| --- Log opened Sat Dec 14 00:00:36 2024 |
| 00:16 | | Elainefoxx [Lilith@Nightstar-pjf.02a.121.209.IP] has joined #code |
| 00:17 | | Elainefo1x [Lilith@Nightstar-pjf.02a.121.209.IP] has joined #code |
| 00:45 | | Elainefo1x [Lilith@Nightstar-pjf.02a.121.209.IP] has quit [[NS] Quit: ttfn] |
| 00:45 | | Elainefoxx [Lilith@Nightstar-pjf.02a.121.209.IP] has quit [Connection closed] |
| 00:47 | | Elainefoxx [Lilith@Nightstar-pjf.02a.121.209.IP] has joined #code |
| 01:15 | | Pinkhair [Pink@Nightstar-sp9qe9.sub-97-181-200.myvzw.com] has joined #code |
| 01:17 | | PinkAFK [Pink@Nightstar-sp9qe9.sub-97-181-200.myvzw.com] has quit [Ping timeout: 121 seconds] |
| 01:30 | | Elainefoxx [Lilith@Nightstar-pjf.02a.121.209.IP] has quit [Ping timeout: 121 seconds] |
| 03:32 | | Degi_ [Degi@Nightstar-bsa8mu.pool.telefonica.de] has joined #code |
| 03:33 | | Degi [Degi@Nightstar-btp4mh.pool.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 03:33 | | Degi_ is now known as Degi |
| 03:38 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has joined #code |
| 03:38 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 04:31 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has quit [Connection closed] |
| 08:37 | | Elainefoxx [Lilith@Nightstar-66tkno.bchsia.telus.net] has joined #code |
| 08:54 | | Elainefoxx [Lilith@Nightstar-66tkno.bchsia.telus.net] has quit [Ping timeout: 121 seconds] |
| 12:05 | | Emmy [Emmy@Nightstar-qo29c7.fixed.kpn.net] has joined #code |
| 12:49 | | macdjord is now known as macdjord|slep |
| 14:42 | | Elainefoxx [Lilith@Nightstar-66tkno.bchsia.telus.net] has joined #code |
| 15:02 | <&ToxicFrog> | abudhabi: I've been using Piper for TTS with good results |
| 15:02 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has joined #code |
| 15:02 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:29 | < abudhabi> | ToxicFrog: Thanks. I'll check that out if my current foray into cocqui TTS fails. |
| 15:45 | <&ToxicFrog> | *nodnod* |
| 15:45 | <&ToxicFrog> | Piper is specifically what my HA setup uses for household announcements and stuff. |
| 16:40 | < abudhabi> | Well, piper was very easy to install. |
| 16:40 | < abudhabi> | Now let's see if it chokes on 1MB text files. |
| 17:28 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has quit [Connection closed] |
| 17:59 | < abudhabi> | Took a bit over an hour, but it produced a file! Sounds almost good. |
| 18:08 | < abudhabi> | ToxicFrog: Do you know where the reference for the command line arguments is? |
| 18:09 | < abudhabi> | I look at `--noise_w NUM phoneme width noise (default: 0.8)` and wonder what that even means. |
| 19:56 | | macdjord|slep is now known as macdjord |
| 19:59 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has joined #code |
| 19:59 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 20:07 | <@macdjord> | abudhabi: Well, 'phoneme' is a single pronounceable sound, so I'm thinking 'phoneme width noise' will be how much random variation there is in the time it takes to say a single one. |
| 20:08 | <@macdjord> | Idea being that if they're all exactly the same length, the voice will sound robotic and artificial, because every time it says the same word it will sound /exactly/ the same. |
| 21:34 | < Alek> | add the same concept to midi files, produce less artificial-sounding music? |
| 21:35 | <&McMartin> | I mean, MIDI *already* needn't do that, because it's got precise timing and the whole point of MIDI as a protocol is you are capturing the actions of humans playing instruments |
| 21:36 | <&McMartin> | Also the phoneme width thing reminds me of the concept of morae in Japanese, which is more precise about long vowels being about lengths |
| 21:36 | < Alek> | perhaps too-precise timing? |
| 21:36 | < Alek> | ANYWAY |
| 21:37 | < Alek> | *flops* |
| 21:37 | <&McMartin> | (well, it's what you'd do to MIDI generated by a tracker program that says "yes, emit exactly 180 BPM and put all notes exactly on the beat") |
| 21:44 | <@celticminstrel> | "long vowel" in English isn't really a thing. I mean, yeah, it's the term used, and in fact the vowels referred to as "long" do tend to be longer than the ones called "short", but… the actual distinguishing factor is something more like "tense/lax" rather than "long/short". Tho I don't remember which is which off the top of my head (are tense ones long, or short? I forget). |
| 21:45 | <@celticminstrel> | Tho "tense/lax" in itself isn't really a precise term either. |
| 21:46 | <@celticminstrel> | Japanese has not only "long vowels" but also "long consonants", tho TBH I'm not even sure what a long T is… |
| 21:46 | <@celticminstrel> | In practice it seems like it sounds as if they'd extended the preceding vowel…? |
| 21:51 | <&McMartin> | https://en.wikipedia.org/wiki/Mora_(linguistics)#Japanese seems to describe it under "special mora" but this is beyond my ken |
| 22:05 | < Alek> | https://cdn.discordapp.com/attachments/677303567630991360/1317459557043273831/keyboard-interrupt.png?ex=675ec353&is=675d71d3&hm=c27db81b249b0201716e9365744bbcbda1f6712468f62cc4ad1985960b4e89e6&; |
| 22:28 | | * McMartin guesses, clicks, is right! |
| 22:28 | <&McMartin> | A classic |
| --- Log closed Sun Dec 15 00:00:37 2024 |