"Compress without losing quality" is technically a contradiction — compression removes data, and removed data is quality you no longer have. But it's not as absolute as that sounds. The human visual system doesn't notice all losses equally. A lot of the data that JPEG discards is in high-frequency detail that your eyes barely register in the first place.
In practice, "compress without losing quality" means: compress to the point where the loss is invisible at normal viewing size. For most document and form submissions, that threshold is generous — you can cut file size by 80–90% and the result is indistinguishable to a human eye.
How JPEG compression works (the short version)
JPEG splits the image into 8×8 pixel blocks and stores each block as a mathematical approximation rather than exact pixel values. The quality setting controls how closely the approximation has to match the original. At quality 100, it's nearly exact. At quality 50, the approximation is rougher but much smaller.
The visible artifact from heavy JPEG compression is a blocky, mosaic-like pattern — the "compression artifact" look. This becomes obvious below about quality 50. Between quality 70 and 90, most people looking at a typical document or portrait photo can't tell the difference from the original.
Tip
For document photos and ID images: quality 75–80 is usually the sweet spot. The file is 5–10x smaller than the original, and the text remains sharp and readable. Quality 60–70 is fine for preview images and thumbnails, but may show artifacts on text edges at full zoom.
PNG is different — and usually the wrong choice for photos
PNG uses lossless compression — it's mathematically reversible, meaning no data is discarded. For images with sharp edges and flat colours (logos, screenshots, icons), this is great. For photographs, PNG is almost always the wrong format because the file size advantage of lossless compression disappears when the image has millions of slightly different colours.
A photograph saved as PNG will typically be 3–5x larger than the same photo saved as JPEG at quality 80. If someone sends you a large PNG photo, converting it to JPEG at quality 80 is the easiest compression win available.
What "reduce image size" actually means
There are two ways to make an image file smaller, and they're often conflated:
- Reduce pixel dimensions — make the image physically smaller (fewer pixels). A 4000×3000 image scaled to 1600×1200 is a quarter of the pixel count and a fraction of the file size.
- Increase compression — keep the same pixel dimensions but discard more data per pixel. JPEG quality reduction does this.
For email attachments and form uploads, both approaches work. For images that need to remain at a specific size (passport photos, ID images, document submissions), only the second approach is valid — you can't reduce pixel dimensions if the portal specifies a minimum.
Practical targets by use case
- Email attachment: aim for under 1 MB. Most email clients handle this comfortably and the image remains clearly viewable.
- Form or portal upload with a 2 MB limit: quality 75–80 on a smartphone photo usually lands between 300 KB and 800 KB.
- Passport or ID photo: don't compress aggressively — these need to remain sharp. Quality 85–90 is a safe floor.
- Profile photos and thumbnails: quality 70 is fine. At small display sizes, the compression artifacts are invisible.
The Compress Image tool gives you a quality slider with three preset zones: Light, Balanced, and Aggressive. Balanced (around quality 75) is the right starting point for most submissions. If the file is still too large after that, step down to Aggressive and check the result.
One thing that doesn't work
Saving a JPEG, compressing it, then compressing the result again — this is "generation loss" and it compounds. Each round of JPEG compression adds more artifacts on top of the previous ones. If you need to compress, do it once from the original. If the original has already been compressed heavily, you're limited in how much further you can push it without visible degradation.