Detect SOF55 marker to distinguish JPEG-LS from baseline JPEG

[hackermd]

The library currently checks the first three bytes of the file (0xFF, 0xD8, and 0xFF) to determine whether it is a JPEG file:

file-type/core.js

Lines 154 to 159 in bb4f822

	if (this.check([0xFF, 0xD8, 0xFF])) {
	return {
	ext: 'jpg',
	mime: 'image/jpeg',
	};
	}

However, to distinguish between baseline JPEG and JPEG-LS the forth byte should also be checked. It it is 0xF7 ("SOF55" or "JPG7" segment), then the file is a JPEG-LS rather than JPEG file and image/jls should be returned instead of image/jpeg.

[hackermd]

It may also be a good idea to check for 0xE8 ("APP8" segment) to further distinguish SPIFF from baseline JPEG.

[Borewit]

Sounds good to me, why don't you create a PR?
Can you link to some supporting documentation or reference material?

[Borewit]

Small reminder @hackermd

[somebody1234]

not sure this is required, afaict neither jpeg-ls nor spiff have a different file extension (or even mimetype)

[Borewit]

not sure this is required, afaict neither jpeg-ls nor spiff have a different file extension (or even mimetype)

It seems to me that they do have a dedicated registered MIME-type: https://www.iana.org/assignments/media-types/image/jls

Other references:

• http://fileformats.archiveteam.org/wiki/JPEG-LS
• https://en.wikipedia.org/wiki/Lossless_JPEG#JPEG-LS

[Borewit]

Looking to some jls fixtures found here, some of them have a string J1U6G at offset 0xF7 (247), some others not, none of them have the SOF55 or JPG7 marker.

So I am not entirely sure how to detect JPEG-LS yet.

[Borewit]

I found this information: JPEG Lossless Compression (ISO/IEC 14495), File type signifiers and format identifiers.

Which describes the SOF55 maker, which is actually encoded as JPG7 at 0xF7. But the way I read it, more conventional to use the 0xFFF7 at offset 2, which is a marker to indicate the improved lossless baseline.