How to Preserve Metadata During File Conversion

Every file format stores metadata differently. JPEG files use EXIF for camera settings and GPS data, IPTC for captions and credits, and XMP for a broader range of custom fields. PNG files were designed without EXIF support and instead store metadata in text chunks (tEXt, zTXt, iTXt). Word documents embed properties in OPC package fields. MP4 videos carry metadata in atoms. When you convert between formats, the conversion tool has to decide what to do with metadata that exists in the source but has no equivalent container in the destination.

Most tools take the path of least resistance: they drop what does not fit. A JPEG-to-PNG converter typically reads the pixel data, compresses it as PNG, and writes a new file. The EXIF block, which contains your camera model, shutter speed, and location, gets left behind because PNG historically had no standard place to put it. The tool did not maliciously strip your data. It simply never read it in the first place.

Three patterns account for most metadata loss:

• Format incompatibility. The destination format lacks a container for the metadata type. PNG has no native EXIF block. Plain text has no metadata container at all. WAV files support limited metadata compared to FLAC.
• Tool behavior. Many conversion tools prioritize speed and file size over metadata fidelity. Browser-based converters and bulk processing scripts often skip metadata entirely. Even professional tools like Photoshop will strip EXIF from certain export paths unless you check the right option.
• Lossy re-encoding. Some converters rebuild the file from scratch rather than remuxing or transcoding in place. This destroys not just metadata but also any embedded profiles, thumbnails, and sidecar references.

The first step in preserving metadata is understanding which conversion you are performing and whether the destination format can hold what the source format contains.

Not all conversions are equally destructive. Some format pairs share metadata standards, making preservation straightforward. Others are fundamentally incompatible, requiring you to extract metadata separately before converting.

Here is how common conversion pairs handle metadata:

Two takeaways stand out. First, conversions within the same media type (image to image, document to document) preserve more metadata than cross-type conversions (video to audio). Second, formats designed for archiving, like TIFF for images and PDF/A for documents, tend to preserve metadata by design. The Library of Congress specifically recommends PDF/A for long-term document preservation because the ISO 19005 standard requires embedded XMP metadata.

When your conversion pair appears in the "usually lost" column, you will need to extract metadata before converting and reapply it afterward.

Keep Your File Metadata Organized After Conversion

Fast.io's Metadata Views extract and verify metadata fields across PDFs, images, and documents. Describe what you need in plain language and get a searchable, sortable spreadsheet. 50 GB free, no credit card required.

Sign Up Free

The right tool makes the difference between losing metadata and keeping it. Here are the most reliable options across file types.

ExifTool for Images

ExifTool by Phil Harvey is the standard for reading, writing, and transferring metadata across image formats. It supports over 400 file types and can copy metadata between files that a normal converter would treat as incompatible.

To copy all metadata from a source file to a converted file:

exiftool -TagsFromFile source.jpg converted.png

This reads every metadata tag from the JPEG and writes whatever the PNG format can accept. For batch processing across a directory:

exiftool -TagsFromFile %d%f.jpg -All:All %d%f.png

The -P flag preserves file modification timestamps, which is useful when your filesystem dates matter for sorting or compliance.

FFmpeg for Video and Audio

FFmpeg handles metadata transfer during media conversion through the -map_metadata flag. By default, FFmpeg copies global metadata from the first input to the output, but explicit mapping gives you more control:

ffmpeg -i input.mov -map_metadata 0 -c:v libx264 output.mp4

The 0 tells FFmpeg to copy metadata from input index 0. For stream-level metadata (per-track titles, language tags), add -map_metadata:s:0 0:s:0 to map stream metadata explicitly.

One common mistake: using FFmpeg's -an flag to strip audio also removes audio stream metadata. If you need the video metadata but not the audio track, use -map_metadata 0 alongside -an to keep global metadata intact.

LibreOffice for Documents

LibreOffice's command-line interface can convert documents to PDF while preserving standard properties:

libreoffice --headless --convert-to pdf document.docx

This carries over author, title, subject, and keywords. For archival needs, convert to PDF/A instead:

libreoffice --headless --convert-to "pdf:writer_pdf_Export:{'SelectPdfVersion':{'Value':2}}" document.docx

The SelectPdfVersion value of 2 produces PDF/A-2b, which requires XMP metadata and is the Library of Congress recommended format for long-term preservation.

ImageMagick for Batch Image Processing

ImageMagick strips metadata by default during conversion. To preserve it, use the -profile flag:

convert input.jpg -profile "*" output.tiff

Or use the newer magick command with -define preserve-profile:

magick input.jpg -define png:preserve-profile=true output.png

Without these flags, ImageMagick silently drops EXIF, IPTC, and color profiles during every conversion.

Preserving metadata across a batch of files requires a consistent process, not just the right tool. Here is a workflow that works across file types.

1. Audit Your Source Metadata Before converting anything, catalog what metadata your source files contain. ExifTool can generate a CSV report across an entire directory:

exiftool -csv -r /source/folder > metadata_audit.csv

This gives you a baseline. If metadata goes missing after conversion, you know exactly what was lost and can trace it back to a specific step.

2. Extract Metadata as Sidecar Files

For conversions where the destination format cannot hold your metadata, extract it into XMP sidecar files before converting:

exiftool -o %d%f.xmp /source/folder/*.jpg

XMP sidecars are plain XML files that sit alongside your converted files. Adobe applications, Darktable, and most digital asset management systems read them automatically. This approach decouples your metadata from the file format, so you never lose data regardless of how many conversions you perform.

