Comment #11 on issue 54 by
fbarch...@google.com: cwebp is slow
https://bugs.chromium.org/p/webp/issues/detail?id=54#c11
good catch... I had old version in my path.
down to 5.38 seconds.
c:\work>timex c:\libwebp-0.5.0-windows-x86\bin\\cwebp.exe -q 75
bryce_big.jpg -o bryce_big.webp
Saving file 'bryce_big.webp'
File: bryce_big.jpg
Dimension: 11158 x 2156
Output: 4681758 bytes Y-U-V-All-PSNR 36.98 42.66 44.25 38.27 dB
block count: intra4: 88047
intra16: 6183 (-> 6.56%)
skipped block: 903 (0.96%)
bytes used: header: 535 (0.0%)
mode-partition: 503170 (10.7%)
Residuals bytes |segment 1|segment 2|segment 3|segment 4| total
macroblocks: | 15%| 30%| 34%| 19%| 94230
quantizer: | 36 | 29 | 23 | 16 |
filter level: | 11 | 6 | 5 | 11 |
TIMEX 5385.00 ms (5.38 seconds),
c:\libwebp-0.5.0-windows-x86\bin\\cwebp.exe -q 75 bryce_big.jpg -o
bryce_big.webp
My current Z620 (Sandy Bridge Xeon) is a faster machine than when I first
reported it on a Z600, but thats at 2.19x faster than the one I had in my
path.
size/quality doesnt exactly match your test? different command line?
5.38s on Windows vs 3.43 on Mac is 1.56x slower?
I'm using prebuilt Windows binaries from the webp site, but assume they
were built with Visual C 32 bit?
Is the code optimized for Windows 32 bit?