Format guide

Image conversion help

Practical guidance for choosing formats, preserving transparency, balancing file size, and understanding browser-based processing.

Compare real conversion results

A hands-on GllOne Lab experiment, measured on your device—not a table of promised compression rates.

Run the same original 960 × 640 test card through PNG, two JPG quality settings and WebP. The card is created in this tab; it is not uploaded, and the comparison needs no personal files. Downloads happen only when you choose them.

Compare like for like. Dimensions stay at 960 × 640. Resizing, sharpening and colour adjustments are off; JPG uses a white background.
Inspect the border. The checkerboard shows through transparent pixels. JPG must replace transparency with a solid colour.
Inspect text and lines at 100%. A fitted thumbnail can hide compression damage. Open the downloaded images at actual size before choosing the smallest acceptable result.

This card tests sharp edges and gradients, not every photograph or colour profile. Results vary with browser encoders and installed fonts. It does not measure photographic fidelity, preserve editing layers, or certify archival or print quality. A higher quality setting is not a percentage of original detail retained.

No comparison has run yet. The experiment requires JavaScript; the method and guidance above remain readable without it.

PDF resolution and page order: a worked exercise

Use a known source to check what your export actually contains.

A preview can look sharp while hiding the difference between a small image and a print-sized export. This original two-page practice document makes that difference measurable: page 1 is portrait, 288 × 432 PDF points; page 2 is landscape, 432 × 288 points. Both contain numbered labels, small selectable text, colour blocks and thin lines. The unpainted page background also lets you test transparency.

Generated on your device · two pages · no personal data

Exercise 1: 150 DPI versus 300 DPI

Open the PDF-to-image tool and choose the practice PDF. Select all pages and PNG.
Set 150 DPI, rotation 0°, crop 0%, and leave appearance adjustments off. Export and open the images.
Repeat at 300 DPI with the other settings unchanged. Compare the smallest text and thin lines at 100% zoom in an image viewer.

Expected dimensions, with no crop or extra rotation
Page	150 DPI	300 DPI
1 · portrait	600 × 900	1200 × 1800
2 · landscape	900 × 600	1800 × 1200

For this document, pixels = PDF points × DPI ÷ 72. Doubling DPI doubles each side, so the 300 DPI page has four times as many pixels. It does not guarantee four times the file size: encoding and page content affect bytes. The tool’s preview is capped at 96 DPI, so judge export detail from the download.

Exercise 2: rendering is not merging

Open the source PDF and try selecting its page label. Now open an exported PNG: the label has become image pixels, not selectable text. Higher DPI improves the sampling of vector text but does not add OCR or document structure.

To keep document pages instead, open Merge PDFs. Add the practice file once, then add it again as a second item. Merge and inspect the result: it should have four pages in the order portrait 1, landscape 2, portrait 1, landscape 2. Check the page labels, not just the file’s name.

This checks page count and order for a simple document. It does not test form fields, signatures, bookmarks or accessibility tags. Keep the original whenever those features matter; page copying may not retain them.

One more check: the page background

Export PNG with transparency enabled and place it over a dark colour. The unpainted areas of this practice PDF should show through. JPG cannot store transparency and needs a solid background. A scanned PDF usually contains an opaque page image: enabling transparency will not automatically remove its white paper.

The dimensions above are calculated from this document’s known page boxes and checked in the site’s automated tests. File size and rendering details can vary by browser and library version. Generated sample text and graphics may be reused for practice; this is not a professional print or PDF conformance certification.

Format questions and honest limitations

Practical answers for choosing a workflow and checking the result.

Reference guide

Do my files upload to a conversion server?

The converter reads and processes selected files inside the browser tab. The source file is not posted to a GllOne Lab conversion server.

Normal website requests still occur: the hosting provider serves the page and the PDF tools load their engine and helper resources from this website. The AdSense verification meta tag identifies the connected publisher account; it does not load advertising. See the privacy notice for the complete distinction.