3. Convert With Metadata-Aware Tools

Use the tools from the previous section. Avoid browser-based converters and free online tools for files where metadata matters. These services almost universally strip metadata, sometimes for privacy reasons and sometimes because they simply do not handle it.

4. Verify the Output

After conversion, run a metadata comparison between source and destination:

exiftool -a -G source.jpg > source_meta.txt
exiftool -a -G converted.png > converted_meta.txt
diff source_meta.txt converted_meta.txt

This shows exactly which fields were preserved, which were lost, and which were transformed. Pay special attention to date fields (formats vary between standards), GPS coordinates (precision may change), and copyright notices (encoding differences can corrupt special characters).

5. Reapply Missing Metadata For any fields that did not survive the conversion, reapply them from your sidecar files or the original audit:

exiftool -TagsFromFile source.jpg -All:All converted.png

This final step closes the loop. You started with an audit, converted with the best available tool, verified what survived, and patched what did not.

Some conversion scenarios deserve extra attention because they are either unusually tricky or unusually common.

JPEG to PNG (and Back)

This is the most common metadata-destroying conversion. PNG was designed for lossless web graphics, not photography, so EXIF support was added late and inconsistently. If you must convert JPEG to PNG and need metadata preserved, your best option is to write XMP data into the PNG's iTXt chunk using ExifTool after conversion. Not all viewers will read it, but the data will be there.

Going the other direction (PNG to JPEG) is simpler. JPEG natively supports EXIF, IPTC, and XMP, so ExifTool can write whatever metadata you supply.

DOCX to PDF for Legal and Compliance

When converting Word documents to PDF for legal filing or regulatory compliance, use PDF/A rather than standard PDF. Standard PDF conversion through Microsoft Word's "Save as PDF" preserves basic document properties (author, title, creation date) but drops revision history, comments, and custom properties. PDF/A-2b or PDF/A-3b preserves standard metadata and, critically, requires the conversion tool to embed XMP metadata. This makes the resulting file self-describing, which is why courts and archives prefer it.

The GoldFynch eDiscovery blog documents cases where organizations destroyed legally relevant metadata by using "Print to PDF" instead of proper PDF export. The difference matters: Print to PDF creates a new document with a new creation date, while Export to PDF preserves the original document properties.

Video Format Conversions

Video containers (MP4, MKV, MOV, AVI) each store metadata in different internal structures. MP4 uses atoms, MKV uses EBML elements, and MOV uses a variant of the MP4 atom structure. Converting between containers with FFmpeg using -map_metadata 0 handles most global metadata, but stream-level metadata (track names, language codes, chapter markers) requires explicit stream mapping.

Extracting audio from video is particularly destructive. When you pull an MP3 from an MP4, the video-level metadata (resolution, frame rate, video codec) is irrelevant, but creation date, copyright, and custom tags often get dropped too. Use -map_metadata 0 and verify the output.

Bulk Conversion Pipelines

For organizations processing hundreds or thousands of files, manual metadata checks are not practical. Build verification into your pipeline: extract metadata before conversion, convert, extract metadata after, and flag any files where critical fields (copyright, author, date) are missing from the output. Tools like Apache Tika can extract metadata from over a thousand file formats, making it useful as a universal verification layer.

If your team stores converted files in shared workspaces, Fast.io's Metadata Views can automate post-conversion verification. Describe the metadata fields you care about in plain language, and the AI extraction layer checks each file against your schema. This replaces manual spot-checking with systematic coverage, and you can add new verification columns without reprocessing existing files.

The most reliable way to preserve metadata is to avoid unnecessary conversions in the first place. Before converting a file, ask whether the conversion is actually needed or whether the consuming application can handle the original format.

When conversion is unavoidable, these practices reduce the risk of metadata loss:

• Choose metadata-rich destination formats. TIFF over PNG for archival images. PDF/A over standard PDF for documents. FLAC over MP3 when audio quality and metadata both matter. WebP over PNG when you need web-optimized images with EXIF support.
• Use XMP as your universal metadata layer. XMP is supported across images, documents, video, and audio. It is extensible, human-readable (it is just XML), and can be stored as a sidecar file when the destination format cannot embed it. Adobe created XMP specifically to solve the cross-format metadata problem.
• Keep originals. Never delete source files after conversion until you have verified that all critical metadata survived. Storage is cheap compared to the cost of recreating lost metadata, especially for GPS coordinates, copyright assignments, and regulatory timestamps.
• Standardize your toolchain. Pick one conversion tool per file type and stick with it. Inconsistent tools produce inconsistent metadata results. Document which flags and settings your team uses so that conversions are reproducible.
• Automate verification. Even with the right tools and settings, individual files can fail silently. A script that compares source and destination metadata counts after each batch run catches problems before they compound.

For teams managing large file libraries, combining local conversion tools with cloud-based metadata extraction creates a safety net. Convert files locally with ExifTool or FFmpeg, upload to a shared workspace, and let Fast.io's Metadata Views verify that the fields you care about are present and correctly formatted. The extraction works across PDFs, images, Word documents, spreadsheets, and scanned pages, so it covers the full range of conversion outputs.