---
title: "Discussion on \"Preparing audio for speech-to-text with ffmpeg: what to change and what to leave alone\" | Hashnode"
description: "Discussion on \"Preparing audio for speech-to-text with ffmpeg: what to change and what to leave alone\". Containers, channels, loudness, long files, and the two edits that break timecodes"
url: "https://hashnode.com/posts/preparing-audio-for-speech-to-text-with-ffmpeg-what-to-change-and-what-to-leave-alone/6ad64431297b162a033e9c1b"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Preparing audio for speech-to-text with ffmpeg: what to change and what to leave alone](https://audiovtext.hashnode.dev/preparing-audio-for-speech-to-text-with-ffmpeg-what-to-change-and-what-to-leave-alone?utm_source=hashnode&utm_medium=feed)

Containers, channels, loudness, long files, and the two edits that break timecodes

Oct 19 · 6 min read · audiovtext.hashnode.dev  
Read the article

-   [#ffmpeg](https://hashnode.com/tag/ffmpeg)
-   [#audio](https://hashnode.com/tag/audio)
-   [#speech-recognition](https://hashnode.com/tag/speech-recognition)
-   [#transcription](https://hashnode.com/tag/transcription)
-   [#whisper](https://hashnode.com/tag/whisper)

## Written by

[Andrey Petrov](https://hashnode.com/@audiovtext)

@audiovtext

  
[View profile](https://hashnode.com/@audiovtext)

AI-powered audio & video to text transcription

AudioVText is an online AI service that transcribes audio and video into text. AudioVText supports 99 languages with Russian as its primary language, adds timecodes, and separates speakers.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.