video-generator/scripts/combine_video_audio.py — one-step final video generation, or two-step audio replacement
video-generator/scripts/make_voiceover_macos.py — creates timed WAV/optional MP3 voiceover on macOS
video-generator/scripts/make_voiceover_windows.py — creates timed WAV/optional MP3 voiceover on Windows
video-generator/subtitles/ — optional place to keep SRT files
video-generator/output/ — optional place to save generated audio/video files

Install

Windows

winget install Gyan.FFmpeg

Install Python from python.org or Microsoft Store, then reopen PowerShell and check:

ffmpeg -version
python --version

macOS

brew install ffmpeg

Python is already available on most Macs. If needed, install it from python.org.

One-step flow: create final video from SRT

Use this when you already have the tutorial video and the .srt narration file. It creates temporary voiceover audio, removes the video's original audio, exports the final MP4, and deletes temporary voice parts automatically.

Windows

python video-generator/scripts/combine_video_audio.py `
  --video "video-generator/input/your_video.mp4" `
  --srt "video-generator/subtitles/your_subtitles.srt" `
  --voice "Microsoft David Desktop" `
  --rate 0 `
  --output "video-generator/output/final_video.mp4"

macOS

python video-generator/scripts/combine_video_audio.py \
  --video "video-generator/input/your_video.mp4" \
  --srt "video-generator/subtitles/your_subtitles.srt" \
  --voice "Daniel" \
  --rate 185 \
  --output "video-generator/output/final_video.mp4"

--voice and --rate are optional. If omitted, the script uses the default voice and speed for your operating system.

Two-step flow: create voiceover

Use this when you want to review or edit the generated audio before combining it with the video.

Windows

python video-generator/scripts/make_voiceover_windows.py `
  --srt "video-generator/subtitles/your_subtitles.srt" `
  --voice "Microsoft David Desktop" `
  --rate 0 `
  --wav "video-generator/output/voiceover.wav" `
  --mp3 "video-generator/output/voiceover.mp3"

macOS

python video-generator/scripts/make_voiceover_macos.py \
  --srt "video-generator/subtitles/your_subtitles.srt" \
  --voice "Daniel" \
  --rate 185 \
  --wav "video-generator/output/voiceover.wav" \
  --mp3 "video-generator/output/voiceover.mp3"

--mp3 is optional. Use WAV when combining with video.

Two-step flow: combine video + audio

This removes the original video audio and uses the generated voiceover.

Windows

python video-generator/scripts/combine_video_audio.py `
  --video "video-generator/input/your_video.mp4" `
  --audio "video-generator/output/voiceover.wav" `
  --output "video-generator/output/final_video.mp4"

macOS

python video-generator/scripts/combine_video_audio.py \
  --video "video-generator/input/your_video.mp4" \
  --audio "video-generator/output/voiceover.wav" \
  --output "video-generator/output/final_video.mp4"

Use absolute paths if files are outside this folder.

Voice and speed options

Windows voice list

Add-Type -AssemblyName System.Speech
$synth = New-Object System.Speech.Synthesis.SpeechSynthesizer
$synth.GetInstalledVoices() | ForEach-Object { $_.VoiceInfo.Name }

Good tutorial voices: Microsoft David Desktop, Microsoft Mark if available.

Windows speed uses -10 to 10:

python video-generator/scripts/combine_video_audio.py `
  --video "video-generator/input/your_video.mp4" `
  --srt "video-generator/subtitles/your_subtitles.srt" `
  --output "video-generator/output/final_video.mp4" `
  --voice "Microsoft David Desktop" `
  --rate 2

To add voices: Settings > Time & language > Speech > Manage voices > Add voices.

macOS voice list

say -v '?'

Good tutorial voices: Daniel, Reed (English (US)), Eddy (English (UK)).

Speed example:

python video-generator/scripts/combine_video_audio.py \
  --video "video-generator/input/your_video.mp4" \
  --srt "video-generator/subtitles/your_subtitles.srt" \
  --output "video-generator/output/final_video.mp4" \
  --voice "Daniel" \
  --rate 190

Optional advanced options

These options are not needed for normal usage.

--audio-output "output/voiceover.wav" — saves a copy of the generated voiceover WAV in one-step mode.
--keep-temp — keeps temporary voice parts for debugging. By default, temporary files are deleted.
--volume 1.2 — changes voiceover volume. Default is 1.0.
--silence-padding 0.15 — adds a small pause after each caption if the SRT time slot has enough space. Default is 0.0.
--sample-rate 48000 — changes generated audio sample rate. Default is 44100.
--channels 1 — changes generated audio channels. Use 1 for mono or 2 for stereo. Default is 2.
--reencode-video — re-encodes video with H.264 if the copied video output has compatibility issues.

Example:

python video-generator/scripts/combine_video_audio.py \
  --video "video-generator/input/your_video.mp4" \
  --srt "video-generator/subtitles/your_subtitles.srt" \
  --output "video-generator/output/final_video.mp4" \
  --audio-output "video-generator/output/voiceover.wav" \
  --volume 1.2 \
  --silence-padding 0.15 \
  --sample-rate 48000 \
  --channels 2

Notes

If a caption has too much text for its time slot, the script keeps the SRT timing and trims the speech. Fix that by increasing the subtitle duration, shortening the sentence, or increasing speed.