Should I use Volume Normalisation for audiobook that is in English?

[open wyvern]

AI-generated Answer for iamtypec

If you're experiencing issues with downloading or converting your audiobook, it might be helpful to disable the "Normalize volume to meet audiobook standards" option in your project settings. This setting can sometimes affect the audio quality or cause errors during the download or conversion process. If you continue to have issues, you can try other troubleshooting steps like clearing your cache, using Incognito mode, or trying a different web browser.

[karmic trail]

if i disable them, i have to regenerate the audio?

[open wyvern]

AI-generated Answer for iamtypec

Yes, if you disable the "Normalize volume to meet audiobook standards" option, you will need to regenerate the audio. Make sure to click 'Convert' to save your audio after making any changes.

[karmic trail]

then I will get charged again?

[open wyvern]

AI-generated Answer for iamtypec

Yes, if you disable the "Normalize volume to meet audiobook standards" option and regenerate the audio, it will use your character quota, as charges are incurred when generating audio. Each time you edit or regenerate text, the character quota is deducted.

[karmic trail]

but it's a 8000 words book.. :/

[open wyvern]

AI-generated Answer for iamtypec

If you disable volume normalization and need to regenerate the audio, you will indeed incur charges again, as credits are deducted each time speech is generated. However, you can regenerate your voiceover up to two more times at no cost if the text, voice, and model remain unchanged, and the regeneration is done within 2 hours of the initial generation. Make sure not to refresh the page during this process.

For large projects like an 8000-word book, consider splitting the text into smaller sections to manage the regeneration process more effectively. If you have further questions or need assistance, feel free to reach out.

[karmic trail]

then should I use volume normalisatio or no?

[open wyvern]

AI-generated Answer for iamtypec

If you're experiencing issues with downloading or converting your audiobook, it might be helpful to disable the "Normalize volume to meet audiobook standards" option. Disabling this feature has been known to resolve some problems related to audio processing.

However, if you disable it, you will need to regenerate the audio, which could incur additional charges. If you're concerned about the cost, consider testing with smaller sections of your book first to see if the issue persists.

Ultimately, whether to use volume normalization depends on your specific needs and whether it affects the quality or downloadability of your audiobook. If the normalization feature is causing issues, it might be best to disable it temporarily.

[karmic trail]

I want to know how much credit is charged per words for each model

[open wyvern]

AI-generated Answer for iamtypec

For non-turbo models like English V1, Multilingual V1, and Multilingual V2, one text character equals 1 credit. For Turbo models, one text character equals 0.5 credits for self-serve plans.

If you disable features like volume normalization and need to regenerate the audio, credits will be deducted again for the new generation. To manage credits efficiently, consider breaking the text into smaller sections and regenerating only the parts that need changes.

[karmic trail]

so if i have 110,000 credits, I can generate 110,000 words?

[open wyvern]

AI-generated Answer for iamtypec

Credits are typically deducted based on the number of characters, not words For most models, 1 credit is equal to 1 character. However, some models, like Turbo models, cost less—specifically, 1 credit for every 2 characters.

If you have 110,000 credits, you can generate 110,000 characters with standard models. If you use Turbo models, you could generate up to 220,000 characters. Keep in mind that spaces and punctuation also count as characters.

[karmic trail]

I have 52520 credits

[open wyvern]

AI-generated Answer for iamtypec

The number of words you can generate with your credits depends on various factors, including the complexity of the text and the number of regenerations. Typically, 100,000 credits allow for around 100 minutes of audio output.

To manage your credits efficiently, consider breaking your text into smaller sections and regenerating only the parts that need improvement. This approach can help you make the most of your available credits.

[karmic trail]

since I have 52250 characters, its better I use turbo v2?

[open wyvern]

AI-generated Answer for iamtypec

With 52,520 characters, using the Turbo V2 model be a good choice, especially if you're focusing on English and need faster generation. Turbo models are optimized for speed and cost 1 credit for every 2 characters, allowing you to maximize your credits.

[karmic trail]

compare turbo v2 , turbo v2.5 & multilingual v2 for 52,250 characters.

[open wyvern]

AI-generated Answer for iamtypec

For generating audio with 52,250 characters, here's a comparison of the models:

1. Turbo v2: This model is optimized for low-latency English text-to-speech. It costs 0.5 credits per character, so generating 52,250 characters would use 26,125 credits.

2. Turbo v2.5: Similar to Turbo v2, this model also costs 0.5 credits per character and supports 32 languages. It is ideal for real-time applications. Generating 52,250 characters would also use 26,125 credits.

