# Is PNG a Lossless Format? What It Covers (2026) > Is PNG a lossless format? Yes, it decodes to the exact pixels it was given. What that covers, where photos still lose quality, and when JPEG or HEIF fits. - Canonical HTML: https://revelraw.com/blog/is-png-a-lossless-format.html - Markdown mirror of this page: https://revelraw.com/blog/is-png-a-lossless-format.md - Published: 2026-11-04 - Language: English - Machine-readable site index: https://revelraw.com/llms.txt · Full site as one file: https://revelraw.com/llms-full.txt Published November 4, 2026 · Daniel Tello · RevelRaw **Quick answer:** Yes. PNG compresses with DEFLATE, which is fully reversible, so a PNG decodes to exactly the pixels that were written into it, however many times you open and re-save it. What lossless doesn't cover is everything that happened before the save: dropping to 8 bits, converting to a smaller color space, resizing, reducing to a 256-color palette or starting from a JPEG all change pixels first, and PNG then keeps that result exactly. For photos you deliver, JPEG or HEIF is usually far smaller; PNG suits editing copies, transparency, screenshots and flat graphics. PNG is a lossless format: decoding returns exactly the pixel data that was encoded, and its compression method is DEFLATE. So why can a photo still lose quality when you export it as a PNG? The answer is the difference between lossless compression and a lossless workflow. Your editor can change tonal precision, colors, dimensions and detail before compression begins. PNG faithfully keeps that result, including any compromises already made. This guide covers what the format guarantees, where photos still lose quality, and when JPEG or HEIF is the better export. Edit RAW photos free on your Mac. RevelRaw · macOS 26 or later · Apple M1 or later [Try it on a RAW file](https://apps.apple.com/app/apple-store/id6773942240?pt=120914369&ct=BlogOutrank&mt=8) ## What Does Lossless Mean for a PNG? PNG stores the pixel values written into the file rather than approximating them to save space. That matters when a small change to an edge, letter, mask or transparent boundary would show. The qualification is what was written. A rendered photograph isn't the camera capture, and an exported image isn't necessarily identical to the document you were editing. Lossless and lossy modes in common image formats: | Format | Compression mode | Transparency | Bit depth | | --- | --- | --- | --- | | PNG | Lossless only, using DEFLATE | Yes, per-pixel alpha | 8 or 16 bits per channel for color photos; 1 to 8 bits for palette images | | JPEG | Lossy in ordinary photo exports | No alpha channel | 8 bits per channel in practice | | HEIF | Usually lossy; lossless where the encoder offers it | Yes, where the app supports it | 8 or 10 bits per channel are common | | WebP | Lossy or lossless | Yes | 8 bits per channel | | TIFF | Uncompressed, lossless ZIP or LZW, or lossy JPEG inside the TIFF | Yes, when the receiving app supports it | 8 or 16 bits per channel, set in the export dialog | This is a practical export comparison, not a promise that every app exposes every option. A HEIF exporter may offer only lossy output, and a TIFF preset can use JPEG compression without making that obvious. For broader terminology around raster graphics and delivery files, PostSyncer's [design file formats decoded](https://postsyncer.com/blog/graphic-design-terms) is a useful glossary. > **Practical rule:** Check the compression mode, transparency setting and bit depth in the export dialog. Don't rely on the file extension alone. If you need exact rendered pixels and transparency, PNG fits the job. If you need a compact photograph for delivery, compare photographic formats first. See what RAW recovers on one of your own photos. Pull back blown highlights and lift the shadows on a RAW you already have. Your first full-quality export is free. macOS 26 or later · Apple M1 or later [Try it on a RAW file](https://apps.apple.com/app/apple-store/id6773942240?pt=120914369&ct=BlogOutrank&mt=8) ## How Does PNG Keep Every Pixel? Think of someone copying a document with repeated sentences. Instead of writing each repeat in full, they note where it appeared earlier. Anyone reading those notes can rebuild the original word for word. PNG uses the same idea on recurring byte patterns. It never needs to blur an edge or simplify a texture to make those references work. 1. **Filter the rows.** Before compression, the encoder can describe each row's bytes as differences from neighboring bytes. These filters are reversible and often turn similar areas into more repetitive data. 2. **Replace repeated patterns with references.** DEFLATE combines LZ77-style back-references with Huffman coding and looks back over a sliding window of up to 32,768 bytes, according to the [W3C PNG compression specification](https://www.w3.org/TR/PNG-Compression.html). A reference points to an earlier pattern; it doesn't delete anything. 3. **Encode frequent symbols efficiently.** Huffman coding gives shorter codes to symbols that occur often. That changes how the information is written, not what it says. 4. **Store the compressed stream.** PNG packages the stream into chunks, each with a checksum. The checksums catch corruption; they don't add detail. 5. **Reverse the process when opening.