> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kits.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Recording High-Quality Datasets

> 30-60 minutes of clean and varied audio will result in the highest-quality voice models.

## Clean audio

The highest-quality voice models are recorded:

* with a [quality microphone](https://docs.kits.ai/train/microphones-audio.md) into an [audio interface](https://docs.kits.ai/train/microphones-audio.md)

* in a [noise-free setting with limited room reverberations](https://docs.kits.ai/train/recording-environment.md)

* with [correct and consistent mic placement](https://docs.kits.ai/train/microphone-placement.md)

* with [volume peaks between -9db and -3db](https://docs.kits.ai/train/volume-level.md)

And [processed](https://docs.kits.ai/train/audio-post-processing.md):

* with consistent dynamics across the whole dataset

* with light EQ to remove any muddiness, hiss, etc.

* with compression/limiting to smooth out peaks

* with no reverb, delay, or doubling


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.