Key takeaways
The three points that matter most
best scan settings for sheet music OCR
Use a lossless or high-quality source
Prefer scanner output or a sharp original over screenshots and repeatedly compressed images.
best scan settings for sheet music OCR
Start around 300 dpi
This is a practical baseline for printed notation; extremely low resolution removes staff and symbol detail.
best scan settings for sheet music OCR
Geometry matters
Keep staves horizontal, pages flat, margins intact, and lighting even.
Explanation 01
Resolution, color, and compression
For ordinary printed scores, around 300 dpi is a useful starting point rather than a guaranteed optimum. Higher resolution may help small print, but it also increases file size and processing cost. Preserve grayscale information when pages are faded or uneven; let the OMR pipeline choose binarization instead of crushing detail too early.
- Avoid screenshots of browser-rendered pages when the original PDF is available.
- Prefer PNG or TIFF for intermediate images; use high-quality JPEG only when necessary.
- Do not sharpen so aggressively that staff lines and noteheads merge.
- Check the actual pixels, not just the DPI metadata field.
Explanation 02
Deskew, crop, and page geometry
Staff lines are strong geometric anchors. Rotation, perspective, curvature, and shadows make it harder to separate staff lines from beams and text. Keep every full staff visible and avoid cropping clefs, key signatures, barlines, or ledger lines.
- Rotate the page until staves are horizontal.
- Photograph straight-on with even light and no fingers or binding shadow.
- Keep page order and orientation consistent in multi-page PDFs.
- Split pages with very different sizes or scan conditions before testing.
Explanation 03
Run a representative-page test
The cleanest cover page is not a useful qualification sample. Test a page containing the densest voices, smallest staff, lyrics, tuplets, or cross-staff notation you expect. If that page needs heavy correction, plan the project around human review or manual entry.
Step-by-step
Follow this order
Inspect the physical page
Flatten folds, remove shadows, and confirm you have the right to digitize the score.
Scan at a practical baseline
Start near 300 dpi in grayscale or color when tonal detail matters.
Deskew and crop safely
Straighten staves while keeping all musical context and margins needed for recognition.
Check at 100% zoom
Look for broken staff lines, blocked noteheads, compression blocks, blur, and missing edges.
Test the hardest page
Compare its candidate with the source before processing the rest of the document.
FAQ
Frequently asked questions
Is 600 dpi always better than 300 dpi?
No. It can help very small print, but also increases processing cost and cannot repair blur, shadows, or perspective.
Should I convert the page to black and white first?
Only if the result preserves fine symbols. Grayscale may retain useful information for adaptive binarization.
Can I use a phone photo?
Yes for a careful straight-on photo, but a flat scan is usually easier to recognize because it avoids perspective, curvature, and uneven light.
Verification sources
Primary and upstream sources
These links verify formats, software behavior, and public limitations; they do not imply commercial endorsement of ScoreTransposer.
Continue reading