How do I change things

[Toolfolks]

1. Change open AI voices.
2. Can I use more than 2 voices.
3. "conversation_style": ["casual", "humorous"], what are the parameters available.
4. What is the correct syntax for local PDF ?

Cheers

[souzatharsis]

Hi, 2. No. Currently only a two person conversation is possible. See how to guide for 1 and 3. https://github.com/souzatharsis/podcastfy/blob/main/usage/how-to.md 4. There is a bug on pdf to podcast generation. I'll fix it when I'm back to keyboard in a couple of days. Apologies for that.

…

On Sun, Oct 20, 2024, 3:14 PM Toolfolks ***@***.***> wrote: 1. Change open AI voices. 2. Can I use more than 2 voices. 3. "conversation_style": ["casual", "humorous"], what are the parameters available. 4. What is the correct syntax for local PDF ? Cheers — Reply to this email directly, view it on GitHub <#90>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADTMY3NBZJUQUVHIRGVNFL3Z4PQFPAVCNFSM6AAAAABQIWDWOWVHI2DSMVQWIX3LMV43ERDJONRXK43TNFXW4OZXGM2DGOBTGA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Toolfolks]

I'm struggling with the syntax

openai: default_voices: question: "onyx" answer: "echo" model: "tts-1-hd"

In app.py

custom_config = { "word_count": 300, "conversation_style": ["casual", "humorous"], "podcast_name": "Steve Warbys Info Podcast", "creativity": 0.7, "default_question_voice": "onyx", # Previously under openai.default_voices.question "default_answer_voice": "echo", # Previously under openai.default_voices.answer "tts_model": "tts-1-hd" }

What is the correct syntax please.

Also when using my app.py how do I set it to use 

custom_config.yaml

Cheers

[souzatharsis]

In podcastfy.ipynb there is a customization section where it shows how to pass conversation config as a dictionary in python application.

…

On Sun, Oct 20, 2024, 4:33 PM Toolfolks ***@***.***> wrote: I'm struggling with the syntax openai: default_voices: question: "onyx" answer: "echo" model: "tts-1-hd" In app.py custom_config = { "word_count": 300, "conversation_style": ["casual", "humorous"], "podcast_name": "Steve Warbys Info Podcast", "creativity": 0.7, "default_question_voice": "onyx", # Previously under openai.default_voices.question "default_answer_voice": "echo", # Previously under openai.default_voices.answer "tts_model": "tts-1-hd" } What is the correct syntax please. Also when using my app.py how do I set it to use custom_confi.yaml Cheers — Reply to this email directly, view it on GitHub <#90 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADTMY3LTE2KFKSIMZMASDODZ4PZOLAVCNFSM6AAAAABQIWDWOWVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTAOJZG44TMMQ> . You are receiving this because you commented.Message ID: ***@***.***>

[Toolfolks]

I have the following custom_config.yaml

`#word_count: 2000
conversation_style:

• "engaging"
• "fast-paced"
• "enthusiastic"
  roles_person1: "main summarizer"
  roles_person2: "questioner/clarifier"
  dialogue_structure:
• "Introduction"
• "Main Content Summary"
• "Conclusion"
  #podcast_name: "PODCASTFY"
  podcast_tagline: "Listen While you work"
  output_language: "English"
  engagement_techniques:
• "rhetorical questions"
• "anecdotes"
• "analogies"
• "humor"
  creativity: 1

text_to_speech:
default_tts_model: "openai"
elevenlabs:
default_voices:
question: "Chris"
answer: "Jessica"
model: "eleven_multilingual_v2"
openai:
default_voices:
question: "onyx"
answer: "onyx"
model: "tts-1-hd"
edge:
default_voices:
question: "en-US-JennyNeural"
answer: "en-US-EricNeural"
audio_format: "mp3"
temp_audio_dir: "data/audio/tmp/"
ending_message: "Bye Bye!"`

app.py

`from podcastfy.client import generate_podcast

custom_config = {
"word_count": 100,
"conversation_style": ["casual", "humorous"],
"podcast_name": "Steve Warbys Info Podcast",
"creativity": 0.7

}

generate_podcast(

urls =["https://www.dewalt.co.uk/products/storage","https://www.dewalt.co.uk/products/storage/backpacks","https://www.dewalt.co.uk/products/storage/mobile-tool-storage","https://www.dewalt.co.uk/products/storage/storage-combo-kits","https://www.dewalt.co.uk/products/storage/tool-bags-pouches-belts","https://www.dewalt.co.uk/products/storage/tool-organizers-bins","https://www.dewalt.co.uk/products/storage/trolleys","https://www.dewalt.co.uk/products/storage/workshop-storage"],

conversation_config=custom_config

)`

The generated voices are not changed.

What am I not understanding here please

[Toolfolks]

I have worked out that it uses conversation_config.yaml to set the defaults.
Also I can change some of these with

custom_config = {
"word_count": 6000,
"conversation_style": ["casual", "humorous"],
"podcast_name": "Steve Warbys Info Podcast"

}

conversation_config=custom_config,

How do I get it to use story_custom.yaml settings ?

[pedrobfig]

Hey everyone.

https://cloud.google.com/text-to-speech

Any chance or suggestion to use Google TTS? As far as I can see, pricing is much more interesting. They even offer the first million characters every month. We can also adjust pitch and speed of voices, something that I would like to adjust for different types of podcasts.

We can find there those two amazing voices from NotebookLM. ;-)

Cheers.

[souzatharsis]

Yes, there's a PR from a contributor that implemented exsctky this. I'll merge ASAP. Cutious: What are the two voices from notebooklm you found? Are those the journey voices?

…

On Wed, Oct 23, 2024, 3:15 PM pedrobfig ***@***.***> wrote: Hey everyone. https://cloud.google.com/text-to-speech Any chance or suggestion to use Google TTS? As far as I can see, pricing is much more interesting. They even offer the first million characters every month. We can also adjust pitch and speed of voices, something that I would like to adjust for different types of podcasts. We can find there those two amazing voices from NotebookLM. ;-) Cheers. — Reply to this email directly, view it on GitHub <#90 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADTMY3JA6P2Y7Z6Y27YRG7DZ47RVFAVCNFSM6AAAAABQIWDWOWVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCMBTGI3DQMQ> . You are receiving this because you commented.Message ID: ***@***.***>

[pedrobfig]

Hi again.

You can find them as "Journey" voices:

https://cloud.google.com/text-to-speech/docs/voice-types

Journey Voices, powered by the AudioLM engine, lets you create more engaging and empathetic speech for conversational applications.

You should also check "Casual" voices, that are very similar to the NotebookLM voices.