Why did PNG to JPG remove transparency?

JPG has no alpha channel. Every transparent or partly transparent pixel must be flattened onto a solid colour. Choose a background that matches the destination, then inspect fine edges for a light or dark halo.

What quality should I choose for JPG or WebP?

There is no universal percentage. Start around 88–92 for ordinary web photographs, compare the download at 100% zoom, and pay particular attention to hair, foliage, text, gradients, and high-contrast edges.

Reducing pixel dimensions to the size actually displayed often saves more data than pushing the quality slider too low.

Will camera metadata and colour profiles be preserved?

Do not rely on it. The still-image converters render pixels through a browser canvas and create a new file, which normally removes EXIF, GPS, camera thumbnails, and most embedded metadata. Colour can also vary slightly between software when profiles are interpreted differently.

What happens to an animated GIF or WebP?

These are still-image workflows. An animated source is treated as a single rendered frame; the animation timeline, frame timing, and looping information are not preserved.

Can raster-to-SVG tracing recreate the original vector file?

No. A raster image contains pixels, not the original paths, layers, text objects, or design history. Tracing estimates new shapes from the visible pixels.

Clean logos and flat icons are the best candidates. Photographs usually create a large, complex SVG that is less useful than the source image.

Does PDF-to-image perform OCR?

No. It renders the appearance of each selected page to pixels. Text becomes part of the image even if it was selectable in the PDF. Use a dedicated OCR tool when searchable or editable text is required.

Why can a large image, PDF, or batch fail?

Decoding, previewing, rendering, encoding, and ZIP packaging all use browser memory. Large pixel dimensions, high PDF DPI, and many files multiply that load. Test a smaller batch, lower the output dimensions or DPI, close other heavy tabs, and try again.

Why does a less-common format work in one browser but not another?

Browsers ship different image decoders, and one extension can contain several encoding variants. PNG, JPG, and WebP are the dependable outputs offered here. TIFF, AVIF, ICO, GIF, and BMP input remains dependent on whether the current browser can decode that particular file.

Which format should I choose?

Use the destination and the image content to decide.

PNG

Good for

Logos
Text/UI
Transparency
Screenshots

Avoid for

Large photo libraries (can be big)

Notes

Lossless
Transparency supported
Often bigger than JPG/WebP for photos

JPG

Good for

Photos
Gradients
Small-ish file sizes

Avoid for

Transparency
Hard-edge text/logos (can show artifacts)

Notes

Lossy
No transparency
Use 85–95 quality for most needs

WebP

Good for

Web performance
Photos
Transparent graphics

Avoid for

Very old software that doesn't support WebP

Notes

Modern format
Often smaller at similar quality
Can be lossless or lossy

SVG

Good for

Logos
Icons
Illustrations
Infinite scaling

Avoid for

Photographs (can become huge when traced)

Notes

Vector shapes not pixels
Raster→SVG is tracing/estimation

A four-check export routine

Use this before replacing an original or publishing an important file.

1
Open it. Confirm that the downloaded file opens in the destination browser, editor, document tool, or publishing service.
2
Inspect it. View the image at 100% and check fine edges, gradients, small text, transparency, and colour.
3
Measure it. Verify pixel dimensions, file size, page count, and filename rather than relying on the extension alone.
4
Keep the source. A converted delivery file is not a replacement for the best original, especially after lossy compression or rasterisation.

What conversion can change

Format conversion is a new export, not a neutral rename.

Pixels and dimensions

Resizing changes the pixel grid. Downscaling discards detail; upscaling creates more pixels without recovering original information.

Compression

JPG and WebP quality settings can remove visual information. Repeating lossy exports may make artefacts accumulate.

Transparency

PNG and WebP can store transparent pixels. JPG cannot, so transparent areas need a solid background.

Metadata and features

EXIF, GPS, animation, layers, editable text, PDF interactivity, and vector paths may be removed depending on the workflow.