** The decoder decompresses the data and undoes the filters. The original pixel values come back exactly. ### Why photographs compress differently A flat background is full of predictable repetition. A photograph is full of grain, noise and gradual tonal change, which is hard to describe compactly, so a RAW-derived photo stays large even at the strongest PNG compression setting. Size also depends on dimensions, bit depth, alpha, metadata, the filter choices and the encoder. To test it yourself, export a PNG, reopen it and save another PNG without editing, resizing or converting its profile. Compare the two in an editor with a difference blend mode. The pixels match, although the file bytes and metadata needn't. ## How Do JPEG and HEIF Trade Detail for Size? JPEG and HEIF exist to make photographs small enough to deliver. They describe the image approximately, keeping the tonal information the eye notices most and simplifying the rest. JPEG divides the picture into small blocks and stores each block as frequency components, encoding fine variation more coarsely than the broad shapes beneath it. For a finished file viewed at its intended size, that approximation is the right deal. A delivered portrait rarely needs every subtle variation held in the editing master. Push the quality setting too low and the simplification shows: blocks in smooth gradients, halos around lettering, rough transitions across skin. ### A portrait through repeated exports Re-encoding is where the cost lands. Opening a JPEG leaves its stored pixels untouched, but saving it again runs the approximation over an already simplified image, so the second export starts from the first one's compromises rather than the original skin texture and edges. Repeat the cycle and artifacts can build up, more or less visibly depending on the image and settings. That is why the capture decision matters upstream. The guide to [shooting in RAW versus JPEG](https://revelraw.com/blog/shooting-in-raw-vs-jpeg.md) explains why keeping an editable source protects everything downstream. HEIF delivers photographs efficiently, but a lossy HEIF export still discards information. The format can do lossless; most photo exports aren't. JPEG has no alpha channel. If you use PDFWix to [change PNG files into JPGs](https://www.pdfwix.com/png-to-jpg), transparent areas get filled with a background color, so pick one deliberately and treat the result as a delivery copy. Use JPEG or HEIF for the delivered portrait. Keep PNG, or TIFF, for a copy you still need to edit, because a lossless format only preserves what earlier exports left intact. ## Where Can a Lossless PNG Still Lose Quality? Exporting a high-precision edit at a lower bit depth rounds tonal values before PNG compression even starts. A sky that looked smooth while editing may then band when someone adds contrast later. The [W3C PNG specification](https://www.w3.org/Graphics/PNG/png-spec8.html) allows 1, 2, 4, 8 or 16 bits per sample depending on the color type; full-color (truecolor) images use 8 or 16 bits, and they can carry per-pixel alpha. The precision belongs to the setting you choose, not to the letters PNG. Common workflow losses: 1. **Reducing tonal precision.** Going from a 16-bit edit to an 8-bit file rounds values. Later contrast or color changes make that rounding easier to see. 2. **Reducing to a palette.** An indexed (palette) PNG holds at most 256 colors. Size-saving tools that do this are fine for icons and flat graphics and visibly band photographs, even though the result is still technically a PNG. 3. **Converting to a narrower color space.** Colors outside the destination gamut get mapped or clipped. PNG stores the converted result, not the colors that no longer fit. 4. **Resizing or sharpening.** Both change pixel values on purpose. Follow a controlled [workflow for resizing an image without losing quality](https://revelraw.com/blog/how-to-resize-an-image-without-losing-quality.md) instead of repeatedly resampling intermediate copies. 5. **Starting from a compressed preview.** Turning a JPEG into a PNG stops further compression loss. It doesn't bring back texture the JPEG already threw away. 6. **Clipping highlights or shadows.** Once the rendered output contains featureless white or black, no format can recover the missing detail. ### Precision and transparency do different jobs Choose 16-bit PNG when another editor will still do substantial compositing, masking or tonal work. 8-bit suits final delivery, where compatibility and size matter more than adjustment headroom. PNG's alpha is unassociated, also called non-premultiplied: color and opacity are stored separately, which allows soft, partly transparent edges instead of only fully visible or fully invisible pixels. > Lossless describes what happens after those export decisions. It doesn't certify that the decisions themselves kept everything. A higher PNG compression level changes encoding effort and file size. It can't undo rounding, gamut conversion, clipping or earlier edits. ## When Is a PNG Too Large for the Job? PNG grows large for a mechanical reason: its compression finds repeated patterns and writes them once. Sharp text, flat color and hard edges repeat constantly, so they shrink well. A photograph works against the