Please look through the mod description and comments before writing a comment. Odds are that the description holds the information or someone else have asked about it before. Thanks <3
First of all: Love the mod, thanks for creating it!
But I have a question/issue: The voice lines seem to be generated on a low frequency (appearently 16khz)? Is there a way to increase the frequency and thus the quality of the generated lines? I know it's not really important for the standard voice but NaturalVoiceSAPIAdapter really would benefit from higher quality (44khz) lines.
A workaround for the low audio quality would be to set the default voice in the Text to Speech control panel dialog. Search "text to speech" in the Windows search, open "Speech Properties" dialog, and select any of the natural voice to be the default voice.
The default voice affects the default speech audio quality, and as the mod changes the voice without changing the audio quality, if the default voice only supports a lower quality, the selected voice will be limited to this quality as well. So changing the default voice will fix that. I'll also check how to fix that in code.
This is huge and prob one of the top 5 my mods i will not be able to play without. I know this is a big ask but do you think it would be possible to implement diferent tts voices ? (mostly askining because some of those tts seem rather old ones and not as crisp as it could be with today's ai tts spech)
I can't open the options window in the ModManager, I can see the underlined text and also the pay buttons and I can press them but there is no voice, does anyone have an idea why?
This is amazing for reading the books and notes around the world and voicing whatever lines you wish. Sometimes my eyes strain too much to read so thank you, so much!
Phonetic Dictionary seems to be broken in 1.1.2, the pause for ellipses ("...": " <silence msec=\"500\"/> ") broke the entire thing, it would only pronounce a couple letters at the end of the test lines. After removing that one, the rest of the narration now works, but the entries in the dictionary still don't. Did the structure change somehow perhaps?
Yes it changed to be entirely RegEx based. This is due to the fact that a lot of people had issues with it and that it seems most people are using NaturalVoiceSAPIAdapter which limits the use for a PhoneticDictionary. It still works, but you might have to encapsulate words in `\/b` to make them work.
And they still won't work... The ellipses and everything related to punctuation I got working (using regex), but, somehow, the actual phonemes are still not clicking...
Do you have any insight as to why? Sorry to bother, by the way!
(I have another question, and this is kind of a dumb one: is there a way to change the audio quality? In the NaturalVoicesAPI tts application, it lets you change the audio format to 48kHz 16 Bit stereo, and other options as well, and I was wondering if it's possible here? Thank you for your patience anyway!)
432 comments
Offline natural voices may stop working due to Microsoft changing their decryption key.
As a workaround, you can download an older but working version of the voice packs, using the instruction here.
But I have a question/issue: The voice lines seem to be generated on a low frequency (appearently 16khz)? Is there a way to increase the frequency and thus the quality of the generated lines? I know it's not really important for the standard voice but NaturalVoiceSAPIAdapter really would benefit from higher quality (44khz) lines.
The default voice affects the default speech audio quality, and as the mod changes the voice without changing the audio quality, if the default voice only supports a lower quality, the selected voice will be limited to this quality as well. So changing the default voice will fix that. I'll also check how to fix that in code.
I know this is a big ask but do you think it would be possible to implement diferent tts voices ? (mostly askining because some of those tts seem rather old ones and not as crisp as it could be with today's ai tts spech)
The male/female talker + narrator = great setup.
Until the KC speaks, then it's just weird he/she can't have a separate voice.
After removing that one, the rest of the narration now works, but the entries in the dictionary still don't.
Did the structure change somehow perhaps?
This is due to the fact that a lot of people had issues with it and that it seems most people are using NaturalVoiceSAPIAdapter which limits the use for a PhoneticDictionary. It still works, but you might have to encapsulate words in `\/b` to make them work.
"\/bArendae(?:'s)?\/b": "A-ren-day",
"\/bDeskari(?:'s)?\/b": "Deh-ska-ree",
"\/bMinagho(?:'s)?\/b": "Mee-nah-go",
"\/bKitrane(?:'s)?\/b": "Kee-trah-nee",
"\/bHulrun(?:'s)?\/b": "Hull-run",
"\/bSilaena(?:'s)?\/b": "Sill-ay-nuh",
"\/bSeelah(?:'s)?\/b": "See-lah",
"\/bAroden(?:'s)?\/b": "Ah-ro-den",
"\/bNabasu(?:'s)?\/b": "Nah-bah-soo",
"\/bDrezen(?:'s)?\/b": "Dreh-zen",
"\/bSosiel(?:'s)?\/b": "So-see-el",
"\/bDesna(?:'s)?\/b": "Dess-nah",
"\/bRimvi(?:'s)?\/b": "Rim-vee",
"\/bKemh(?:'s)?\/b": "Kehm",
And they still won't work... The ellipses and everything related to punctuation I got working (using regex), but, somehow, the actual phonemes are still not clicking...
Do you have any insight as to why? Sorry to bother, by the way!
(I have another question, and this is kind of a dumb one: is there a way to change the audio quality? In the NaturalVoicesAPI tts application, it lets you change the audio format to 48kHz 16 Bit stereo, and other options as well, and I was wondering if it's possible here? Thank you for your patience anyway!)