3. Multilingual v2: This model supports 28 languages and is known for its accuracy and stability. It costs 1 credit per character, so generating 52,250 characters would use 52,250 credits.

Given your credit balance, using either Turbo v2 or Turbo v2.5 would be more cost-effective for generating 52,250 characters.

[karmic trail]

whats the difference between turbo v2 & turbo v2.5?

[open wyvern]

AI-generated Answer for iamtypec

Turbo v2.5 is optimized for low-latency applications and supports 32 languages, making it suitable for real-time conversational use cases. It is 25% faster than Turbo v2 and 300% faster than Multilingual v2. Turbo v2, on the other hand, focuses exclusively on English and is ideal for speed-critical English-only use cases. Both models offer different character limits per request, with Turbo v2.5 allowing up to 40,000 characters and Turbo v2 up to 30,000 characters.

[karmic trail]

what?? turbo v2.5 only allow maximum 40k character?!

[open wyvern]

AI-generated Answer for iamtypec

Yes, when using the API, the Turbo v2.5 model allows a maximum of 40,000 characters per request, which is approximately 40 minutes of audio. If you have 52,250 characters, you would need to split your text into smaller sections to fit within this limit. Turbo v2 allows up to 30,000 characters per request.

For longer texts, consider using the Projects feature, which can handle extremely long-form content more efficiently.

[karmic trail]

oh you mean per generate maximum 40k?

[open wyvern]

AI-generated Answer for iamtypec

Yes, that's correct. With the Turbo v2.5 model, you can generate up to 40,000 characters in a single request. If you have more than 40,000 characters, you'll need to split your text into multiple requests.

[karmic trail]

so if my text has 52k character, I can still use it with turbo v2.5

[open wyvern]

AI-generated Answer for iamtypec

Yes, you can use Turbo v2.5 for your 52,250-character text. However, since the maximum character limit per generation is 40,000, you'll need to split your text into smaller segments and generate them separately.

[karmic trail]

what voice is recommended for turbo v2.5? I want a Narrative voice with American Accent and talks in english.

[open wyvern]

AI-generated Answer for iamtypec

For a narrative voice with an American accent using the Turbo v2.5 model, you can explore the Voice Library to find a suitable voice. While specific recommendations aren't provided, you can use filters in the Voice Library to narrow down voices based on language and accent preferences. If you need further assistance in selecting a voice, feel free to reach out.

[karmic trail]

But I choose one of the voice, it's recommended for Turbo v2.. Im using turbo v2.5

[open wyvern]

AI-generated Answer for iamtypec

If the voice you selected is fine-tuned for Turbo v2, it might not perform optimally with Turbo v2.5. For better results, it's recommended to use a voice specifically trained on Turbo v2.5. You can verify this in the My Voices section or via the API. If you are generating content solely in English, you might consider using the Turbo v2 model, as it is optimized for English and offers similar performance.

[karmic trail]

but on the voice library it didnt say which voice is finetuned for which model.

[open wyvern]

AI-generated Answer for iamtypec

To find a voice that is fine-tuned for the Turbo v2.5 model, you can check the My Voices section or use the API to verify the compatibility. If a voice is specifically trained on Turbo v2.5, it should perform better with that model. If you're unsure, you might want to try different voices to see which one gives you the best results with Turbo v2.5.

[karmic trail]

release date for elevenlabs turbo v2.5

[open wyvern]

AI-generated Answer for iamtypec

The ElevenLabs Turbo v2.5 model was updated 3 months ago. It is optimized for low-latency applications and supports 32 languages, including English.

[karmic trail]

released?

[open wyvern]

AI-generated Answer for iamtypec

Turbo v2.5 was released around three months ago. It is optimized for low-latency applications and supports 32 languages, including English.

[karmic trail]

so I must used voices that was created 3 months ago? so that its fine tuned for turbo v2.5?

[open wyvern]

AI-generated Answer for iamtypec

The release date for ElevenLabs Turbo v2.5 is not specified in the provided content. However, you can use any voice available in the Voice Library, regardless of when it was created, as long as it suits your needs. If you have specific requirements for a narrative voice with an American accent, you can explore the Voice Library to find a suitable option.

[karmic trail]

so I must used voices that was created 3 months ago? so that its fine tuned for turbo v2.5?

