when I zip the file and throw it into MO2, then go to Arngeri, the most obivious NPC for testing this, and talk to him about any dialogue he might say "dragonborn", he says the original phrase he usually says with DRAGONBORN, proceeds with his normal dialogue, and then in his last sentence instead of saying what he should, he says the phrase he should say again, but with the name I choose. why is this happens? is it an error via MO2, or is it an error in his lines on the voice files?
So I'm not sure what I've done wrong. I followed the instructions, got the Data folder in the say_my_name folder. I then made a zip of that folder renamed to SayMyName_"chosen name", and installed the zip as a mod in MO2. When I talk to Arngeir however, I'm still hearing Dragonborn. I checked in the batch file, the line does exist for the replacement "you have learned so much, "___" growing your gift too quickly could be dangerous"
Should I have not renamed the zip file? or do I need to zip a subfolder instead of the entire zip folder?
I've gone through the creation process and made absolutely sure that I had all the correct boxes checked. Now, when I tell Arngeir "I'd like to learn more about the Voice", he'll say "You have learned so much already, Dragonborn..", continue for a couple lines, then instead of finishing his speech, the audio says "You have learned so much already, Nathan, growing your gift too quickly would be dangerous". While he says that, the subtitles continue with his normal speech.
I also tried using one of the samples, Bob McBobface, and the same thing happened
I have followed the instructions in the video. Whenever I run the "Say my name" script, xVASynth ignores the enabled ARPAbet custom dictionary and uses some alleged pronunciation for my custom string, which is not included in the CMU database but sounds exactly how an English speaker would attempt to pronounce it. As if some sort of AI overrides all dictionaries. I have also tried adding my string to the default dictionary to no avail. If I change the pronunciation of an existing word in the default dictionary, that setting is ignored as well.
I'm on v3.0.3 and run xVASynth as a normal user (running as an admin is known to cause issues). My custom string matches that in the batch file and includes no special characters. I have the latest voice libraries installed for Skyrim. What am I doing wrong? And where does xVASynth even take its alleged pronunciations from?
Notwithstanding the above, I appreciate your work put into this mod. As long as I can find a custom string that an English speaker would pronounce similary to how I want my character's non-English name to sound, the result is almost as intended. But then I don't see why we have ARPAbet in the first place.
If the AI voice was only during the name, It would be an instant download. But the fact that it has to replace the whole line just makes the awkward robot voice too noticeable, I'm sorry. If it was doable to make a version that only replaced that specific word, then it would be a must-use.
Man, I really don't wanna go through all of this just to get rid of the dragonborn lines.
Could any of you with a female character upload your own generated files. Changing my character's name to yours would be enough, I just wanna be a non-dragonborn.
I'm an casual mod user and an mod list specialist so how do you actually work say my name and i also have an problem lol see my skyrim absorbed 4 mods i don't use anymore path of sorcery thane weapons reborn arcanum i really hope those 2 mods that my skyrim absorbed into the game interface anyway i done the batch file correctly no errors i just wanna know how to enable it so it would work i use nexus mod manager thanks.
I keep getting an error when I try to synthesize the batch.
Traceback (most recent call last): File "server.py", line 335, in do_POST File "python\xvapitch\model.py", line 326, in infer_batch File "C:\Users\Marco\Downloads\xVASynth v3.0.0 Main app-44184-3-0-0-1684850032\v3.0.0\.\resources\app\python\xvapitch\xvapitch_model.py", line 223, in infer_advanced return self.infer_using_vals(logger, plugin_manager, cleaned_text, text, lang_embs, speaker_embs, pace, None, None, None, None, None, None, pitch_amp=pitch_amp) File "C:\Users\Marco\Downloads\xVASynth v3.0.0 Main app-44184-3-0-0-1684850032\v3.0.0\.\resources\app\python\xvapitch\xvapitch_model.py", line 292, in infer_using_vals x, x_emb, x_mask = self.text_encoder(input_symbols, x_lengths, lang_emb=None, stats=False, lang_emb_full=lang_emb_full) File "torch\nn\modules\module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "C:\Users\Marco\Downloads\xVASynth v3.0.0 Main app-44184-3-0-0-1684850032\v3.0.0\.\resources\app\python\xvapitch\xvapitch_model.py", line 602, in forward x = torch.cat((x_emb, lang_emb_full), dim=-1) RuntimeError: Sizes of tensors must match except in dimension 2. Expected size 45 but got size 1 for tensor number 1 in the list.
Please if anyone knows what stupid thing I've done, I'd be grateful. I've been trying to get this to work for hours.
172 comments
Should I have not renamed the zip file? or do I need to zip a subfolder instead of the entire zip folder?
Any help would be appreciated
I also tried using one of the samples, Bob McBobface, and the same thing happened
I'm on v3.0.3 and run xVASynth as a normal user (running as an admin is known to cause issues). My custom string matches that in the batch file and includes no special characters. I have the latest voice libraries installed for Skyrim. What am I doing wrong? And where does xVASynth even take its alleged pronunciations from?
Notwithstanding the above, I appreciate your work put into this mod. As long as I can find a custom string that an English speaker would pronounce similary to how I want my character's non-English name to sound, the result is almost as intended. But then I don't see why we have ARPAbet in the first place.
Could any of you with a female character upload your own generated files. Changing my character's name to yours would be enough, I just wanna be a non-dragonborn.
Would appreciate it a lot.
Traceback (most recent call last):
File "server.py", line 335, in do_POST
File "python\xvapitch\model.py", line 326, in infer_batch
File "C:\Users\Marco\Downloads\xVASynth v3.0.0 Main app-44184-3-0-0-1684850032\v3.0.0\.\resources\app\python\xvapitch\xvapitch_model.py", line 223, in infer_advanced
return self.infer_using_vals(logger, plugin_manager, cleaned_text, text, lang_embs, speaker_embs, pace, None, None, None, None, None, None, pitch_amp=pitch_amp)
File "C:\Users\Marco\Downloads\xVASynth v3.0.0 Main app-44184-3-0-0-1684850032\v3.0.0\.\resources\app\python\xvapitch\xvapitch_model.py", line 292, in infer_using_vals
x, x_emb, x_mask = self.text_encoder(input_symbols, x_lengths, lang_emb=None, stats=False, lang_emb_full=lang_emb_full)
File "torch\nn\modules\module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Users\Marco\Downloads\xVASynth v3.0.0 Main app-44184-3-0-0-1684850032\v3.0.0\.\resources\app\python\xvapitch\xvapitch_model.py", line 602, in forward
x = torch.cat((x_emb, lang_emb_full), dim=-1)
RuntimeError: Sizes of tensors must match except in dimension 2. Expected size 45 but got size 1 for tensor number 1 in the list.
Please if anyone knows what stupid thing I've done, I'd be grateful. I've been trying to get this to work for hours.
This! Banging my head trying to figure out why I have to brush up on my Python, when I just need to learn to read!