JPEG Quality 80 vs 90: What's the Real Difference?

Updated: July 10, 2026

For most web photographs, JPEG quality 80 is a practical starting point: it produces smaller files while looking acceptable to the majority of viewers at normal screen sizes. JPEG quality 90 retains more image information and is usually larger—but the difference in both file size and visual appearance is not governed by a fixed rule. The actual outcome depends on image content, pixel dimensions, the encoder you are using, chroma subsampling settings, and whether metadata is embedded. Before settling on either number, export both from the same tool and compare them at your actual display size. Pick the smaller file where you cannot see a meaningful difference.

Key Takeaways

• Quality 80 is a starting point, not a universal standard. It works well for most web photography, but visual inspection matters more than the number.
• Quality 90 retains more detail but has no guaranteed size premium. File size depends on image content, encoder, and settings—not a fixed 1.5× to 2× ratio.
• The quality number is not portable. Different encoders, and different export dialogs within the same tool, produce different results from the same quality value.
• For screenshots and UI, test lossless formats first. PNG or lossless WebP often outperforms any JPEG setting for content with hard edges and text.
• Keep a lossless master for re-editing. Repeated JPEG re-saves compound artifacts through generation loss. JPEG at any quality level—including 100—is a lossy format.
• Modern formats often beat JPEG. WebP and AVIF frequently deliver smaller files at comparable visual quality; whether you need a fallback depends on your target audience's browser support.

How the JPEG Quality Number Actually Works

The quality value (typically 1–100) is not part of the official JPEG standard. It is a convention popularized by the Independent JPEG Group's libjpeg reference encoder and later adopted by other tools. It controls the quantization tables—how aggressively the encoder discards high-frequency detail captured during the discrete cosine transform step.

A higher number uses finer quantization and discards less information; a lower number uses coarser quantization and discards more. The relationship between the quality value and both file size and perceived sharpness is curved rather than linear. Points near the top of the scale preserve diminishing amounts of additional detail at a growing byte cost, while points near the bottom have a much larger perceptual impact.

One important caveat: the number is not portable across encoders. Photoshop has two different export paths—Save As and Export As—that use different quality scales. libjpeg-turbo, MozJPEG, and ImageMagick each map the same number to their own quantization tables, producing similar but not identical results. Treat any quality value as a starting range for a specific tool, not as an absolute standard.

JPEG Quality 80 vs 90: File Size Depends on the Image and Encoder

File size is the most practically significant difference between quality 80 and quality 90—but the gap is not predictable without testing your specific image in your specific tool. Quality 90 consistently retains more information than quality 80 and is generally larger, but there is no fixed ratio that applies across image content, encoders, pixel dimensions, chroma subsampling choices, or metadata handling.

The table below describes what you can expect qualitatively rather than prescribing numbers:

Dimension	Quality 80	Quality 90	What to do
File size	Usually smaller	Usually larger; gap varies	Export both and measure actual bytes
Typical photography	Acceptable starting point for most web use	Retains more detail; useful for critical content	Compare at your actual display size
Text and UI content	May show edge artifacts (ringing)	May reduce artifacts but JPEG is still lossy	Test PNG or lossless WebP alongside
Re-editing or archiving	Not suitable as a master	Still not suitable as a master (lossy)	Retain original RAW, PSD, TIFF, or PNG

Variables that drive the size difference more than the quality slider itself:

• Image content. Smooth areas (sky, gradients) compress efficiently at both settings; high-frequency detail (foliage, fabric, crowd scenes) results in larger files at any quality.
• Pixel dimensions. Resizing to the display width before exporting usually saves more bytes than any quality adjustment. A full-resolution image at quality 80 can easily be larger than a correctly sized image at quality 90.
• Encoder and version. The same quality value produces different quantization in libjpeg, MozJPEG, Squoosh, Photoshop Export As, and ImageMagick.
• Chroma subsampling. Many encoders use chroma downsampling for photographs, but the default varies by tool and quality level. ImageMagick's documentation notes that it switches subsampling behavior at quality 90. For images with text or sharp color edges, check which subsampling your encoder applies.
• Metadata. Embedded EXIF, IPTC, and ICC profiles can add kilobytes independently of the quality setting.