codec. Grain, sensor noise and gradual tonal change produce few repeats, so the lossless guarantee costs real bytes exactly where the eye tolerates approximation. Content type and the usual format choice: | Content type | PNG fit | Usual delivery choice | | --- | --- | --- | | Screenshots with text | Strong, keeps text and edges sharp | PNG | | Logos and flat graphics | Strong, repeated colors compress well | PNG, or SVG from a vector original | | Photo with transparent cut-out | Good, keeps the alpha | PNG, or WebP or HEIF with alpha where supported | | Photo with smooth sky or gradient | Keeps every pixel, but large | JPEG or HEIF for delivery, 16-bit PNG or TIFF for editing | | General web photo | Usually inefficient | WebP or AVIF where supported, JPEG as fallback | The trade-off shows up across the web. The [2025 Web Almanac performance chapter](https://almanac.httparchive.org/en/2025/performance) looks at the formats sites use for their largest above-the-fold image, the one that sets Largest Contentful Paint, and PNG still appears there alongside JPEG; newer formats such as WebP and AVIF generally compress photographs better, which helps those images load sooner. File weight is a delivery decision; PNG's fidelity stays intact either way. A heavy PNG still decodes to exactly the pixels it was given. Graphics with sharp edges and flat color suit PNG; photographs with smooth gradients usually fit a lossy delivery format better. WebP and AVIF also offer lossless modes, so lossless isn't exclusive to PNG. ### Compare files for the actual destination Export the same image at the same dimensions as PNG and as a lossy format, then view both at the size the destination shows. Faces, foliage, text overlays and gradients reveal artifacts first. Compare the sizes and keep the smaller file that passes. For a LinkedIn image, FoPost's [LinkedIn image resizing tool](https://fopost.com/tools/linkedin-photo-resizer) can prepare the platform's dimensions. Size the image first, then choose the format: resizing changes pixel values, while the format only changes how those pixels are stored. > **Delivery check:** Keep the smallest file that meets the destination's visual and functional needs, rather than choosing PNG by default. Two export choices inflate PNG files before the format even matters. Saving 16 bits when the destination only shows 8 doubles the pixel data before compression, and the noisy low bits compress poorly. Keeping a wide-gamut color space that the destination converts anyway stores color the viewer never sees. A technically exact file can still be needlessly heavy for its use. ## Which Export Settings Work From RAW to PNG? Treat PNG as a derived image, not a replacement for the capture. The [W3C PNG specification](https://www.w3.org/TR/png-3/) defines lossless raster storage and ways to record color information, but none of that preserves a RAW sensor capture or an editor's non-destructive edit history. The workflow starts by protecting the master and ends by inspecting the result. 1. **Keep the RAW or DNG master untouched.** Keep its edit settings or sidecar with it. Converting a PNG back to a RAW-like filename doesn't recover the capture. 2. **Finish the creative work at full resolution.** Set white balance, tone, color and local corrections before making delivery copies, so a reduced export never becomes the starting point for later changes. 3. **Choose the bit depth for the next task.** 16-bit for further editing, when your exporter and the receiving app support it; 8-bit for final delivery when no more tonal work is expected. 4. **Match the destination's color space.** Use sRGB for the web and most clients unless the destination asks for something else, and embed the profile rather than trusting the recipient to guess. The [Display P3 vs sRGB guide](https://revelraw.com/blog/dci-p3-vs-srgb.md) covers when the wider space is worth it. 5. **Resize near the end.** Set final dimensions after the creative decisions, then apply output sharpening for that size. Resizing still changes pixels; the point is to do it once, on purpose. 6. **Save a separate derivative.** Give the export a clear name and keep it apart from the master. If you use Lightroom, the guide to [Lightroom export settings](https://revelraw.com/blog/export-settings-for-lightroom.md) walks through the same output decisions. 7. **Inspect the exported image.** Open it at 100% and check gradients, edges, transparency and color, then check it again in the app that will show it. Color deserves that last check: apps can read or ignore a PNG's color information differently, so identical pixel values don't guarantee an identical look on every screen. ## Which PNG Myths Cost Photographers Quality? These persist because lossless sounds like a guarantee for the whole photograph. A file can look unchanged on screen and still hold less editing room than its source. - **PNG is always smaller than JPEG at the same quality.** For photographs it's usually much larger. Exact preservation and a visually acceptable approximation aren't the same target. - **Higher PNG compression improves quality.** It changes encoding effort and file size. Every compression level decodes to the same pixels. - **A PNG exported from RAW keeps the RAW data.