voice missing for bark?

[danielb2]

what could be causing the issues below?

$ ./ebook2audiobook.sh --headless --ebook ~/books/Frédéric\ Bastiat\ -\ The\ Law\ \(2011\).epub --tts_engine bark

EnvironmentLocationNotFound: Not a conda environment: /Users/daniel/work/ai/ebook2audiobook/python_env

v25.4.25 native mode
GPU is not available on your device!
Available Processor Unit: CPU
Running command: /Applications/calibre.app/Contents/MacOS/ebook-convert /Users/daniel/work/ai/ebook2audiobook/tmp/ebook-12cdc6ce-7ad5-4a7a-a4c1-51877393e0c3/d3f6fa782bd5786e6234c6fc0665e2f1/Frédéric Bastiat - The Law (2011).epub /Users/daniel/work/ai/ebook2audiobook/tmp/ebook-12cdc6ce-7ad5-4a7a-a4c1-51877393e0c3/d3f6fa782bd5786e6234c6fc0665e2f1/__Frédéric Bastiat - The Law (2011).epub
Conversion options changed from defaults:
  verbose: 1
  page_breaks_before: "//*[name()='h1' or name()='h2']"
  input_encoding: 'utf-8'
  smarten_punctuation: True
  disable_font_rescaling: True
  flow_size: 0
  output_profile: 'generic_eink'
  epub_version: '3'
1% Converting input to HTML...
InputFormatPlugin: EPUB Input running
on /Users/daniel/work/ai/ebook2audiobook/tmp/ebook-12cdc6ce-7ad5-4a7a-a4c1-51877393e0c3/d3f6fa782bd5786e6234c6fc0665e2f1/Frédéric Bastiat - The Law (2011).epub
Parsing all content...
Parsing OEBPS/index.html ...
Parsing OEBPS/cover.html ...
Parsing OEBPS/stylesheet.css ...
Parsing OEBPS/halftitle.html ...
Parsing OEBPS/title.html ...
Parsing OEBPS/copyright.html ...
Parsing OEBPS/foreword.html ...
Parsing OEBPS/chapter001.html ...
Parsing OEBPS/backcover.html ...
Reading TOC from NCX...
34% Running transforms on e-book...
Merging user specified metadata...
Detecting structure...
Flattening CSS and remapping font sizes...
PropertyValue: Missing token for production Choice(ColorValue, Dimension, URIValue, Value, variable, MSValue, CSSCalc, function): ('HASH', '#00007', 1, 131)
No content to parse.
PropertyValue: Unknown syntax or no value: #00007
CSSStyleDeclaration: Syntax Error in Property: background-color:#00007
PropertyValue: Missing token for production Choice(ColorValue, Dimension, URIValue, Value, variable, MSValue, CSSCalc, function): ('HASH', '#00007', 1, 131)
No content to parse.
PropertyValue: Unknown syntax or no value: #00007
CSSStyleDeclaration: Syntax Error in Property: background-color:#00007
Source base font size is 9.60000pt
Removing fake margins...
Found 2 items of level: div_1
Found 2 items of level: p_2
Found 603 items of level: p_1
Ignoring level p_2
div_1  left margin stats: Counter()
div_1  right margin stats: Counter()
p_1  left margin stats: Counter({'0': 561, '4em': 39, '1.5em': 3})
p_1  right margin stats: Counter({'0': 564, '4em': 39})
Removing page margins specified in the Adobe page template
Cleaning up manifest...
Trimming unused files from manifest...
Creating EPUB Output...
67% Running EPUB Output plugin
Splitting markup on page breaks and flow limits, if any...
                Splitting on page-break at id=ind
                Splitting on page-break at id=calibre_pb_0
                Splitting on page-break at id=fore
                Splitting on page-break at id=ch1
Removing anchor from TOC href: OEBPS/chapter001.html#ch1
The cover image has an id != "cover". Renaming to work around bug in Nook Color
Upgrading to EPUB 3...
EPUB output written to /Users/daniel/work/ai/ebook2audiobook/tmp/ebook-12cdc6ce-7ad5-4a7a-a4c1-51877393e0c3/d3f6fa782bd5786e6234c6fc0665e2f1/__Frédéric Bastiat - The Law (2011).epub
Output saved to   /Users/daniel/work/ai/ebook2audiobook/tmp/ebook-12cdc6ce-7ad5-4a7a-a4c1-51877393e0c3/d3f6fa782bd5786e6234c6fc0665e2f1/__Frédéric Bastiat - The Law (2011).epub


            ***************************************************************************************
                                            NOTE: THE WARNING
                                "Character xx not found in the vocabulary."
            MEANS THE MODEL CANNOT INTERPRET THE CHARACTER AND WILL MAYBE GENERATE AN HALLUCINATION
            TO IMPROVE THIS MODEL IT NEEDS TO ADD THIS CHARACTER INTO A NEW TRAINING MODEL.
            YOU CAN IMPROVE IT OR ASK TO A MODEL TRAINING DEVELOPER.
            ***************************************************************************************

