@Tamasg Just wanted to congratulate you on version 2.99 of TGSpeechbox! The impulse pitch mode is almost, almost eloquence now. The only real issue is the "o" sounds in words like social and mode being a bit too long haha. But I can now go as fast as eloquence for predictable work. I've got to slow it down quite a bit for reading, but I'm not sure how much of that is just that I haven't fully adapted to the new voice, yet.
@fastfinge@Tamasg I'm trying it, and to my ears impulse pitch mode makes the engine much more robotic. Also I still have the t vs p issue - they do sound the same.
@fastfinge oh this is good to know! Yeah, the impulse rewrite was maybe 5 hours of straight work, lots of ear tuning. Now it's Klatt mode that sounds too flat, but there's languages that might benefit from that kind of prosody too. I just fixed a bunch of English tuning (not live yet, but will be soon) on words like card, shard, party, ETC. Way too fronted, now it's more central. So tuning is a slow drip-feed of boring ear-by-ear listening, plucking a phoneme and making it sound better based on what people describe. Ripping out Espeak's phonemizer only was surprisingly easy, so I'm not having to borrow the entire car just to get the GPS out of it, making it have really good responsiveness there, so I was quite happy to discover that once going MIT. Doesn't make it the best IPA phonemizer, but for sure the most diverse.
@fastfinge ah still Espeak, but not the entire lib and voice processing engine. It now statically compiles straight into the libs on mobile, so all the phonemizer stufff was like 17 C files that get linked into the project. Definitely would have been a big no-no with the GPLV2 V3 mismatch earlier.
@Tamasg@fastfinge I asked because, I can see with the latest version, words such as "sharp" and "wrap" sounds like "sha" and wra" with tgspeech box UK english, but the US english sounds right. When I tried with Espeak, both english sounds good.
@kaveinthran@Tamasg@fastfinge It sounds different to me, but then again I'm using TGSpeechbox with rate 50. I'm just getting used to it, so I'm doing high rates only on the test flight releases.
@kaveinthran@Rosalyn@Tamasg@fastfinge I could have done that, but a couple of things. My iPhone is on charge, currently I'm on the windows machine, and secondly I only have Rosalyn's telegram which ain't gonna work for iOS at the very least. Cheers for that.
@Rosalyn@TomGrant91@Tamasg@fastfinge You can add the TG speech voice to your rotor. The method is the same as how you add other voices to the rotor. The speech box should show up in the rotor settings.
@kaveinthran@TomGrant91@Tamasg@fastfinge I tried the impulse speaking style since Sam said that it reminded him more of eloquence, but I'm actually finding that I like the e-speak and legacy styles more. I'm on legacy right now.
@Rosalyn@TomGrant91@Tamasg@fastfinge I'm quite confused on how the settings apply. It would be great if you can educate me better. Let's say if I have two voices added to the rotor. One is US and another one is UK. If I go and change the speech mode, does it change both the voices that I have in the rotor, or how does it work generally?
@Rosalyn@TomGrant91@Tamasg@fastfinge Another thing is that when I switch the speech mode, I couldn't hear the changes take effect while I have the speech running. How do we let the change take effect?
@kaveinthran Good questions! So right now, yes, engine settings are global. If you change something like voice tilt or pitch mode, it applies to every voice you've added to the rotor. Per-voice settings are something I want to add down the line but haven't gotten to yet. On the second one, you actually found a bug! The sliders (voice tilt, breathiness, etc.) should take effect on the very next thing VoiceOver speaks, no restart needed. But I just checked and pitch mode specifically wasn't notifying the AU extension that something changed, so it would just silently ignore the switch. That'll be fixed in the next build. Thanks for catching it! @Rosalyn That's the audio component label, yeah. Right now it shows up as "TGSp" which is a four-character manufacturer code that iOS pulls from the extension metadata. I can see how that's confusing, especially with a screen reader. Renaming it to "TGSpeechBox" in the next update so it's actually readable. @TomGrant91@fastfinge
@kaveinthran@Rosalyn@TomGrant91@fastfinge Thanks all again for the feedback overnight. The voice picker and the rename to TGSpeechBox in your VoiceOver list of voices is complete in Build 12, along with that pitch selection bug fixed. Please check Testfflight for the latest to get it.
@Tamasg Also test flight link? I will test flight! Seriously though, you've really worked your entire ass off, here. And it shows in the quality. Once it's on mobile, I intend to really give a serious attempt at using this 24/7 as my daily driver. I suspect I'll succeed, when IOS isn't just constantly pulling me back to eloquence. So congrats on your success! Obviously the entire goal of your project was to make a speech engine that Sam likes. Because everything is always about me forever LOL JK.