[open wyvern]

AI-generated Answer for iamtypec

To ensure the best performance with the Turbo v2.5 model, it's recommended to use voices that are specifically fine-tuned for it. You can verify which voices are trained for Turbo v2.5 by checking the "My Voices" section in your account. If a voice is not fine-tuned for Turbo v2.5, it might not perform optimally.

If you're unsure about the fine-tuning status of a voice, you can check the recommended models for each voice in your account. This will help you select the most suitable voice for your needs.

[karmic trail]

On the Add a new Project page, if I use a voice as default voice, can I still change it after that? or i will be unable to change anymore

[open wyvern]

AI-generated Answer for iamtypec

When you set a default voice for a new project, it will apply to newly created chapters. However, you can still change the voice for existing chapters by manually adjusting the voice settings for each paragraph. If you want to change the voice for the entire project, you might consider creating a new project with the desired voice.

[karmic trail]

what i mean is this

[open wyvern]

AI-generated Answer for iamtypec

When you set a default voice for a new project, it will apply to new chapters you create. However, you can still change the voice for existing chapters by adjusting the voice settings for each paragraph individually. If you want to change the voice for the entire project, you might need to recreate the chapters with the new voice settings.

[karmic trail]

Can I delete a project?

[open wyvern]

AI-generated Answer for iamtypec

Yes, you can delete a project. To do this, navigate to the Projects Page. Once there, you should be able to access the "Remove Project" option. If you're using a keyboard for navigation, ensure you follow the correct keystrokes to access project options. If you encounter any issues, try again later, as sometimes functionality may resolve itself.

[karmic trail]

Does it cost my credit if I import my document file in projects??

[open wyvern]

AI-generated Answer for iamtypec

Importing a document file into a project does not consume credits. Credits are only used when you generate or regenerate audio from the text.

[karmic trail]

Are there any preview voice in project? like preview first before I generate

[open wyvern]

AI-generated Answer for iamtypec

Currently, there is no option to preview audio without using credits. When you generate audio, credits are deducted. However, you can manage your credits more effectively by using the "Projects" feature, which allows you to regenerate specific paragraphs instead of the entire text. This can help minimize credit usage.

[karmic trail]

For Voice Settings, Should I turn on Use Global Voice Settings? In Projects

[open wyvern]

AI-generated Answer for iamtypec

The "Use Global Voice Settings" option in Projects allows you to apply the same voice settings across your entire project. If you want consistency in voice settings throughout, it's a good idea to turn it on. However, if you need different settings for specific sections, you might want to keep it off and adjust settings individually.

[karmic trail]

Voice settings like Stability & Clarity + Similarity Enhancement?

[open wyvern]

AI-generated Answer for iamtypec

In Projects, using the "Global Voice Settings" option allows you to apply the same voice settings across all your text-to-speech generations within that project. This can be useful for maintaining consistency in the voice output.

Stability and Clarity + Similarity Enhancement are key settings you can adjust:

• Stability: Controls the emotional range and consistency of the voice. Lower settings result in more variation and emotion, while higher settings produce a more stable, monotone voice.

• Clarity + Similarity Enhancement: This setting helps the AI closely match the original voice. High settings may replicate artifacts from low-quality audio.

Whether to use global settings depends on your specific needs for the project. If you want consistent voice characteristics throughout, enabling global settings is beneficial. If you need different settings for different parts, you might want to adjust them individually.

[karmic trail]

What are the tips for Voice Settings?

[open wyvern]

AI-generated Answer for iamtypec

Here are some tips for adjusting voice settings to achieve better results:

1. Stability: Increasing stability can lead to more consistent output, but it might make the voice sound more monotone. Adjust this setting based on your preference for emotional range versus consistency.

2. Similarity: Set the similarity to around 75-90 to maintain a consistent voice that closely matches the original. Be cautious, as high settings might replicate artifacts from low-quality audio.

3. Style Exaggeration: Keep this setting at 0 to avoid mispronunciations and maintain stability.

4. Text Segmentation: Break your text into shorter sections, preferably under 1,000 characters, to maintain voice quality and consistency.

5. Professional Voice Cloning: Consider using a Professional Voice Clone or high-quality voices from the Voice Library for more consistent and high-quality results.

Experiment with these settings to find the best combination for your specific needs.

[karmic trail]

How to download voice from library? Preview voice

[open wyvern]

AI-generated Answer for iamtypec

To download your previously generated voiceovers, navigate to the Speech Synthesis tab and click on history. There, you'll find a list of your past voiceovers available for download in MP3 or WAV format at no extra cost. If you have any additional questions, feel free to reach out for assistance.

[karmic trail]

not my voiceover... I want to download voice from library.

