离线 · 无上传 · 浏览器本地完成Offline · no uploads · runs in the browser

BMP 转 HTML——离线、字节原样内嵌、不解码BMP to HTML offline — raw bytes, no decode

这条路和 BMP 的另一条出口**形态相反**，这一点值得先说清楚。`bmp>html` 是搬运：`FileReader.readAsDataURL` 把文件的原始字节编成 data URI 写进一个 `<img src>`，再包进自带样式表的 HTML——中间**没有解码、没有上画布、没有重编码**。而 `bmp>pdf` 是处理：它把同一份位图解码后铺在**白底画布**上重编码成 PNG 流（这是刻意的，jsPDF 自带的 BMP 解码分支会强走 quality-100 JPEG、无视白底、不给质量旋钮，写出的流还比源文件大）。所以「同一张 BMP 会不会被动过像素」取决于你走哪条：HTML 这条一个字节都不动，PDF 那条必然重走一遍画布。对 BMP 这种通常不带压缩的容器，搬运的代价要如实说：base64 是每三个字节写成四个字符，再加约 2.6 KB 的文档外壳（它随文件名长度增减），而源文件本身通常就比同画面的 PNG 大——画面越平、差距越大，内容越接近噪声、差距越小——所以**这一路绝不是用来变小的**，它是用来把图交给只收网页的地方。反过来，如果你要的是体积、或者要让背景落在白底上，PDF 那条才有压缩存放。其余与位图转网页同形：`alt` 与 `title` 取**文件名主干**（去扩展名、无名时 `converted`）并按属性位置转义；内层 CSS 居中、`max-width: 100%`，版心实测约 800 px，宽图显示会收而文件不动；`data:image/bmp` 在剥远程引用的白名单之内，所以打开这份 HTML 不发任何请求，再往下一步转 Markdown 也不会被剥掉。BMP 在这套界面里**只能作为输入**（浏览器不提供 BMP 编码器），所以没有「从别的格式转成 BMP」这条路。This route is shaped **opposite** to BMP’s other exit, and that is worth saying first. `bmp>html` is carrying: `FileReader.readAsDataURL` turns the file’s raw bytes into a data URI inside an `<img src>`, wrapped in a document with the bundled stylesheet — **no decode, no canvas, no re-encode anywhere**. `bmp>pdf` is processing: it decodes the same bitmap, paints it onto a **white-filled canvas** and re-encodes it as a PNG stream — deliberately, because jsPDF’s own BMP branch forces a quality-100 JPEG, ignores the white fill, offers no quality knob, and writes a stream larger than the file it came from. So whether a BMP’s pixels are touched depends on which page you take: this one changes not a byte, that one necessarily redraws the canvas. For a container that usually ships uncompressed, state the cost honestly: base64 writes four characters per three bytes, plus a wrapper of about 2.6 KB that grows with the length of your file name, and the source is usually bigger than a PNG of the same picture — the flatter the artwork, the wider the gap, and the closer it is to noise, the narrower — **this route is not a way to get smaller**, it is a way to hand the picture to something that only accepts a web page. If what you want is size, or a white ground behind the pixels, the PDF page is the one that stores compressed. Everything else matches the other pictures-to-HTML pages: `alt` and `title` carry the **basename** (extension dropped, `converted` when there is no usable name) escaped for attribute position; the inner CSS centres the image and caps it at `max-width: 100%` inside a measure measured at about 800 px, so a wide picture scales on screen while the file stays untouched; and `data:image/bmp` is inside the allow-list of references allowed to survive, so opening the page issues no request and one more step to Markdown keeps the picture too. One boundary belongs in plain sight: BMP is **input-only** in this interface — browsers cannot encode it — so there is no route that produces a .bmp.