![]() I can tell you one thing, I have no intention of spending $4000 on a top of the line TR, $500 on a motherboard, hundreds more for ram, hundreds for a beefy power supply not to mention cooling, fans and a case for that monstrosity, when I can spend $150 buying an i5 9400 with a $50 motherboard and some cheap ram and call it a day. Pay attention to the x265 encode times and compare them against the QSV HEVC encode times. ^^^Tests done by the people behind Handbrake, using a 22C/44T Xeon, take a look at the x265+medium 3840x1714 10Mb/s encode at a speed of 14 fps for comparison my i3 7100 2C/4T encodes a 4096*1714 10Mb/s (the same test file as they used, just not cropped) using Handbrake+QSV HEVC+Ubuntu 19.10 at 25 fps and at normal bit rates you will not be able to tell the difference between the 2. Very interesting read about why software doesn't scale with higher core counts. ![]() Software encoders run into severe bottlenecks as the number of threads goes up: If you use the Net Flix standard 16 Mb/s you would be hard pressed to see the difference and QSV hevc is very good and beats x264 under most circumstances and comes close to x265's quality but way faster speeds and with enough bit rates you are not going to see the difference. It's finally happened, the final version of Blender 2.8 is here UI redesigned, EEVEE realtime renderer, new cycles features, 2D grease pencil support and s. #Blender 2.8 release date how toReality: If you are talking very low bit rate AVC, then yes, encoders like x264 will beat qsv, nvenc or vce, if you use one of the slower presets, but I'm talking bit rates that no one actually uses in practice, for instance 10 Mb/s for 4096*1714. How to make Animated Pixel Art in Blender 2.8,Argentina vs Atletico Madrid Qatar Copa del Mundo,Real. ![]() CPU encoding means better quality compared to GPU encoding so for preview its maybe ok, but for the final rendering CPU is always better.Īnd these days we have Ryzen/Threadripper with many many cores.I take it you are one of the many people, like the rest that posted in this thread up to now, that probably do not do all that much encoding at all or do not know how to read benchmarks.Ĭlaim 1: Software based encoders produce better quality than hardware based encoders: ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |