There's a great deal of confusion on this subject, which is not surprising because there are several different compression methods all known as "JPEG". The commonly used method is "baseline JPEG" (or its variant "progressive JPEG"). The same ISO standard also defines a very different method called "lossless JPEG". And if that's not confusing enough, a new lossless standard called "JPEG-LS" is about to hit the streets.

When I say "lossless", I mean mathematically lossless: a lossless compression algorithm is one that guarantees its decompressed output is bit-for-bit identical to the original input. This is a much stronger claim than "visually indistinguishable from the original". Baseline JPEG can reach visual indistinguishability for most photo-like images, but it can never be truly lossless.

Lossless JPEG is a completely different method that really is lossless. However, it doesn't compress nearly as well as baseline JPEG; it typically can compress full-color data by around 2:1. And lossless JPEG works well only on continuous-tone images. It does not provide useful compression of palette-color images or low-bit-depth images.

Lossless JPEG has never been popular --- in fact, no common applications support it --- and it is now largely obsolete. (For example, the new PNG standard outcompresses lossless JPEG on most images.) Recognizing this, the ISO JPEG committee recently finished an all-new lossless compression standard called JPEG-LS (you may have also heard of it under the name LOCO). JPEG-LS gives better compression than original lossless JPEG, but still nowhere near what you can get with a lossy method. It's anybody's guess whether this new standard will achieve any popularity.

It's worth repeating that cranking a regular JPEG implementation up to its maximum quality setting *does not* get you lossless storage; even at the highest possible quality setting, baseline JPEG is lossy because it is subject to roundoff errors in various calculations. Roundoff errors alone are nearly always too small to be seen, but they will accumulate if you put the image through multiple cycles of compression (see section 10).

Many implementations won't even let you get to the maximum possible setting, because it's such an inefficient way to use regular JPEG. With the IJG JPEG software, for example, you have to not only select "quality 100" but also turn off chroma downsampling to minimize loss of information. The resulting files are far larger and of only fractionally better quality than files generated at more reasonable settings. And they're still slightly lossy! If you really need lossless storage, don't try to approximate it with regular JPEG.