Batch Image Processing: Uniform Size Plus Compression
Thirty product shots, one platform, one width, one size cap. Opening an editor thirty times is history. Resize in batch, compress in batch, convert formats in batch — one pipeline, three steps.
Unify size before anything else
Resizing has two modes: by percentage (everything to 50%) and by target width (everything to 800px wide, height follows proportionally). Store listings and article images want the fixed-width mode — the display area is fixed, and uniform width is what makes a grid look tidy. Aspect ratio is always preserved; nothing stretches.
Order matters: resize first, compress second. Resizing is the biggest size lever — halving the width cuts the pixel count to a quarter, which dramatically lowers the compression burden. Compressing first and shrinking second compresses twice and pays the quality tax twice.
The batch-compression parameters
Quality 0.82 is the universal sweet spot for photographs: visually lossless at roughly half the original size. Below 0.6, dark areas start showing blocking; for product shots prefer 0.85 over saving a few hundred kilobytes. Screenshots, charts, and flat graphics tolerate 0.7 easily.
Format choice follows content: JPEG for photos (smallest), PNG for anything with transparency such as logos, WebP for marketing imagery (another 25-35% smaller at equal quality, fully supported by major platforms). Note the fallback: GIF, BMP, and AVIF inputs saved as "keep original format" come out as PNG — that is a compatibility floor, not a converter.
Batching limits and acceptance checks
Batch processing fits images from one batch with one parameter set: one shoot, one export. Mixed sources (phone photos plus screenshots) should be split into groups — the optimal quality for photos and screenshots differs a lot.
Acceptance is three spot checks: open a few at random to confirm sharpness held, verify every width matches, and confirm the total size meets the limit. Local batch processing means no queue and no upload — dozens of images resolve in under a minute.