Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
oughly 4 times that of 4K resolution content (Table 4B vs Table 4A), average decoding speed of 720p<br />
resolution content is almost a factor of 1.6 of 1080p content (Table 4C vs Table 4B), and that of SD/XD<br />
resolution content is close to a factor of 2 as compared to decoding speed of HD720p content (Table 4D<br />
vs Table 4C).<br />
<strong>Deliver</strong> <strong>High</strong> <strong>Quality</strong>, <strong>High</strong> <strong>Performance</strong> <strong>HEVC</strong> <strong>via</strong> <strong>Intel®</strong> <strong>Media</strong> <strong>Server</strong> <strong>Studio</strong><br />
36<br />
Further, Fig. 10 shows a combined graph of variation of <strong>Media</strong> <strong>Server</strong> <strong>Studio</strong> <strong>HEVC</strong> Software Decoder<br />
performance (fps) across various resolutions test sets. On the Haswell platform, including the rendering<br />
overhead, one 60 fps UHD4K stream or two 30 fps UHD4K streams, or up to eight HD1080p streams can<br />
be easily decoded in realtime.<br />
<strong>Intel®</strong> <strong>Media</strong> <strong>Server</strong> <strong>Studio</strong> <strong>HEVC</strong> Software Encoding Computational<br />
Scalability Tests (on E5 <strong>Server</strong>)<br />
For measurement of computational scalability beyond 4 core i7-4770K system used in tests thus far, an E5<br />
<strong>Server</strong> (<strong>Intel®</strong> Xeon® Processor E5-2697v2, 12 Core, 2.7 GHz, 2 processors per board, total of 24 physical<br />
cores) is employed. Currently such servers do not inherently support <strong>Intel®</strong> graphics acceleration. The<br />
purpose of this test is to demonstrate threading scalability of the Intel <strong>Media</strong> <strong>Server</strong> <strong>Studio</strong> <strong>HEVC</strong><br />
software encoder when many more cores are available. This also addresses an application where very<br />
high resolution video (4K and above) is needed to be <strong>HEVC</strong> encoded in realtime in software on many<br />
available cores.<br />
<strong>Quality</strong> Evaluation Results for Computational Scalability Tests for 4:2:0 8-bit<br />
Using the test set of 4:2:0 8-bit content as discussed earlier <strong>via</strong> Tables 0A-0B and test procedure establish<br />
earlier, we first evaluate video quality of the <strong>Intel®</strong> <strong>Media</strong> <strong>Server</strong> <strong>Studio</strong> <strong>HEVC</strong> for the three main TU modes, ie,<br />
TU1, TU4, and TU7, for the encoding configuration used to provide high degree of scalability.<br />
The results of quality tests are shown in Tables 5A and 5B for UHD4K and HD1080p test sets respectively.<br />
Table 5A Relative <strong>Quality</strong> Evaluation Results on UHD4K test set for <strong>Intel®</strong> <strong>Media</strong> <strong>Server</strong> <strong>Studio</strong> <strong>HEVC</strong> Software Codec<br />
in TU1, TU4, and TU7 modes with respect to MPEG <strong>HEVC</strong> HM14 Codec<br />
UHD4K Test<br />
Sequence<br />
TU1 BD rate, %age TU4 BD rate, %age TU7 BD rate, %age<br />
Y U V Y U V Y U V<br />
1 Park_joy -0.18 1.92 -0.63 13.50 7.92 16.47 26.97 25.10 33.01<br />
2 Ducks_take_off -0.93 6.81 9.79 15.92 -16.84 6.27 28.53 -0.84 23.39<br />
3 CrowdRun 0.37 2.99 2.14 17.33 18.18 20.35 35.46 46.38 48.19<br />
4 PeopleOnStreet 0.15 -7.15 -6.22 20.65 21.06 25.66 40.38 52.82 59.84<br />
5 Traffic 0.32 -9.01 -15.64 20.86 16.15 23.76 36.52 41.70 42.65<br />
6 Nebuta_Festival 6.91 47.07 42.79 20.02 16.68 5.40 46.90 33.35 16.72<br />
Average -0.05 -0.89 -2.11 17.65 9.30 18.50 33.57 33.03 41.41
Change language