Image Format Converter

Source image:

Drop, paste, or browse one PNG, JPEG, WebP, GIF, SVG, AVIF, BMP, ICO, HEIC, or HEIF file.

{{ sourceDropTitle }} {{ sourceDropLine }}

Target format:

The browser reports when a native encoder falls back to a different MIME type.

Image sizing:

Sizing changes happen before encoding and are reflected in every result artifact.

Output frame:

Enter width and height in pixels.

× px

Exact-canvas fit:

Scale:

{{ normalizedScalePercent }}%

Allow upscale:

Turn this option on when allow upscale is required.

Encoding quality:

{{ normalizedQuality }}%

Transparency:

Matte color:

Icon payloads:

Resampling quality:

This setting matters only when decoded pixels are resized.

Preparing the current output preview.

Measure	Value	Review note	Copy
{{ row.label }}	{{ row.value }}	{{ row.note }}

Format guidance

{{ item.title }}{{ item.status }}

The destination should decide the image format. A transparent logo, a photographic upload, a favicon, and a legacy desktop asset may show the same subject but need different combinations of transparency, compression, dimensions, and reader support.

Raster conversion rebuilds an image from decoded pixels. The source file is opened, its pixels are drawn onto a new canvas, and that canvas is encoded into another format. This creates a genuinely new file. Renaming an extension does not perform those steps, and it does not change the data inside the file.

Common image format strengths and tradeoffs
Format	Useful for	Main tradeoff
PNG	Interface graphics, screenshots, and transparency	Lossless storage can be larger than modern photographic formats.
JPEG	Photographs and broad compatibility	Lossy compression and no alpha transparency.
WebP or AVIF	Compact web delivery with modern codec support	The receiving software must support the chosen encoder and format.
BMP	Older Windows-oriented workflows	Uncompressed output can be substantially larger.
ICO	Windows icons and favicons with several embedded sizes	Some older readers expect BMP-style payloads rather than PNG payloads.

Dimensions and file size describe different things. Width and height set the pixel grid, while the encoder and quality setting decide how that grid is stored. Cutting both dimensions in half leaves one quarter as many pixels, but the encoded byte size will not necessarily fall by exactly 75 percent because image detail and codec overhead also matter.

Transparency deserves an explicit decision. PNG, WebP, AVIF, and ICO can carry alpha in the supported output path. JPEG and BMP cannot, so transparent pixels need a matte color. The same flattening occurs when a user deliberately chooses it for an alpha-capable target.

Re-encoding is not archival preservation. Metadata, GPS coordinates, color profiles, animation timing, extra animation frames, depth maps, bursts, and companion files are not intentionally retained. Keep the original when any of those details may matter later.

How to Use This Tool:

Choose the output requirement first, then adjust geometry and compression while watching the actual encoded result.

Drop, paste, or browse for one supported source image. Standard images must be below 40 MiB; HEIC and HEIF sources must be below 80 MiB.
Select the Target format that matches the destination's transparency and compatibility needs.
Choose original size, a maximum box, an exact canvas, or percentage scaling under Image sizing. For an exact canvas, select contain, cover, or stretch and decide whether enlargement is allowed.
Set Encoding quality for JPEG, WebP, or AVIF, and choose how transparent pixels should be handled. ICO output also needs a payload-size preset.
Inspect the preview, actual output format, dimensions, transparency status, and byte-size comparison before downloading.

Interpreting Results:

Use the Actual output label rather than assuming that every browser encoded the requested format. If native encoding is unavailable, a browser may return another MIME type and the downloaded extension follows the actual result.

Open the preview at useful zoom and check edges, text, gradients, and fine detail after resizing or lossy encoding.
A smaller file is not automatically better if it introduces visible artifacts, removes transparency, or loses required metadata.
Cover means source edges may be cropped; stretch changes proportions; contain can leave matte or transparent padding.
Animation inputs produce one browser-decoded still frame, not a converted animation.

Technical Details:

The conversion operates on one decoded raster frame. Geometry is planned first, pixels are centered and drawn to the planned rectangle, and encoding happens last. This ordering means every format receives the same selected dimensions and fit behavior before codec-specific compression is applied.

Transformation Core

The observable transformation is source file to decoded pixels to resized canvas to encoded output. A matte is painted before drawing whenever the target lacks alpha or flattening is selected.

Image conversion stages and effects
Stage	Operation	Material effect
Decode	Open one source image as browser-readable pixels	Animated and multi-image sources reduce to one still frame.
Plan	Resolve canvas and draw dimensions	Contain, cover, stretch, scale, and upscale policy determine geometry.
Draw	Center and resample the decoded frame	Cover clips overflow; contain can leave padding; stretch changes aspect ratio.
Encode	Write PNG, JPEG, WebP, AVIF, BMP, or ICO	Codec support, quality, alpha, and ICO payload selection determine the file.

Formula Core

For an exact canvas, contain uses the smaller axis ratio and cover uses the larger one. Without permission to upscale, any requested scale above 1 is reduced to 1.

s_{contain} = \min (\frac{W_{target}}{W_{source}}, \frac{H_{target}}{H_{source}})

s_{cover} = \max (\frac{W_{target}}{W_{source}}, \frac{H_{target}}{H_{source}})

Draw dimensions are rounded to whole pixels after scaling.

W_{draw} = round (W_{source} \times s), H_{draw} = round (H_{source} \times s)

The byte comparison uses encoded file sizes, not pixel counts.

Δ size % = \frac{B_{output} - B_{source}}{B_{source}} \times 100

Geometry Rules and Bounds

Image sizing rules and validation bounds
Setting	Rule	Limit or consequence
Percentage	Scale both axes by 1% to 400%	Values above 100% enlarge only when upscale is allowed.
Maximum box	Use the smaller width or height ratio	Aspect ratio is preserved.
Exact contain	Fit the whole image inside the canvas	Padding may remain.
Exact cover	Fill the canvas with the larger axis ratio	Overflowing edges are cropped.
Exact stretch	Scale width and height independently	Aspect ratio may change.
Canvas	Whole-pixel dimensions	Maximum 12,000 px per side and 64 megapixels.
Lossy quality	40 to 100	Used for JPEG, WebP, and AVIF; actual codec behavior remains browser-dependent.

Privacy and Accuracy Notes:

Source decoding, resizing, encoding, preview, and download occur in the current browser tab; the image is not uploaded for conversion. Local processing does not make the result archival or identical across browsers.

Browser codec availability can change the actual output type, especially for newer formats.
Re-encoding does not intentionally preserve EXIF, GPS, profiles, animation, depth data, or companion files.
HEIC and HEIF collections, bursts, depth maps, and Live Photo sidecars are outside the still-image result.
Keep the original and test the downloaded file in the destination application before deleting any source asset.

References:

HTML Standard: Canvas, WHATWG.
ICO Format Overview, Microsoft Learn.
High Efficiency Image File Format, HEIC and HEIX brands, Library of Congress.