Since I've helped quite a bit in the community, I figured I give something back.
I managed to track down the couple versions of the DECTalk mail listing archive. This should have all or most of the files from around 2005 as well as later files.
The mailing listing messages are also included in MBOX format within the archive.
https://mega.nz/file/o2sAWRTY#ysX6riODONqg9yvTjSSWvMvUj0RTTF3q2Bt3fOrgIA8
=> More informations about this toot | More toots from rommix0@mastodon.social
@rommix0 As someone with their slightly more than fair share of obscure speech synthesis experience and knowledge, I really, really appreciate your work and especially your willingness to share and converse with all of us. There aren't very many people at all with your drive and resources among the few of us who are blind and care about this stuff for one reason or another. Thank you for everything and for being so easy-going with everyone!
=> More informations about this toot | More toots from guilevi@dragonscave.space
@rommix0 On a completely different note, here's something that you might enjoy and is not easy to come by.
This is SPG Lectores, and as far as I know it will only run in its own little speak window type thing. It was made by Silverio Pardo, a very peculiar blind TVI (teacher for the visually impaired) here in Spain, not as long ago as it may seem by the aesthetics of the program and the sound of the synthesizer. The UI is entirely in Spanish and the synth will only speak Spanish, but I figured you may find it interesting to mess around with anyway. The synthesis method is quite different, it's a kind of strange mix of recorded consonants and heavily-processed-but-also-technically-prerecorded periodic waves for the voiced sounds. Really interesting stuff, but his way of explaining things is... Nonstandard at best!
There's technical documentation (also in Spanish) in the archive, and I have no doubt that Silverio would send over the source code were I to contact him, assuming he's still alive and well.
Also copying in @datajake1999 and @x0 because this looks right up their alley, and I don't recall if Jake has this archived yet!
https://spg_util.droppages.com/download/spg_lp.zip
=> More informations about this toot | More toots from guilevi@dragonscave.space
@guilevi @rommix0 @x0 I remember an older version of this program from the DECtalk Archive Resilio Sync folder.
=> More informations about this toot | More toots from datajake1999@dragonscave.space
@datajake1999 @rommix0 @x0 There were a good few versions, and the sound has actually changed pretty significantly across them. This is likely the last version we'll see from Silverio.
=> More informations about this toot | More toots from guilevi@dragonscave.space
@guilevi now undergoing a challenge to make a voice for this
=> More informations about this toot | More toots from spacepup@dragonscave.space
@spacepup OH god. Good luck. Seriously the voice creation part was really, really really confusing to me. The way he writes documentation and structures his UIs just baffles me.
=> More informations about this toot | More toots from guilevi@dragonscave.space
@guilevi i'm getting it so far. The easiest part are the vowels. The hardest? Gosh, the consonents, they must be just that short otherwise it would lag the speeech behind
=> More informations about this toot | More toots from spacepup@dragonscave.space
@spacepup I must be really dumb because I could never get half the options to even work
=> More informations about this toot | More toots from guilevi@dragonscave.space
@guilevi i figured most of it out. 1. You have to record all your phonemes at least a second long, even the vowels. Save that phoneme wav file in 22.5khz, mono
=> More informations about this toot | More toots from spacepup@dragonscave.space
@guilevi oh and you press control shift v to start the voice creation process
=> More informations about this toot | More toots from spacepup@dragonscave.space This content has been proxied by September (ba2dc).Proxy Information
text/gemini