Compress JPEG Files

When to Compress JPEG

• Website images — Google recommends images under 100KB for fast page loads
• Email attachments — Most email providers limit attachments to 10-25MB total
• Social media uploads — Platforms re-compress large images, often with worse quality than you'd choose
• E-commerce products — Product photos need to load fast on mobile without sacrificing zoom quality
• Blog posts — Each unoptimized hero image can add 2-5 seconds to page load time

JPEG Quality Guide

90-100: Near-lossless. Minimal size reduction (10-20%). Use for archival or professional print.

75-89: The sweet spot for most uses. 50-70% size reduction with differences invisible to most viewers. Recommended for web, email, and social media.

50-74: Noticeable quality loss on close inspection. 70-85% size reduction. Acceptable for thumbnails and previews.

Below 50: Visible artifacts. Only use when file size is critical (under 20KB targets).

When to Compress JPEG Files

JPEG is the most common image format on the web, and also the most common reason pages load slowly. A photo straight from a DSLR or modern smartphone sits at 3–5MB. That's fine for archiving — it's a problem for everything else.

Website performance. Unoptimized JPEGs are the single biggest cause of slow page loads. A hero image should land under 200KB; content images under 100KB. A 4MB photo in a WordPress blog post burns three to four seconds of load time on a mobile connection. Google's Core Web Vitals score that against you.

Email. Most providers cap attachments at 25MB. A batch of ten vacation photos at 4MB each exceeds that before you hit send. Compressing first means the email actually arrives.

E-commerce. Amazon recommends product photos under 1MB. Oversized images increase bounce rate — shoppers don't wait. Compress and resize every product photo before upload; the conversion impact is measurable.

Social media. Instagram, LinkedIn, and Twitter re-compress any image that exceeds their thresholds — and their algorithms are aggressive. Pre-compressing to your target quality means you control what the output looks like, not the platform.

Local storage. A photo library of 10,000 images at 4MB each is 40GB. Batch compression at quality 80 cuts that to 8–12GB with no visible quality change.

How JPEG Compression Works

JPEG compression is lossy by design. Understanding how it works helps you make smarter tradeoffs between file size and visual quality.

Discrete Cosine Transform (DCT). The encoder divides your image into 8×8 pixel blocks, then converts each block from pixel values into frequency components — similar to how audio compression works. High-frequency components represent fine detail and sharp edges. Low-frequency components represent broad color areas. The quality setting controls how aggressively the high-frequency data is discarded.

Quality parameter (1–100). At quality 80–85, JPEG typically achieves 60–70% file size reduction versus quality 95, with no perceptible visual difference on screen. Below quality 60, you start seeing blocky artifacts, especially around sharp edges and text. Below quality 40, the image looks degraded to any viewer.

Chroma subsampling. Human eyes are more sensitive to brightness than color. JPEG exploits this by storing color at lower resolution than brightness. The 4:2:0 mode (used by most JPEG encoders at quality settings below 90) discards half the color data in both dimensions. You almost never notice it. It saves roughly 33% on top of DCT compression.

Progressive vs. baseline. A baseline JPEG loads top-to-bottom, one row at a time. A progressive JPEG loads in passes — first a low-resolution version of the whole image, then progressively sharper. For web use, progressive feels faster even when the total bytes are identical. Progressive encoding also typically saves 2–8% on file size.

EXIF metadata. Every camera embeds metadata in JPEGs: GPS coordinates, camera model, lens, aperture, shutter speed, timestamp. This data adds 5–50KB per image and, more importantly, leaks location data if the photo was taken on a phone. Stripping EXIF is good practice before any public-facing upload.

Re-compression risk. Every time you open a JPEG and save it at a lower quality, you add another generation of artifacts on top of the existing ones. Always compress from the original high-quality source, never from an already-compressed copy.

JPEG Compression Tools

Notes on the table:

• TinyPNG's free tier limits you to 20 files per session and 5MB per file.
• Squoosh is excellent for evaluating quality tradeoffs before committing to a setting — but you have to process files one at a time.
• ImageMagick is the right choice if you're compressing hundreds of images in a build pipeline. Pixotter handles the browser workflow.

JPEG Quality Settings Guide

The 80–85 range is the practical sweet spot for most web use. You get 60–70% file size reduction from a quality-95 source, and the visual difference is imperceptible on screen. Start there and only go lower if your file size target requires it.

Tips for Best Results

• Start at quality 80. Reduce from there only if the file size target isn't met. Don't go below 60 for anything public-facing.
• Resize before compressing. A 4000×3000 image scaled to 1200×900 and then compressed at quality 80 will be dramatically smaller than the original compressed at quality 80 at full resolution. Both steps matter. Pixotter's resize tool handles both in one pass.
• Strip EXIF metadata. Cuts 5–50KB per image. Removes GPS data from photos taken on phones. Always strip before uploading user-generated content to a public page.
• Use progressive encoding for web. Better perceived load time, marginally smaller file. No downside for browser delivery.
• Batch-process e-commerce photos with identical settings. Consistent quality settings produce visual consistency across a product catalog. Mixed settings look unprofessional.
• Compress from the original, every time. Never re-compress an already-compressed JPEG. Start from the RAW or the highest-quality export you have.
• Check the result at 100% zoom. Zoom to 100% after compressing. Blocky artifacts around text or sharp edges at 100% zoom are a sign you've gone too far.