Skip to content

P4-6: Compress/optimize images #85

Closed
#93
Closed
P4-6: Compress/optimize images#85
#93
Labels
go:needs-researchNeeds investigationmediaMedia/image/video operationsoptimizationOptimization workphase-4Phase 4: Presentation OptimizationsquadSquad triage inbox — Lead will assign to a membersquad:copilotAssigned to @copilot (Coding Agent) for autonomous worksquad:shiherlisAssigned to Shiherlis (Tester)type:featureNew capability
Milestone
Phase 4: Presentation Optimization
@jongalloway

Description

@jongalloway
jongalloway
opened on Mar 19, 2026

Description

Downscale images that are larger than their display dimensions warrant. Allow target DPI selection (e.g., 150 DPI for screen, 300 DPI for print). Support format conversion (BMP/TIFF to PNG/JPEG for better compression). Return compression statistics before/after per image.

Technical Approach

  • Query all images in the presentation via P4-2 patterns
  • For each image part: extract pixel dimensions, current size
  • Query slide shapes to determine display size (EMU to pixels)
  • Calculate optimal resolution vs. display size (DPI-based)
  • Use image processing library (likely SkiaSharp or System.Drawing) to downscale/re-encode
  • Replace the image part in the package with optimized version
  • Validate with OpenXmlValidator
  • Return compression stats: before/after per image, total savings

Architectural Decision: Image Processing Library

Choose between:

  1. SkiaSharp (cross-platform, modern, no Windows-only dependencies)
  2. System.Drawing.Common (legacy, Windows-only on .NET, but pre-installed)
  3. ImageSharp (pure .NET, slower than SkiaSharp)

Recommendation: SkiaSharp. Pro: cross-platform, high quality. Con: new NuGet dependency. Document in README and NuGet dependency tracking.

Acceptance Criteria

  • Tool accepts file path, target DPI, optional format conversion flag
  • Scans all images in presentation
  • For each image: measures display vs. pixel dimensions
  • Downscales images that exceed target DPI relative to display size
  • Optionally converts BMP/TIFF to PNG/JPEG
  • Re-encodes images with appropriate compression (JPEG quality ~85%)
  • Updates image parts in the package
  • Validates with OpenXmlValidator
  • Returns: images processed, original size, compressed size, total savings
  • Unit tests: 3+ cases (no compression needed, significant downscaling, format conversion)
  • E2E test: compress a presentation with mixed image sizes, open in PowerPoint, verify visual quality and size reduction

Tier

Tier 2 — Write Operations (optimization)

Size Estimate

Large (6–8 hours, includes dependency research and E2E validation)

Dependencies

None direct; pairs well with P4-2 for media analysis

Risk Mitigation

  • Test round-trip: compressed images must display correctly in PowerPoint
  • Document JPEG quality trade-offs (85% is recommended starting point)
  • Log before/after metadata so users can validate compression is acceptable

Metadata

Metadata

Assignees

No one assigned

    Labels

    go:needs-researchNeeds investigationmediaMedia/image/video operationsoptimizationOptimization workphase-4Phase 4: Presentation OptimizationsquadSquad triage inbox — Lead will assign to a membersquad:copilotAssigned to @copilot (Coding Agent) for autonomous worksquad:shiherlisAssigned to Shiherlis (Tester)type:featureNew capability

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions