diff --git a/test/3pdata/nia-checksums-of-blinksuite.commentary.md b/test/3pdata/nia-checksums-of-blinksuite.commentary.md index 9a4835f2..8f4178fa 100644 --- a/test/3pdata/nia-checksums-of-blinksuite.commentary.md +++ b/test/3pdata/nia-checksums-of-blinksuite.commentary.md @@ -26,6 +26,21 @@ a3cecef6](https://github.com/google/wuffs/commit/a3cecef69aa04f1145f0c84b92481e8 performance trade-offs if Wuffs were to try to exactly match on all input, not just on well-formed JPEGs. +As pointed out by the a3cecef6 commit message, matching libjpeg-turbo exactly +doesn't necessarily give better output. Here's two decodings of JPEG images +(one sequential, one progressive, both truncated to 1 KiB) by libjpeg-turbo: + +![Decoding by djpeg](../data/peacock.default.truncated.decoded-by-djpeg.png) +![Decoding by djpeg](../data/peacock.progressive.truncated.decoded-by-djpeg.png) + +Here's the decodings on the same input by Wuffs: + +![Decoding by Wuffs](../data/peacock.default.truncated.decoded-by-wuffs.png) +![Decoding by Wuffs](../data/peacock.progressive.truncated.decoded-by-wuffs.png) + +Zoom in on all four images (in four different tabs, if viewing in a browser), +and the first pair isn't obviously better or worse than the second pair. + --- `test/3pdata/blinksuite/large-size-image-crash.jpeg` is simply a very large