[open wyvern]

AI-generated Answer for iamtypec

Currently, you can only download voiceovers that you have generated. The Voice Library itself does not offer a direct download option for preview voices. If you have any other questions or need further assistance, feel free to ask!

[karmic trail]

tell me about stability, similarity and style exaggeration

[open wyvern]

AI-generated Answer for iamtypec

The voice settings in ElevenLabs allow you to tailor the performance of AI-generated voices. Here's a brief overview:

1. Stability: This slider controls the emotional range and consistency of the voice. Lower stability settings introduce more variation and emotion, which can result in a more lively performance. However, setting it too low might lead to erratic speech patterns. Higher stability settings produce a more consistent and monotone output, which is useful for serious or formal content.

2. Similarity: This setting dictates how closely the AI should replicate the original voice. A high similarity setting ensures the AI adheres closely to the original voice, but if the original audio quality is poor, it might reproduce unwanted artifacts or background noise.

3. Style Exaggeration: This setting amplifies the style of the original speaker, adding more character to the voice. However, it can affect the stability of the voice, leading to mispronunciations or inconsistent speech speed. It's generally recommended to keep this setting at 0 to avoid these issues.

Experimenting with these settings can help you achieve the desired voice performance for your project.

[karmic trail]

Can I turn on Volume normalisation after creating projects?

[open wyvern]

AI-generated Answer for iamtypec

Yes, you can turn on Volume Normalization after creating a project However, if you make changes to the settings, you may need to regenerate the audio, which could incur additional credit charges.

[karmic trail]

Does it matter if I use bulletpoints in my project vs if I dont use bulletpoints in my projects?

·        Self-regulation: Managing your emotional reactions so you don’t respond impulsively. This involves taking a step back when you’re upset and thinking about the consequences of your actions before reacting.

Will the voice generated be different?

[open wyvern]

AI-generated Answer for iamtypec

Using bullet points in your project text should not significantly affect the voice generation. The AI focuses on the content and structure of the text rather than the formatting. However, the way the text is structured can influence the pacing and intonation slightly, so it's always a good idea to preview the generated audio to ensure it meets your expectations. If you notice any issues, you can adjust the text or settings accordingly.

[karmic trail]

Why is my download canceled halfway when I download a chapter in projects??

[open wyvern]

AI-generated Answer for iamtypec

If your download is being canceled halfway, here are a few steps you can try to resolve the issue:

1. Disable Volume Normalization: Go to your Project settings and uncheck the "Normalize volume to meet audiobook standards" option. This has helped in similar cases.

2. Browser Troubleshooting: Try using a different browser or clear your current browser's cache. You can also try downloading in Incognito mode.

3. Check for Problematic Text: Sometimes, specific text elements like very large paragraphs, special characters, or syntax commands can cause issues. Try editing the text slightly and regenerating the audio.

4. Screen Recording: If the problem persists, consider creating a screen recording of the issue using a tool like the Jam Chrome extension. This can help in diagnosing the problem further.

If these steps do not resolve the issue, please reach out for further assistance.

[karmic trail]

if i disable volume normalization and download project, i will get charged again?

[open wyvern]

AI-generated Answer for iamtypec

Disabling the "Normalize volume to meet audiobook standards" feature should not result in additional charges when downloading a project. If you encounter issues with downloads, such as cancellations, try disabling this feature as it has resolved similar issues for others. If problems persist, consider other troubleshooting steps like clearing your browser cache or trying a different browser.

[karmic trail]

are there any way to slow down voice?

[open wyvern]

AI-generated Answer for iamtypec

Currently, there is no direct feature to adjust the speech speed in the text-to-speech application. However, you can try a few workarounds to slow down the voice:

1. Add punctuation: Incorporating punctuation marks such as commas, dashes, and full stops can create pauses and help control the perceived speed.

2. Utilize a Voice Clone: Opt for a voice clone that speaks at a slower pace as an alternative.

3. Create pauses manually: Use the syntax <break time="1.5s" /> to introduce natural pauses in speech. The AI can handle pauses of up to 3 seconds, but excessive breaks may cause the speech to speed up or introduce noise.

A speed control feature is planned for future updates, but no specific date is available yet.

[karmic trail]

what does the grey line in project means?

[open wyvern]

AI-generated Answer for iamtypec

I'm unable to provide specific details about the grey line in projects. Could you please provide more context or details about where you're seeing this grey line?

[karmic trail]

infront of the text I want to generate voice. theres a grey bar