Where You May See a Visual Difference

At quality 80, JPEG artifacts can appear but are often subtle on ordinary photographs viewed at screen size. At quality 90, the same artifacts are usually less visible. The perceptual gap between the two settings is typically smaller than the file-size gap—but neither statement holds universally.

Areas where artifact differences between quality 80 and 90 may be more noticeable:

• Smooth gradients (skies, sunsets, background fills): banding or color noise may appear earlier.
• Text and UI elements: ringing along letter edges can make text look soft or dirty. Quality 80 may be insufficient here, though PNG or lossless WebP is often a better choice than raising JPEG quality.
• Saturated reds and blues: color bleeding and blockiness can appear in highly saturated regions.
• Fine repetitive detail (foliage, hair, woven fabric): detail may flatten at lower quality.
• Skin tones in close-up portraits: compression blockiness can produce an artificial appearance.

These patterns are commonly observed but are not guaranteed. The extent of visible artifacts depends on image content, the specific encoder, chroma subsampling, display scale, and viewing distance. The only reliable check is a side-by-side comparison of actual exported files at the size and context where they will be displayed.

Which JPEG Quality Should You Choose?

Use this as a starting range, not a fixed prescription. Adjust based on what you see in your actual output:

Situation	Suggested starting range	Notes
Web photography (blog, hero images)	78–82	Compare at display size; check page weight
Thumbnails and previews	65–75	Small display size hides most artifacts
Screenshots and UI with text	Test lossless first (PNG, WebP lossless)	JPEG is lossy; higher quality may not solve ringing
E-commerce product detail	82–88 as a starting point	Buyers may zoom in; visual check is essential
Print-ready delivery copy	90–95 as a starting point	Still lossy; not a substitute for the original
Master for re-editing	Original RAW / PSD / TIFF / PNG	JPEG at any quality degrades through re-saves

A practical decision path:

1. Cannot see a difference between Q80 and Q90 at your actual display size → use Q80.
2. Difference only visible at 100% zoom, not at the display size → use Q80.
3. Can see Q80 artifacts at actual display size → test Q85 or Q90 and re-compare.
4. Content is a screenshot, diagram, or contains text → test PNG or lossless WebP.
5. File will be re-edited or archived → keep the original lossless source; export JPEG only for delivery.

How JPEG File Size Can Affect LCP

Images are often among the heavier resources on a web page, and for pages where the Largest Contentful Paint (LCP) element is an image, file size can affect how quickly it loads. If the JPEG is the LCP resource and network transfer is the bottleneck, reducing file size may shorten resource load duration.

However, LCP is determined by several factors including resource discovery, fetch priority, server response time, render-blocking resources, and rendering delay—not file size alone. Switching from quality 90 to quality 80 may or may not produce a measurable LCP improvement on a given page and network condition. Use PageSpeed Insights or Lighthouse to identify the actual LCP element and the specific diagnostics for your page, then measure with real data before drawing conclusions.

Similarly, Core Web Vitals are one signal among many in Google's ranking systems. Good scores improve user experience and can contribute to better rankings, but no single change—including compressing an image—guarantees a ranking outcome.

Should You Use JPEG, WebP, or AVIF Instead?

If your project can serve modern image formats, WebP and AVIF are worth evaluating alongside JPEG. According to Google's WebP research, WebP achieves roughly 25–34% smaller file sizes than JPEG at equivalent SSIM on the test datasets used—results that reflect specific conditions and may not generalize to every image. AVIF results vary by encoder, content, and settings; independent evaluations using AOM and browser-based encoders generally show meaningful file size advantages over JPEG, particularly for gradients and complex detail, but exact results require your own tests on your own content.

Whether you need a JPEG fallback depends on your audience's browser support. The <picture> element lets you serve AVIF or WebP to supporting browsers while falling back to JPEG for environments that require it. If you do need a JPEG fallback, quality 80 remains a reasonable starting point for that version.

