PNG's lossless compression works by finding repetition, flat areas of identical color, sharp edges, the same pattern repeated across a screenshot or icon. That's why a BMP full of graphics or text can shrink by half or more once converted. A photographic BMP behaves very differently, since natural photos are full of subtle noise and gradients with almost no exact repetition, so the same conversion might only trim 10 or 15 percent, occasionally landing larger than the same photo saved as a high quality JPEG.
Check what kind of image you're actually converting
Before committing to PNG, think about what's in the BMP. A scanned photo, a picture of a person, or anything with sky, skin tones or natural texture will almost always end up smaller and just as usable as a JPEG or WebP through BMP to JPG or BMP to WebP. Save PNG for screenshots, icons, diagrams and anything with flat color or text.
Not every 32-bit BMP actually has transparency
The BMP format reserves a fourth byte per pixel in its 32-bit variant for an alpha channel, but not every program that writes 32-bit BMPs actually stores real transparency data there. Some just pad that byte with zeros or leftover values instead of a genuine alpha mask. This tool preserves whatever is actually present in the file, so if the resulting PNG looks fully opaque or oddly transparent, the original BMP likely never carried real alpha information despite its 32-bit format.
Simple graphics are where PNG earns its keep
PNG compresses by predicting each pixel from its neighbors before applying its main compression pass, a step that pays off enormously on flat regions where neighboring pixels often match exactly. Icons, UI screenshots and line art hit that sweet spot consistently, which is why they're the clearest win for a lossless format like PNG over anything lossy.