Someone with Threadripper 1950x or Skylake X 7900x Please Test This

tripleflip18 wrote on 8/29/2017, 11:10 PM

Hi, i have Samsung NX1 camera and I love the quality it has, however my computers 6 core Intel I7 5820k clocked at 4.1ghz and E5 2699 v4 22 Core at 2.6ghz can't playback smoothly in vegas 14 or the 15. its playing back at 29.97fps with small hick ups few times per second (its really painful for my head) and I have to trans code footage before i can edit with pleasure. I've just sold my E5 2699 and can afford the new Threadripper or skylake X, IF the timeline playback is butter smooth without any hick ups. Can someone Please test this footage for me with those processors if its perfect, i don't want to build a system and have the exact problem because its software related and not hardware

Here is test footage http://www.filedropper.com/sam0288 150mb file size..

My comp also has nvidia 1080 as well as 950pro m2 ssd... and rest of the spec is top of the line....

pretty please, thank you so much!

Back to post

Comments

Former user wrote on 9/3/2017, 5:20 AM

Hi bitman, I'm using the original Red car FHD test, not a 4K version😊

Perhaps that could explain the difference?

I've no problem doing all tests again if theres a "ready to go" uprezzed 4k version available for download in a zip.

I just discovered from one of Nicks faqs that Oldsmoke left a link, but it doesnt work.

I prefer to use an already made up version in case of any errors in following oldsmokes instructions, by myself or others. Maybe Nick, while he's doing nothing else😀 could make it available for all.

NickHope wrote on 9/3/2017, 6:43 AM

OldSmoke's instructions for converting the 1080 benchmark to 4k are here. I still have the folder he shared in 2014 on my HDD but it's 9.75 GB, I have Thailand upload speeds, and I'm about to dump my Google Drive 1TB allowance and go back to my lowly 15GB.

NickHope wrote on 9/3/2017, 7:00 AM

Actually the files in my 9.75GB folder from 2014 were not rendered according to OldSmoke's instructions from 2017 because they are AVC/MP4 files, not XAVC-I/MXF. He obviously changed his method.

Actually, using XAVC-I/MXF at the moment is probably better than AVC because it gets decoded by mxfxavc.dll in both VP14 and VP15, so eliminating the current issues surrounding compoundplug.dll to so4compoundplug.dll.

fr0sty has already rendered and uploaded 4k XAVC-I/MXF media according (I assume) to OldSmoke's instructions on the other thread so we can all use those. I suggest all the test reports using that set of media are posted on that thread from now on as the title is more appropriate.

Former user wrote on 9/3/2017, 8:34 AM

Sounds good Nick. Its important that the same render settings are used by all. For example the Intel QSV quality setting is in a range of 1 to 7, where 7 is best quality, 4 is default.

But thats on my system .. Haswell Core i7-4790K which was I think the 3rd. generation of QSV. The other generations probably used a different number of steps. So its important to pick something and stick with it, say "default".

For Nvenc I dont know what the variables are, over generations, again simplest might be to select the default values.