Lines Matching refs:JDCT_IFAST

973                 JDCT_IFAST: less accurate integer method [legacy feature]
976                 JDCT_FASTEST: fastest method (normally JDCT_IFAST)
979         was measured in minutes.  Thus, JDCT_IFAST provided noticeable
982         milliseconds, and thus the performance benefits of JDCT_IFAST are much
984         instructions, JDCT_IFAST and JDCT_ISLOW have similar performance.  On
985         other types of CPUs, JDCT_IFAST is generally about 5-15% faster than
990         levels above 90, however, the difference between JDCT_IFAST and
992         JDCT_IFAST incurs generally about a 1-3 dB loss in PSNR relative to
994         JDCT_IFAST with quality levels above 97.  The algorithm often
997         libjpeg-turbo, JDCT_IFAST is not fully accelerated for quality levels
1282                 JDCT_IFAST: less accurate integer method [legacy feature]
1285                 JDCT_FASTEST: fastest method (normally JDCT_IFAST)
1288         was measured in minutes.  Thus, JDCT_IFAST provided noticeable
1291         milliseconds, and thus the performance benefits of JDCT_IFAST are much
1293         instructions, JDCT_IFAST and JDCT_ISLOW have similar performance.  On
1294         other types of CPUs, JDCT_IFAST is generally about 5-15% faster than
1301         between JDCT_IFAST and JDCT_ISLOW becomes more pronounced.  With images
1302         compressed using quality=97, for instance, JDCT_IFAST incurs generally
1304         larger for some images.  If you can avoid it, do not use JDCT_IFAST