Images & photos · Image Converter & Compressor
Why PNG has no quality slider and how to shrink one anyway
· How it works
image-formats png image-quality
PNG is lossless, so a quality control would have nothing to trade away; that is why the browser ignores the quality argument for PNG. This post explains what does determine PNG size and the three levers that actually shrink one.
The quality slider that does nothing — why moving it leaves a PNG's size unchanged, and what that tells you about the format
When PNG is selected, ToolAcre hides the quality field instead of leaving a disabled or ineffective slider on screen. The format table marks `qualityApplies` false, and the conversion plan sets quality to undefined. Moving a number that never reaches the encoder cannot explain a size change, so the interface removes the false decision entirely.
This is also a diagnostic clue. If a user expected “quality 50 PNG,” their intended operation was probably fewer pixels, fewer colours through a specialized optimizer, or a lossy destination. Those are different transformations with different consequences. Naming the desired sacrifice is more useful than searching for a universal compression knob.
Lossless means every pixel is kept — PNG's filter-plus-DEFLATE pipeline and why its size depends on content, not a setting
The repository supports one precise claim: PNG’s available encoder is treated as lossless and stores the pixels it is handed. It does not expose PNG row filters, DEFLATE levels or optimizer passes, so the outline’s internal pipeline description is outside product evidence. A canvas export may choose those details without giving ToolAcre a control.
Lossless applies to encoding, not every preceding operation. If the image was resized, the plan calls the conversion lossy because resampling already changed pixel values. If the source was JPEG, PNG faithfully stores pixels decoded from an already-lossy source and cannot restore removed detail. These boundaries prevent the word “lossless” from becoming a blanket promise.
The repository verifies lossless PNG output but does not implement PNG filter or DEFLATE controls
Fewer pixels is the strongest lever this interface exposes for PNG. Percentage mode scales both axes, while exact-pixel mode derives a missing dimension or fits inside a box when aspect ratio stays locked. The planned output appears before conversion, and a device budget can reduce an excessive request with a warning.
For screenshots, resize with legibility in mind. Thin type and one-pixel rules may no longer map cleanly after resampling, even though the final PNG encoder preserves the newly calculated pixels exactly. Test at the display size where readers will use the image. A smaller byte count is not useful if interface labels become hard to read.
Lever two: fewer colours — palette PNG, why flat interface screenshots compress well and photographs do not
A palette with fewer colours can shrink some PNGs, but ToolAcre does not provide colour quantization, palette selection or dithering. It decodes to a canvas and asks for `image/png`; no setting chooses indexed colour or a palette size. Presenting palette reduction as an available lever would therefore misstate the interface.
If a palette workflow is required, use a dedicated optimizer whose output mode and dithering are documented, then inspect the result. Keep this converter’s role clear: it can preserve PNG alpha, resize pixels and write a PNG. Product boundaries are valuable because they stop an easy local tool from being mistaken for a complete PNG optimization suite.
Palette reduction is not offered by this browser converter
A different destination trades different properties. JPEG is always lossy and cannot store alpha, so it suits opaque photographic content only after a matte decision. WebP keeps alpha but ToolAcre describes its browser output as lossy at every quality. Either can be smaller for some images, yet no fixed saving is promised.
Content should drive the trial. A photograph embedded in a PNG is a candidate for JPEG or WebP; a crisp diagram with transparency may deserve PNG despite its bytes. Export one representative copy, inspect edges and gradients, and compare the measured size. Converting by extension alone ignores what the image must preserve.
Worked example: a 2.4 MB full-screen PNG — comparing the three levers and their effect on legibility
Take a full-screen documentation screenshot that is too large for its destination. First export PNG at the original dimensions to establish what a plain canvas re-encode produces. Next reduce both axes to the real display requirement and inspect small type. Finally return to the original source and create a same-dimension WebP for a format comparison.
Write down the three measured files rather than inserting a fictional 2.4 MB starting point or universal reduction. The resized PNG isolates pixel count; the WebP isolates lossy format choice. If either harms the details readers need, the larger PNG is the correct result. A workflow is successful when it meets the use case, not when it wins an arbitrary percentage.
Worked example without invented source sizes or promised savings
External optimizers may explore PNG filters, compression effort, palette PNG and metadata chunks. Those controls are not exposed here, and metadata loss through canvas is a side effect rather than a guaranteed cleaning operation. Use the Image Metadata Privacy Tool when verified metadata removal is the actual objective.
The converter also does not promise byte-identical PNG round trips. Lossless pixel encoding and identical container bytes are different claims. Browser encoders can arrange the file differently while decoding to the same raster. Compare visible pixels for image fidelity and use format-aware tools if container-level reproducibility matters.
External PNG optimizers and palette modes remain outside the tool
PNG has no quality slider in this panel because there is no lossy-quality decision to forward. To shrink an output, reduce dimensions deliberately or choose a destination whose documented losses fit the content. The tool keeps those choices separate and reports when resizing itself changes pixels.
Keep the source, work from it for every experiment, and inspect downloaded files at their intended scale. That procedure avoids repeated degradation and makes every saved byte traceable to a real decision. A hidden or ineffective slider would offer comfort; explicit geometry and format choices offer control.