syng (still not 100% sure, though i was able to get the reaktor one to sound a lot better after jamming some round pegs in square holes :D ). anyway, i had written the most concise step by step description i could manage, before deciding, whatever my intentions, trying to give dsp tips in this particular forum would probably come off like a teenager trying to give a lesson on manners. which is something i try to avoid, so good to know it was my better judgement, and not paranoia informing the situation. as soon as i get my weeks and weeks of back pay from my dangerously stupid, lazy, and possibly criminal employers, its a sure bet that i will load up one of those prepaid visas i always use for online purchases (proactive paranoia has its place).

would it be rude to send virsyn a virtual singing telegram song about how and why E CANTOR has been hopelessly undercut and rendered obsolete while they were serenely licking the last bit gravy off their plates?

Post

this is a big surprise :hyper: I missed something, because there is no news on the release of syng in the xoxos forum?there is only the beta thread?
I'm not a great expert on vocal synthsesis but I'm going to find and try .congratulations on the release :tu:

Post

there may be updates adjusting the particularities of various phonemes if people get on me or i have convincing experiences - i've sequenced maybe 28? measures with it and my experience is that it will need modulation for a satisfactory performance.. sometimes the same phoneme is too loud, other times not enough. some transitions especially need tucking and fluffing.

but! also nice results can be achieved.

a few nights ago i tried to sequence some rap for a demo. never got to modulation, the first part sounded pretty straight, and could have passed for a pitch shifted vocal. the second part, which rhymed with the first part, was hard to distinguish, even though it used mostly the same sounds. with effort put into enunciation i'm confident i could have found something close to expressivity i had in mind, some compromise of intent and medium, but i couldn't care that much about it honestly.

might get you there, crap to use.. much less candy like than vocaloid. don't know if other people can really interface with it enough to yield. but it's there if you are sufficiently wealthy to indulge such interests.

syng3 thread
http://www.kvraudio.com/forum/viewtopic ... 1&t=437138

an FFT model would be much better.. i was thinking, non-vst app that allows user to record or edit a selection of files into a data set, then the instrument uses the same source as this (BLIT osc and filtered noise) with FFT with frame adaption/nodging for the plosives.

but a lot of work, would need a "pro" model to record a source to sell it with, and if i did all that work the industry would still piss on me so f**k em. :)
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.

Post

i think i beat you to the punch but now i just wish i knew something about spectral morphing

Post Reply

Return to “xoxos”