Vegas 16 Benchmarks

fr0sty wrote on 8/27/2018, 3:44 PM

I've seen posts claiming 16 offers no improvement on render speed, so I decided to give it a test. I figured I'd start a thread where we could post benchmarks and their results. Vegas 16 does come with a demo project that should be useful for benchmarking performance across different systems, I will post later tonight with my results rendering that project. Unfortunately, it uses assets that are not backwards compatible with 15, so we'll have to figure out something else for cross version testing. Here is a 15 vs. 16 4K NVENC render test I did, with rather surprising results of a (roughly) 25% performance boost in 16.

Test specs:

media: GH5 4K 10 bit VLOG 150mbps codec. No fx or changes of any sort applied on timeline.

Rendered 1 minute of it in both versions to Magix AVC 4K NVENC at 24p

System:

Ryzen 7 1800x

64GB DDR4 3000mhz

GTX970 4GB

Windows 10

Comments

Former user wrote on 8/27/2018, 4:06 PM

what was the comparative gpu & cpu percentages?

fr0sty wrote on 8/27/2018, 6:48 PM

Here are the render times (Magix AVC codec, untouched GH5 4K 10 bit 150mbps VLOG clip, 8 bit project settings, best render quality, smart resampling):

Vegas 15 4K NVENC: 1 minute 14 seconds

Vegas 15 CPU only: 4 minutes 41 seconds

Vegas 16 NVENC: 55 seconds (faster than real time!)

Vegas 16 CPU Only: 4 minutes 41 seconds

Former user wrote on 8/27/2018, 7:27 PM

About those cpu & gpu comparison figures?

Also, an increase in speed at a decrease in quality, is not an upgrade. Not that i'm saying that, I"m not going anywhere near vp16 at this stage, but you need to confirm it's maintaining quality. Perhaps you could upload some short clips to a file dump site?

fr0sty wrote on 8/27/2018, 7:51 PM

I listed the CPU/GPU comparison in the post directly above yours?

If you'd like to compare the output files, here they are:

test1 = Vegas 15 NVENC

test2 = Vegas 16 NVENC

Test 3 = Vegas 15 CPU only

Test 4 = Vegas 16 CPU only

https://drive.google.com/open?id=1QW1e1QKNNXv2rk4YzFoitfX0iPd40HU1

Former user wrote on 8/27/2018, 7:58 PM

oh sorry, I"m not good at communicating, but actually those figures are very interesting, especially how vp15 & 16 software encode is still identical.

What I meant to ask is a comparision of gpu & cpu utalisation between vp15 & vp16, did you notice an increase in both with vp16?

fr0sty wrote on 8/27/2018, 8:12 PM

ah, I will have to do some more tests to see... I'll try to post some more detailed tests, as well as a project file and some media we all can use to benchmark tomorrow.

Former user wrote on 8/27/2018, 10:02 PM

I couldn't see any difference. I also used the MSU Quality Measurement Tool to compare them, It can't see a difference, although that's not really how you're supposed to use it. should use source file to compare to encoded files. It's interesting how they can speed up nvenc. I guess they've sped up the serving of files to nvenc. Interesting whatever they've done.

Only thing different is that the vp16 file doesn't report as being VBR. A missing meta tag?