如何在Colab中用FFmpeg将.m2ts/.vob转MKV并保持画质及指定AVC编码等级

Got it, let's fix this properly. The All-in-One script you used likely defaulted to low-quality encoding presets, which explains the poor output and incorrect AVC profile. We'll use FFmpeg directly in Colab—it's the most reliable tool for this kind of task, and we can control every parameter precisely to meet your exact needs.

Step 1: Confirm Google Drive Access #

You mentioned you already have Drive mounted, but double-check it's accessible by running this snippet:

from google.colab import drive
drive.mount('/content/drive')

Follow the prompt to authenticate if the connection drops.

Step 2: Update FFmpeg (Optional but Recommended) #

Colab comes with FFmpeg pre-installed, but updating to the latest version ensures we have all the latest encoding features:

!apt update && apt install ffmpeg -y

Verify the installation works with:

!ffmpeg -version

Step 3: Core FFmpeg Conversion Command #

This command will convert your video while preserving resolution, maintaining source-level quality, and enforcing the High@L4.1 AVC profile:

!ffmpeg -i "/content/drive/MyDrive/Path/To/Your/Input.m2ts" \
  -c:v libx264 \
  -profile:v high \
  -level:v 4.1 \
  -crf 18 \
  -preset medium \
  -c:a copy \
  -c:s copy \
  -map 0 \
  "/content/drive/MyDrive/Path/To/Your/Output.mkv"

Let's break down the key parameters so you understand what each does:

• -c:v libx264: Uses the industry-standard x264 encoder for AVC/H.264 video
• -profile:v high -level:v 4.1: Explicitly sets the High@L4.1 profile you require
• -crf 18: Controls visual quality—18 is visually lossless (matches source quality almost exactly). Lower values = better quality but larger files; 18-22 is ideal for preserving source fidelity.
• -preset medium: Balances encoding speed and file size. Use fast or faster if you want quicker conversions (no quality loss, just slightly larger output files).
• -c:a copy -c:s copy: Copies audio and subtitle streams directly without re-encoding (preserves original audio quality and subtitle tracks)
• -map 0: Copies all streams (including multiple audio tracks, subtitles, and chapters) from the input to the MKV output, so nothing gets lost.

Step 4: Batch Conversion (For Multiple Files) #

If you have a folder full of .m2ts/.vob files, use this bash script to process them automatically:

#!/bin/bash
INPUT_DIR="/content/drive/MyDrive/Your/Input/Folder"
OUTPUT_DIR="/content/drive/MyDrive/Your/Output/Folder"

# Create output directory if it doesn't exist
mkdir -p "$OUTPUT_DIR"

# Process all .m2ts and .vob files in the input folder
for file in "$INPUT_DIR"/*.{m2ts,vob}; do
  if [ -f "$file" ]; then
    filename=$(basename -- "$file")
    filename_no_ext="${filename%.*}"
    output_file="$OUTPUT_DIR/$filename_no_ext.mkv"
    
    echo "Converting $file to $output_file..."
    ffmpeg -i "$file" \
      -c:v libx264 \
      -profile:v high \
      -level:v 4.1 \
      -crf 18 \
      -preset medium \
      -c:a copy \
      -c:s copy \
      -map 0 \
      "$output_file"
  fi
done

To run this in Colab:

1. Create a new text file in Drive (e.g., batch_convert.sh) and paste the script
2. Update INPUT_DIR and OUTPUT_DIR to match your actual folder paths
3. Execute it with:

!bash "/content/drive/MyDrive/Path/To/batch_convert.sh"

Step 5: Verify the Output #

After conversion, confirm the output meets your requirements by running:

!ffmpeg -i "/content/drive/MyDrive/Path/To/Output.mkv"

Look for lines like these in the output to confirm the profile and level:

Stream #0:0(eng): Video: h264 (High), yuv420p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], 23.98 fps, 23.98 tbr, 1k tbn, 47.95 tbc (default)
Profile: High
Level: 4.1

Pro Tip: Speed Up with GPU Acceleration #

Colab's GPU runtime will drastically cut down encoding time. Switch to it by:

1. Going to Runtime > Change runtime type
2. Selecting GPU under Hardware accelerator
3. Clicking Save

内容的提问来源于stack exchange，提问作者Adnan Katper