mirror of
https://github.com/ddidderr/wuffs.git
synced 2026-11-18 16:32:11 +01:00
test/3pdata: tweak blinksuite JPEG commentary
This commit is contained in:
1 parent
ca50faaa92
commit
6184059c05
1 file changed
+7
-7
@@ -27,17 +27,17 @@ 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:
|
||||
|
||||

|
||||

|
||||
|
||||
Here's the decodings on the same input by Wuffs:
|
||||
doesn't necessarily give better output. Here's two JPEG images (one sequential,
|
||||
one progressive, both truncated to 1 KiB) decoded by Wuffs:
|
||||
|
||||

|
||||

|
||||
|
||||
Here's the same input, this time decoded by libjpeg-turbo:
|
||||
|
||||

|
||||

|
||||
|
||||
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.
|
||||
|
||||
|
||||
Reference in new issue
Block a user