#Hello I have a problem with the quality of the generated audio why is very good on the site but whe

1 messages · Page 1 of 1 (latest)

wet umbraBOT
#
AI-generated Answer for hasna.chouikhi_04357

The difference in audio quality between the website and the API could be due to a few factors:

  1. Model Used: If you didn't specify a model in your API call, it defaults to Eleven English V1. Different models can produce varying results. You might want to check which model is being used in both cases.

  2. Voice Settings: Generative AI can produce slight variations in output. Settings like stability and similarity can impact the consistency of the audio. Lower stability and similarity settings can lead to more variation.

To improve consistency, ensure that the same model and settings are used across both platforms. Adjusting the stability and similarity settings might also help achieve more uniform results.

formal tundra
#

it is the same model and the same stability and similarity value

wet umbraBOT
#
AI-generated Answer for hasna.chouikhi_04357

The difference in audio quality between the website and the API could be due to a few factors. One possibility is that different models might be used if not explicitly set in your API call. By default, the API uses the Eleven English V1 model.

Another factor could be the inherent variability in AI-generated audio. Even with the same settings, slight differences can occur due to the non-deterministic nature of the AI. Adjusting the stability setting might help reduce these variations, but it could also lead to a more consistent yet potentially monotonous output.

If you continue to experience issues, consider checking the voice settings and ensuring that the correct model is being used. You can also experiment with different settings to see if it improves the quality.

formal tundra
#

but I specified the model : eleven_multilingual_v2 in the website and in my notebook

wet umbraBOT
#
AI-generated Answer for hasna.chouikhi_04357

There are a few reasons why you might experience differences in audio quality between the website and the API, even when using the same model and settings:

  1. Model Differences: Ensure that the model specified in your API call is exactly the same as the one used on the website. If you didn't set a model in your API call, it defaults to Eleven English V1.

  2. Voice Settings: Variations can occur due to the nature of generative AI. Even with the same input, different generations can produce slightly different results. Settings like stability and similarity can impact the variation in outputs. Lower stability and similarity settings can lead to more variation.

  3. API vs. Web Interface: Sometimes, there might be differences in how the API and web interface handle the same settings, which can affect the output quality.

For more detailed information on models and settings, you can refer to the Help Center articles on models and voice settings. If the issue persists, consider running additional tests or trying different settings to see if the quality improves.

formal tundra
#

any suggestion