And yet another V18 vs V17 rendering test... Positive! :)

adimatis wrote on 8/6/2020, 1:57 PM

Ok, now that I got Vegas 18 able to render, I wanted to do a test for rendering speed as comparison.

Probably not very useful to anyone, but just wanted to share the result with you.

I used the same clips for both, for a total of 1 minute in timeline. These are .mp4, FHD 50 fps, from my Lumix G85 camera.

I did no edits at all, no cuts, crossfades, anything. Just added a LUT - the same of course for both - and adjusted the level at 50%.

I did the test both with Intel QSV and NVDEC as decoders, in both, but this did not seem to produce any difference.

The rendered format is Magix AVC, FHD 50fps. I used the preset Internet HD 1080p 50 fps.

The following data is from graphs and values of Windows Activity Manager and they vary +/- about 2-3% I would say.

My laptop, as by signature, has an Intel UHD 630 (card 1) and a nVidia 1050Ti (card 2).

 

On Vegas 17:

Playing the timeline: CPU 39%, Intel 17%, 1050Ti 6%

Rendering with NVENC: CPU 46% average (spikes at 51%), Intel 6%, 1050Ti 15% (spikes at 19%) Render time 1:54

Rendering without NVENC: CPU 88%, Intel 5%, 1050Ti 3% Render time 3:36

 

On Vegas 18:

Playing the timeline: CPU 33%, Intel 20%, 1050Ti 10%

Rendering with NVENC: CPU 35% average (spikes at 44%), Intel 7%, 1050Ti 16% (spikes at 19%) Render time 1:15

Rendering without NVENC: CPU 90%, Intel 5%, 1050Ti 5% Render time 3:24

 

So, in my little test, for rendering in the format I use mostly, for some reason, v18 seems to be more efficient, although the values of hardware utilization are not terribly different than v17. The rendering time, with NVENC, is better in v18 by about what 35%?

I would say this is a positive thing! :)

More than rendering times, I will explore more the editing experience - first and foremost the stability and smoothness. And of course, the precision of various processes like stabilization, etc.

 

 

Comments

aboammar wrote on 8/6/2020, 6:39 PM

Thanks for sharing. Just wondering, does the faster rendering in V18 affects the output quality in anyway?

HP Z1 AIO Workstation G3

OS: Windows 10 Pro 64bit

Display: 23.6" UHD 4K

CPU: Xeon E3-1270 v5  quad-core @ 3.60GHz, 8MB cache, up to 4GHz with Intel Turbo Boost Technology

GPU: nVidia Quadro M2000M 4GB

RAM: 32GB DDR4 2133MHz ECC memory

System Drive: 1TB M.2 (2500MB/s)

Working Drive: 1TB M.2 (2500MB/s)

Storage Drive: 3GB SSD (500MB/s)

Video: Vegas Pro 16 Suite / DaVinci Resolve 16 Studio

Audio: PreSonus Studio One Pro 5

Graphics: CorelDraw Technical Suite 2020 / Xara Designer Pro X365

Image Editing: Corel PhotoPaint 2020 / Corel PaintShop Pro X9 Ultimate / PHASEONE Capture One Pro 11

3D Graphics: Maxon Cinema 4D Studio 10

Camera: Sony A7S II / A7 III

Website: www.innoviahouse.com

Vimeo: vimeo.com/innoviahouse

Former user wrote on 8/6/2020, 6:58 PM

Hope some of those people who do that quality analysis benchmarks of videos will be able to check this soon

Former user wrote on 8/6/2020, 8:16 PM

I still see the speed difference between vp18's internal nvenc encoder and voukoder. 140fps VP, 180fps VP via voukoder. That's a difference in speed of 28%. So ofcourse I will always use voukoder, I just don't understand how a 3rd party can make a faster encoder though

lan-mLMC wrote on 8/6/2020, 9:08 PM

Rendering has two sections.

1) is that Vegas generate a "Vegas raw frame" and pass it to Encoder, like Nvenc, Mainconcept AVC, x264, etc.

2) is that Encoder compresses "Vegas raw frame" to a kind of format , like AVC, HEVC, etc.

So any section of the two may make low usage of CPU, GPU, Memory.

lan-mLMC wrote on 8/6/2020, 9:11 PM

I still see the speed difference between vp18's internal nvenc encoder and voukoder. 140fps VP, 180fps VP via voukoder. That's a difference in speed of 28%. So ofcourse I will always use voukoder, I just don't understand how a 3rd party can make a faster encoder though

It may be because Vegas uses built-in "nvenc.dll" while Voukoder uses GPU Driver's "nvenc.dll". GPU Driver's "nvenc.dll" is more new if you often upgrade the GPU Driver.

Former user wrote on 8/6/2020, 9:32 PM

Interesting. in VP17 part of the problem was it would serve frames to it's internal encoder slower. this was due to a pause every 30frames or so. I can't tell if there is a delay with vp18, but I would suggest it's still there, but to confirm I need to choose a input file that encodes slower so I check

andyrpsmith wrote on 8/7/2020, 6:44 AM

I get 35% faster render with V18 too.

(Intel 3rd gen i5@4.1GHz, 32GB RAM, SSD, 1080Ti GPU, Windows 10) Not now used with Vegas.

13th gen i913900K - water cooled, 96GB RAM, 4TB M2 drive, 4TB games SSD, 2TB video SSD, GPU RTX 4080 Super, Windows 11 pro

adis-a3097 wrote on 8/7/2020, 8:54 AM

Anyone else having problems rendering to Sony MXF HDCAM SR?

With highlights pushed into clipping, I get very glitchy video which I don't get with other rendering formats.

Vegas Preview window:

MPC snapshot:

Can someone repro this, or is it just me? :)

Edit:

Oh, cancel this! Imported rendered video back into Vegas - no glitches! It's just my MPC player. :)

Marco. wrote on 8/7/2020, 9:28 AM

MPC and VLC are tricky when it's about how they handle color and levels. They offer a lots of variants and many if not most of them output something off the standard.

walter-i. wrote on 8/10/2020, 7:33 AM

I still see the speed difference between vp18's internal nvenc encoder and voukoder. 140fps VP, 180fps VP via voukoder. That's a difference in speed of 28%. So ofcourse I will always use voukoder, I just don't understand how a 3rd party can make a faster encoder though

@Former user
It is crucial that you have both options and can use them from Vegas Pro.
Everything has its advantages and disadvantages .......