JPEG Quality Settings Explained (1 to 100)
Every JPEG export tool asks you to pick a “quality” level — usually a number from 0 to 100, or sometimes a slider. Higher = better quality but bigger files. Lower = smaller files but visible compression artifacts. Most people pick somewhere in the middle and hope for the best. Here’s what each level actually means.
What the quality slider actually controls
JPEG quality determines how aggressively the compressor discards image data. Specifically, it sets the quantization tables — the divisors that decide how much detail to throw away in each 8×8 pixel block.
- Higher quality = smaller divisors = more detail preserved = bigger files
- Lower quality = bigger divisors = more detail discarded = smaller files but visible artifacts
The numbering isn’t linear — going from 100 to 90 barely changes file size, but going from 30 to 20 changes it dramatically. This is intentional: the high-end of the scale is for fine detail you’d only notice on close inspection; the low end is where compression visibly affects the image.
Quality level guide
Here’s what you can expect at each level for a typical photograph:
100 (lossless-feeling, but still lossy)
Files about 20-30% smaller than the same image saved without compression. No visible quality loss to any normal viewer. Use only when storage is unconstrained or when you need to re-encode multiple times (each re-encode at lower quality loses more).
90-95
Near-perfect quality. Files about 50-60% smaller than quality 100. Indistinguishable from uncompressed by all but extreme zoom in. Recommended for archival or master copies.
85 (recommended default)
The sweet spot. Files about 65-75% smaller than quality 100. Quality is excellent for general use. Visible only under very careful inspection. This is the right default for almost everything.
80
Slightly more compressed. Files about 75-80% smaller than 100. Quality still very good. Subtle artifacts may appear in smooth gradients (sky, skin) on very close inspection. Acceptable for web and most casual sharing.
70-75
Noticeable savings, still acceptable. Files 80-85% smaller than 100. Compression starts being visible in detail-heavy areas (texture in foliage, fine patterns, smooth color transitions). Fine for thumbnails or contexts where small file size matters.
60-65
Visible compression artifacts. Files 85-90% smaller. “JPEG blockiness” — the characteristic 8×8 pixel grid pattern — becomes visible in smooth areas. Subtle “ringing” around hard edges (text, signs, contrast boundaries). Use only for small thumbnails or when bandwidth is critical.
50
Clearly compressed. Files about 90% smaller. Artifacts are obvious to most viewers. Only acceptable when image is small (thumbnails under 200×200) or when speed/size is more important than quality.
30-40
Heavy compression. Most photographic detail starts disappearing. Smooth gradients turn into blocky bands. Edges show clear ringing. Only useful for placeholder thumbnails or when extreme compression is required.
Below 20
Severe degradation. The image still resembles the source but compression artifacts dominate the visual experience. Avoid unless you specifically want the “compressed” aesthetic (some artistic uses).
What artifacts to watch for
Compression artifacts appear in specific patterns:
Blockiness: 8×8 pixel squares become visible, particularly in smooth gradient areas. Sky, skin, blurred backgrounds show this first.
Ringing (Gibbs effect): wavy lines or “halos” around high-contrast edges. Letters in text often show this — slight ghost outlines.
Posterization: smooth gradients break into visible bands of color instead of gradual transitions.
Color shifts: chroma data is compressed more aggressively than brightness data. Subtle color shifts can appear in flat-color areas.
Detail loss: fine texture (fabric weave, foliage detail, hair) gets smoothed away.
At quality 85, these artifacts are imperceptible except under careful examination. At 50, they become clearly visible. At 20, they dominate.
Why JPEG re-encoding compounds artifacts
Each time you save a JPEG, you re-encode it — which means another round of compression artifacts. So:
- Original photo: no JPEG artifacts
- Saved as JPEG at quality 85: minor artifacts (invisible)
- Re-edited and saved again at 85: artifacts from first save + new artifacts
- Re-edited and saved again: cumulative degradation
- After 5-10 re-saves: visible quality loss even though each individual save was “high quality”
This is called generational loss. To avoid it:
- Edit in a non-destructive format (PNG, TIFF) and only save to JPEG at the very end
- Or maintain the original RAW/source as your “master” and re-export to JPEG when changes are needed
- Don’t repeatedly open-and-save JPEGs
For an image you’ll edit many times, don’t keep saving as JPEG. Use PNG (lossless) for working copies; convert to JPEG only for final output.
Quality vs file size relationship
For a typical 4032×3024 (12 MP) iPhone photo:
| Quality | Approximate file size |
|---|---|
| 100 | 8-12 MB |
| 95 | 5-7 MB |
| 90 | 3.5-5 MB |
| 85 | 2.5-3.5 MB |
| 80 | 2-3 MB |
| 75 | 1.5-2.2 MB |
| 70 | 1.2-1.8 MB |
| 60 | 800 KB - 1.2 MB |
| 50 | 500-800 KB |
| 40 | 350-500 KB |
| 30 | 250-350 KB |
| 20 | 150-250 KB |
Varies based on image content (heavily textured photos compress less efficiently).
When higher quality is worth it
- Print at large sizes: at 11×14 inches or bigger, the compression of quality 85 starts becoming visible. Use quality 95+ for large prints.
- Photo editing source material: if you’ll process the image in editing software, start with high quality to avoid magnifying existing artifacts.
- Color-critical work: for art reproduction, color matching, scientific imagery, quality 95+ avoids subtle color shifts.
- Master copies for archival: family photos you want to keep forever should be high quality.
When lower quality is fine
- Web display at small sizes: thumbnails, previews, anything displayed at 800×600 or smaller can use quality 70-80 with no visible loss.
- Social media uploads: most platforms re-compress everything anyway, so saving at 85 vs 95 makes little difference in the final post.
- Bandwidth-critical contexts: emails to slow connections, mobile-data-conscious users.
- Anything users will glance at quickly: news article inline images, blog headers.
How our Image Compressor maps quality
The Image Compressor uses a 0-1 quality scale (some tools use 0-100; same idea, different scale):
| Setting | Equivalent in 0-100 |
|---|---|
| 0.95 | 95 |
| 0.85 (default) | 85 |
| 0.75 | 75 |
| 0.65 | 65 |
| 0.5 | 50 |
Default 0.85 is the sweet spot — minimal visible loss, substantial file size reduction.
PNG and other lossless formats
For images where any loss is unacceptable (logos, screenshots of text, charts):
- PNG is lossless — no quality slider, just compresses pixel-perfectly
- WebP lossless mode — lossless with better compression than PNG
- TIFF — also lossless, often used in professional contexts
Convert JPGs to PNG with JPG to PNG if you need lossless from a starting JPG (note: this can’t recover detail that the original JPG compression already lost — it just stops further degradation).
TL;DR
- Quality slider = how aggressively JPEG discards detail
- 100: near-uncompressed, big files
- 95: archival quality
- 85: the right default for almost everything
- 75-80: good for web, casual sharing
- 60-65: visible artifacts; use only when file size is critical
- Below 50: noticeably degraded; thumbnails only
- Re-encoding compounds artifacts: don’t repeatedly save as JPEG; use PNG for working copies
- For lossless: PNG or WebP lossless; JPG to PNG or JPG to WebP
- Our default: Image Compressor at 0.85 = quality 85