Videos

For most personal music libraries, permanently changing the audio files is unnecessary. First check whether your player supports loudness normalization, such as ReplayGain, and whether it is enabled and working. That is usually enough to make playback volume more consistent.


Important Warning Before Modifying Audio Files!
These are destructive changes. The correction process can permanently alter audio files. If a script is pointed at the wrong folder, given the wrong settings, or allowed to overwrite the originals, the damage may be difficult or impossible to reverse.

Check whether permanent file changes are necessary
Before modifying any music files, check whether the playback software supports ReplayGain or another loudness-normalization system. Determine whether that feature is enabled and working correctly. Then determine whether permanent modification of the audio files is necessary or whether playback normalization alone can achieve the desired result.

Before modifying any audio files, have/make two complete backups of the original music library. Store one backup on a different drive or at a separate location and leave it untouched. The second backup may be used as a read-only source if corrected copies are needed later. Do not modify either backup in place. Corrected files should always be written to a separate working or output location. Test the process on a small group of files first. Verify the output before running it on thousands of tracks. Keep the CSVs, manifests, and logs so every change can be traced. If MP3 files must be re-encoded, use untouched originals whenever possible rather than processing files that have already been altered or re-encoded.

The safest approach is: Original library → two verified backups → test batch → verified corrected copy → working library
Do not begin large-scale correction until the backup has been checked and can actually be restored.

Prompts for Analyzing and Correcting a Music Library
Copy the prompts below into ChatGPT one at a time. Complete each step before moving to the next. When a step creates a CSV or other report, upload that file with the next prompt.
Adjust settings such as the loudness target to fit your system..

1. Create the scanning script
   Prompt – I have a Windows music library containing MP3 files. I want to analyze every audio file recursively without modifying anything. Create a PowerShell .ps1 script that scans the files and produces a CSV containing the full file path, integrated LUFS, peak or true-peak information, scan status and any errors. Before writing the script, tell me what software or command-line tools need to be installed. Include exact instructions for running the script in Windows PowerShell.
2. Analyze the CSV
   Prompt – Analyze this loudness-scan CSV. My target is −14 LUFS. Identify tracks that are at least 1 dB quieter than the target. For each track calculate how much gain would be required to approach −14 LUFS. Also check whether that increase would cause its true peak to exceed −1.5 dBTP. Do not modify any files. Create a table showing current LUFS, current peak, requested gain, maximum safe gain, projected LUFS, projected peak and whether the gain needs to be reduced to prevent clipping.
3. Find problems before changing anything
   Prompt – Examine these scan results for anything that should be reviewed manually before processing. Look for failed scans, missing measurements, unusually low or high loudness, suspicious peak values, files that cannot safely reach the target and anything else that appears abnormal. Separate these from the files that can be processed automatically.
4. Create a preflight
   Prompt – Using the approved correction list, create a PowerShell preflight script. It must not change any audio. Verify that every source file exists, destination paths are valid, there are no duplicate or conflicting filenames, proposed gain for each file is correct and projected peak levels remain within the selected ceiling. Create a CSV manifest showing exactly what would happen. Stop and report errors rather than guessing or silently skipping files. Do not proceed to the correction stage until the preflight results have been reviewed and approved.
5. Create the correction script
   Prompt – Now create the PowerShell script that performs the approved corrections. Apply corrections only to the tracks in the approved list. Raise each selected track toward −14 LUFS, but reduce the gain when necessary so the true peak does not exceed −1.5 dBTP. Read from the designated source files and write corrected files to a separate output location. Never modify the source or backup files in place. Preserve filenames, folder structure, metadata, tags and artwork where possible. Do not silently change the audio format, channel configuration, sample rate or encoding settings without explaining why. Log every successfully processed file, the applied gain, files whose gain was reduced because of peak headroom, and every failure. Do not process files that are not on the approved correction list. Before running the script, explain clearly whether the process requires MP3 re-encoding and what changes will be made to the audio files.
   
   *Important: If the files being corrected have already been modified, normalized, or re-encoded, do not process those altered MP3s again if untouched originals are available. Use the untouched backup files as the source and place the newly corrected versions into the working or streaming library. This avoids an unnecessary additional lossy MP3 encode.
6. Verify the corrected files
   Prompt – Create a second scan of the corrected files and compare it with the original CSV. Show the original LUFS, original peak, applied gain, new LUFS, new peak and difference from the original. Flag any files that did not change as expected, exceeded the intended peak ceiling, failed processing or remain significantly below the −14 LUFS target. Confirm that the number of output files matches the number successfully processed

ChatGPT-5.6 Sol.