How to Test Quality 80 and 90 on Your Own Image

Quality comparisons only make sense on your actual image, in your actual tool, at your actual display size. Here is a repeatable process:

1. Start from the same source file. Use your lossless original—RAW, TIFF, or PNG—not a previously saved JPEG.
2. Fix all other variables. Use the same encoder, the same pixel dimensions, the same chroma subsampling, and the same metadata handling for both exports.
3. Export both versions at quality 80 and quality 90 using your target tool.
4. Check the actual file sizes in bytes—not the quality slider readout.
5. Compare at your actual display size, not zoomed in to 100%. If you cannot see a difference at the size where the image will appear, quality 80 is sufficient.
6. Zoom in on problem areas if needed: gradients, text, saturated colors, fine detail.
7. Decide by the smallest quality where you stop seeing meaningful artifacts for that specific image in that specific context.

For a private, browser-based comparison without uploading files to a server, LessMB's JPEG compressor lets you select a JPEG, adjust the quality slider, and compare the output size and visual result—all processed locally in the browser.

Variables That Matter More Than the Quality Number

These factors often have a larger effect on file size and output quality than the difference between 80 and 90:

• Resizing before export. A 5000 px image at quality 80 is far larger than the same image resized to 1200 px at quality 90. Resize to display dimensions first.
• Chroma subsampling. The encoder's subsampling mode can meaningfully affect color rendering in text and saturated content, independently of the quality value.
• Re-saving from a JPEG source. Each encode/decode cycle compounds artifacts through generation loss. Edit a lossless master; export JPEG once.
• Encoder choice. The quality scale is not standardized. Calibrate your settings to the encoder you are actually using, and test the real output.
• Metadata. EXIF, IPTC, and color profiles add file weight. Strip or retain them intentionally.
• Visual inspection. A number on a slider tells you nothing on its own. Compare actual exported files at the actual display size.

FAQ

Is JPEG quality 80 good enough for websites?

For most web photography, quality 80 is a reasonable starting point. It keeps files smaller while looking acceptable to most viewers at normal screen sizes. If you can see artifacts in the actual display context, test 85 or 90 and compare the output visually.

How much bigger is a JPEG at quality 90 than quality 80?

There is no fixed ratio. The size difference depends on image content, pixel dimensions, the encoder used, chroma subsampling, and whether metadata is retained. Quality 90 retains more information and is usually larger, but the gap varies widely across images and tools.

Is JPEG quality 100 lossless?

No. JPEG at quality 100 is still a lossy format. It discards some information and the file is typically much larger than quality 90, often without visible improvement. Use PNG, TIFF, or your original RAW file for a true lossless master.

Why can two quality-80 JPEGs have different sizes?

The quality number is a convention, not part of the JPEG standard. Different encoders—Photoshop's Export As, libjpeg, MozJPEG, ImageMagick, Squoosh—map the same number to different quantization tables. Image content, pixel dimensions, chroma subsampling, and embedded metadata also affect file size independently of the quality slider.

Should screenshots use JPEG quality 90 or PNG/WebP?

Screenshots and UI images with hard edges and text often show JPEG ringing and color bleeding even at quality 90. Test PNG or lossless WebP alongside any JPEG setting. For content that must stay JPEG, 85–92 is a common starting range, but visual inspection on the actual output is essential.

Does JPEG quality affect SEO?

Indirectly. If a JPEG is the LCP element and network transfer is the bottleneck, a smaller file may shorten resource load duration and potentially help Largest Contentful Paint. However, LCP also depends on resource discovery, prioritization, server response, and rendering delay. Measure with real field or lab data rather than assuming file size alone determines LCP or rankings. Core Web Vitals are one signal among many in Google's ranking systems.

Should I re-save a JPEG multiple times?

Avoid it. Each re-encode compounds compression artifacts through generation loss. Keep a lossless master such as PNG, TIFF, or the original RAW file, and export a fresh JPEG when you need one.

References

1. MDN Web Docs — Image file type and format guide
2. web.dev — Learn Images
3. web.dev — Largest Contentful Paint (LCP)
4. web.dev — Optimize LCP
5. web.dev — Core Web Vitals
6. Google — WebP Compression Study
7. Google — PageSpeed Insights
8. Independent JPEG Group — libjpeg reference encoder
9. ImageMagick — Quality option documentation
10. Adobe — Save files in graphics formats (Photoshop)