** It keeps one rendered version. The sensor data and your edit settings live only in the RAW and its sidecar or catalog. - **Lossless means the image can't change.** Editing, resizing, profile conversion and bit-depth reduction all change it. Lossless compression preserves the result of those steps. - **PNG is the best format for every photo.** It's right for exact graphics and transparent composites. A delivered photograph is usually better served by a smaller format the recipient supports. Ask what the next person needs: an editable master, a transparent asset, an exact interchange file or a lightweight viewing copy. Those are different deliverables, and the format should follow from which one it is. ## How Do You Export a PNG From RevelRaw? We make RevelRaw, a native RAW editor for Macs with macOS 26 or later on Apple M1 or later. It exports PNG, JPEG, HEIF and 8- or 16-bit TIFF straight from the RAW, in sRGB, Display P3 or Adobe RGB. When you want a 16-bit master that another app will keep editing, 16-bit TIFF is the export to choose. A Location option removes GPS from exports, and it is on by default for the Social and Web presets. 1. **Open the RAW and finish the edit.** AI scene detection ranks 50+ looks for the photo as a starting point; tone, color and up to three mask layers are free. 2. **Choose the format for the job.** PNG for an exact copy or a graphic workflow, JPEG or HEIF for delivery, 16-bit TIFF for a master another app will keep editing. 3. **Pick the color space and size.** sRGB for the web, Display P3 or Adobe RGB only when the destination handles it. 4. **Inspect the file.** Check gradients and edges at 100% before you send it. Editing is free and the first full-quality export is included, with no account. Pro adds unlimited and batch export, as an auto-renewing subscription or a one-time lifetime purchase. Processing happens on your Mac; downloading the app, buying Pro and the one-time Apple selection-model download for click-to-select masks need internet. If you have no RAW to practice on, there are 14 [free CC0 sample RAW files](https://revelraw.com/sample-raw-files.md) across 12 formats. [Get RevelRaw on the Mac App Store](https://apps.apple.com/app/apple-store/id6773942240?pt=120914369&ct=BlogOutrank&mt=8). ### Check the profile, not just the pixels When a delivered PNG looks duller or shifted on another screen, the compression is rarely to blame. Open the export in a viewer that reads embedded profiles and confirm the color space matches what you chose. PNG stores the pixels faithfully; how other software shows their color depends on the profile traveling with them. Keep the RAW master whatever you export. A PNG made from a JPEG keeps the JPEG's artifacts exactly as faithfully as a PNG made from RAW keeps its gradients, because lossless preserves whatever it receives. That is why the format decision belongs before the first save, not after it. See what RAW recovers on one of your own photos. Pull back blown highlights and lift the shadows on a RAW you already have. Your first full-quality export is free. macOS 26 or later · Apple M1 or later [Try it on a RAW file](https://apps.apple.com/app/apple-store/id6773942240?pt=120914369&ct=BlogOutrank&mt=8) ## Frequently Asked Questions ### Is PNG a lossless format? Yes. PNG compresses image data with DEFLATE, which is fully reversible, so a PNG decodes to exactly the pixel values that were written into it. Lossless describes the compression only: anything that changed the pixels before saving, such as dropping to 8 bits, converting the color space, resizing or reducing to a 256-color palette, is stored faithfully too. ### Is PNG or JPEG better for photos? For a finished photo you deliver or post, JPEG is usually better: it is far smaller and the difference is hard to see at a sensible quality setting. PNG is the better choice for an editing copy you will adjust again, for images with transparency, and for screenshots, text and flat graphics where JPEG artifacts show. ### Does a PNG lose quality if you save it repeatedly? No. Opening and re-saving a PNG without editing, resizing or converting it gives back the same pixels every time, even though the file size and metadata can change. A JPEG re-saved again and again can build up artifacts, which is why PNG or TIFF suits copies that will be edited more than once. ### Should I export 8-bit or 16-bit PNG? Use 16 bits per channel when another editor will still make strong tone or color changes, because the extra precision keeps smooth skies and gradients from banding. Use 8 bits for final delivery and the web, where 8 bits per channel is the norm and the file is much smaller. ### Does a PNG keep the RAW data, and can RevelRaw export PNG? No. A PNG made from a RAW file holds one rendered version of the photo, not the sensor data or the edit settings, so keep the RAW as your master. RevelRaw exports PNG, JPEG, HEIF and 8- or 16-bit TIFF in sRGB, Display P3 or Adobe RGB, and the first full-quality export is free. It needs macOS 26 or later on an Apple M1 or later Mac. * * * Visit [RevelRaw](https://revelraw.com) to edit a RAW file on your Mac and use the included full-quality export to inspect a rendered PNG. Start with one of the free sample RAW files, check the exported gradients and edges, and keep your original capture separate from the delivery copy.