Fantastic! I'm so happy to see this release! I was stunned by your Narsis release. Just incredible quality. I can't imagine how much work this was. I'm just finishing up my TR audiobook mod using a fork of the Chatterbox TTS and it's been a lot of work but looking forward to getting it out in the open and getting feedback. I've managed to do almost all the remaining books from TR, SHoTN, and PT. Even done the newspapers like the Canyon Echo. Had to learn a lot about Python and stuff but hopefully it will be worth it!
Hopefully, this week before I head off on holiday, then i'll work on another version after feedback when i'm back - my main issue is that the fork of Chatterbox i'm using is very picky about unusual punctuation so things like "--" can throw up errors and the double dash is used a lot in TR so there's quite a bit of text editing before I can run the TTS.
Been playing with your mod and it really brings Cyrodil alive! Still think it's the best sounding mod like this in terms of naturalistic speech and havn't heard any errors or other issues!
Some things I noticed after a brief run around Anvil: some of the deliveries are good (like Lusha the pawnbroker), while others could still do with a bit of polishing. For instance, Nadia the Curio merchant near the gates has a very shrill initial greeting, the services topic only activates the last sentence in the paragraph, and, as another user mentioned earlier, some of the deliveries can be a bit slow.
Found the issue regarding the missing first parts of the topic. Happened for some of the longer texts unfortunately. As for the occasional high pitched/shrill voice and slow paced delivery, those are a bit harder to solve but I'll do my best to fix or at least improve them in an upcoming version.
If you (or anyone else) comes across more issues, please let me know!
just from watching the trailer, a lot of the standard lines seem a little slow in delivery. do you know what i mean? the bespoke character lines actually seemed pretty good, like the bounty hunter, but the standardised lines (which are written in a robotic kind of way) need a little juice. its funny because i think those lines like "looking for someone in particular" or "services" would be just as hard for a human VA to make sound natural. the way i see it their intended more an approximation of a conversation than actual dialogue so i wonder if it would be better in general to avoid voicing those lines or maybe they could eventually be rewritten slightly. haha probably a little beyond the scope of what your doing now but i think in general a middle ground between "fully voiced" and "Quest greetings" would be interesting. just thinking out loud really.. its all such wierd new stuff
I think it would be easier for updates if you just put the OpenMW plugin in a separate archive from the voice files themselves, that way you won't have to upload both everytime you need to fix/improve on something
Easier for me for sure, you're right. Just wanted to keep it simple for the Users. One download, and done. Plus I'm hoping to add more versions in the future (with quest voice greetings added, or as a standalone), and to avoid confusion I'm trying to stick with this "only download one file" approach.
16 comments
Do you already know when you're going to release it?
Been playing with your mod and it really brings Cyrodil alive! Still think it's the best sounding mod like this in terms of naturalistic speech and havn't heard any errors or other issues!
Found the issue regarding the missing first parts of the topic. Happened for some of the longer texts unfortunately. As for the occasional high pitched/shrill voice and slow paced delivery, those are a bit harder to solve but I'll do my best to fix or at least improve them in an upcoming version.
If you (or anyone else) comes across more issues, please let me know!
PD: ill be waiting for skyrim too hehe
its funny because i think those lines like "looking for someone in particular" or "services" would be just as hard for a human VA to make sound natural. the way i see it their intended more an approximation of a conversation than actual dialogue so i wonder if it would be better in general to avoid voicing those lines or maybe they could eventually be rewritten slightly. haha probably a little beyond the scope of what your doing now but i think in general a middle ground between "fully voiced" and "Quest greetings" would be interesting.
just thinking out loud really.. its all such wierd new stuff
Just wanted to keep it simple for the Users. One download, and done. Plus I'm hoping to add more versions in the future (with quest voice greetings added, or as a standalone), and to avoid confusion I'm trying to stick with this "only download one file" approach.
Thanks for pointing it out!