Loading TTS tts_models/multilingual/multi-dataset/bark model, it takes a while, please be patient...
/Users/daniel/Miniforge3/lib/python3.12/site-packages/torch/nn/utils/weight_norm.py:143: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
  WeightNorm.apply(module, name, dim)
convertsion 0.00%: 0/833 Block 1 containing 833 sentences...
convertsion 0.00%: 0/833
'Jamie'
Traceback (most recent call last):
  File "/Users/daniel/work/ai/ebook2audiobook/lib/classes/tts_manager.py", line 489, in convert_sentence_to_audio
    audio_data = self.params['tts'].tts(
                 ^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniel/Miniforge3/lib/python3.12/site-packages/TTS/api.py", line 323, in tts
    wav = self.synthesizer.tts(
          ^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniel/Miniforge3/lib/python3.12/site-packages/TTS/utils/synthesizer.py", line 422, in tts
    outputs = self.tts_model.synthesize(
              ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniel/Miniforge3/lib/python3.12/site-packages/TTS/tts/models/bark.py", line 215, in synthesize
    history_prompt = load_voice(self, speaker_id, voice_dirs)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniel/Miniforge3/lib/python3.12/site-packages/TTS/tts/layers/bark/inference_funcs.py", line 66, in load_voice
    paths = voices[voice]
            ~~~~~~^^^^^^^
KeyError: 'Jamie'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/daniel/work/ai/ebook2audiobook/lib/functions.py", line 861, in convert_chapters_to_audio
    if tts_manager.convert_sentence_to_audio():
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniel/work/ai/ebook2audiobook/lib/classes/tts_manager.py", line 649, in convert_sentence_to_audio
    raise ValueError(e)
ValueError: 'Jamie'
Caught DependencyError: 'Jamie'

[ROBERT-MCDOWELL]

it seems an update issue, missing a wav file with the right name. I'm going to fix it and tell me if bark can detect it as for now it never been because of a weird implementation on coqui-tts

[danielb2]

I can see this voices/eng/adult/male/Jamie.wav. not sure if that's the one you're referring to

voices/eng/adult/male/Jamie.wav
voices/eng/adult/male/bark/Jamie/
voices/eng/adult/male/bark/Jamie/Jamie.npz

Maybe because it's missing in the bark folder?

[ROBERT-MCDOWELL]

no, in fact bark uses .npz file for voice reference, not wav. Jamie.wav has been used to create the .npz but the issue is bark cannot find the path of Jamie.npz for unknown reason. I tried to test many different path, relative or not, same.
as you have an enough power full machine to test bark, would you join me to discord and test in real time the modif I'm doing? (within 2 hours and other days)?

[danielb2]

what's the discord ?

nm, i found it on your profile. maybe should put that in this repo too

[danielb2]

i'm online for a little. about to go afk

[ROBERT-MCDOWELL]

ok let's just continue the discussion here, and will tell you when I updated.
first off all for the next test, better you fork my fork here https://github.com/ROBERT-MCDOWELL/ebook2audiobook
then you will be able to follow the update.

[danielb2]

which branch? using main still has the same error

KeyError: 'Jamie'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/daniel/work/ai/ebook2audiobook/lib/functions.py", line 876, in convert_chapters_to_audio
    if tts_manager.convert_sentence_to_audio():
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniel/work/ai/ebook2audiobook/lib/classes/tts_manager.py", line 658, in convert_sentence_to_audio
    raise ValueError(e)
ValueError: 'Jamie'
Caught DependencyError: 'Jamie'

[ROBERT-MCDOWELL]

there is onlly one branch, v25
gdb can help me to see where and how coqui-tts cannot find the path in it's code.
I passed hours to test various fixes but still persists.

[ROBERT-MCDOWELL]

are you on linux or mac? if linux do you have gdb installed?
try:
gdb --args ./ebook2audiobook.sh --headless --ebook ~/books/Frédéric\ Bastiat\ -\ The\ Law\ (2011).epub --tts_engine bark
gdb:> run
gdb:> bt full

and copy paste the terminal log.

[ROBERT-MCDOWELL]

brew install gdb. are you sure it's already installed? sorry you need to ad /usr/bin/zsh /path/to/ebook2audiobook.sh ......

[danielb2]

still getting the error about it not being in executable format.

$ gdb --args (pwd)/ebook2audiobook.sh --headless --ebook ~/resilio/books/mises.org/Frédéric\ Bastiat\ -\ The\ Law\ \(2011\).epub --tts_engine bark

GNU gdb (GDB) 16.3
Copyright (C) 2024 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "--host=aarch64-apple-darwin24.4.0 --target=x86_64-apple-darwin20".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
"/Users/daniel/work/ai/ebook2audiobook/ebook2audiobook.sh": not in executable format: file format not recognized
(gdb)

[ROBERT-MCDOWELL]

you forgot to add /usr/bin/zsh before /Users/daniel/work/ai/ebook2audiobook/ebook2audiobook.sh

[danielb2]

$ gdb --args /opt/homebrew/bin/zsh (pwd)/ebook2audiobook.sh --headless --ebook ~/resilio/books/mises.org/Frédéric\ Bastiat\ -\ The\ Law\ \(2011\).epub --tts_engine bark

GNU gdb (GDB) 16.3
Copyright (C) 2024 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "--host=aarch64-apple-darwin24.4.0 --target=x86_64-apple-darwin20".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /opt/homebrew/bin/zsh...
(No debugging symbols found in /opt/homebrew/bin/zsh)
(gdb) run
Don't know how to run.  Try "help target".
(gdb) r
Don't know how to run.  Try "help target".
(gdb)

according to gpt:

the problem is you're trying to debug an ARM-native zsh binary (/opt/homebrew/bin/zsh) using a GDB configured for x86_64 targets, which doesn't work natively on Apple Silicon.

[ROBERT-MCDOWELL]

Forgot recent mac is on ARM. it needs an equivalent

[ROBERT-MCDOWELL]

are you on discord to continue the tests?
or is the last git and release version solve it?

[danielb2]

i'm there. you can ping me. i did a pull after you mentioned the last update and it wasn't resolved.

[danielb2]

just tried again, fyi

[ROBERT-MCDOWELL]

ok I'm not surprise btw :(

[ROBERT-MCDOWELL]

could you please try again with last git or v25.5.19?
thanks

[danielb2]

it works 🎉