Vegas 16 Render Speed same...

Stiven wrote on 8/27/2018, 2:59 PM

Hello,

So I have been expecting Vegas 16 to get some improvement in rendering speed.

I record most of the content in 4k, and when I render them 1 minute of 4k video needs 4 min and 15 sec to render using CPU. If I switch on GPU rendering it takes 3 min and 45 seconds.

My system:
CPU: i9 7980xe
GPU: Nvidia GTX 1080ti

Also one more sad thing is that when I am using GPU it use only 12% of GPU Power?

Is this working as intended or not ? I am working with Magix AVC to render my videos.

Comments

fr0sty wrote on 8/27/2018, 3:39 PM

I did a test to see if it was the same, and it was not. I did a side by side test using a 24p 4K 10 bit clip from my GH5, untouched (no processing), using NVENC on a file that was 1 minute long. The total render time was 1:16 in Vegas 15, 56 seconds (about 25% faster) in Vegas 16.

My system:

Ryzen 7 1800x CPU

GTX 970 GPU (4GB)

Windows 10

Video of the test:

Windows seems to incorrectly report GPU usage, when I use GPUz the numbers are far higher for GPU utilization.

Last changed by fr0sty on 8/28/2018, 11:37 AM, changed a total of 3 times.

Systems:

Desktop

AMD Ryzen 7 1800x 8 core 16 thread at stock speed

64GB 3000mhz DDR4

Geforce RTX 3090

Windows 10

Laptop:

ASUS Zenbook Pro Duo 32GB (9980HK CPU, RTX 2060 GPU, dual 4K touch screens, main one OLED HDR)

D7K wrote on 8/27/2018, 5:02 PM

I just did a test of 4K from my Gx8 in both 15 & 16 (i7 7700 /32 gid/AMD 480 8 gig/ SSD render and both are the same - real time for a 30 second file, about 27 seconds.

Former user wrote on 8/27/2018, 6:14 PM

I just did a test as well and my results came out to be the same as D7K and fr0sty. Render time for Sony Vegas may have decreased just a bit but nothing major as some may have hoped. As for the GPU usage, when I render a video, my GPU says that it is using around 10% so your usage seems about average for your render settings.

fr0sty wrote on 8/27/2018, 6:51 PM

I'd call a 25% increase pretty major... it is now on par or faster than any other encode tool I've used.

Last changed by fr0sty on 8/27/2018, 6:51 PM, changed a total of 1 times.

Systems:

Desktop

AMD Ryzen 7 1800x 8 core 16 thread at stock speed

64GB 3000mhz DDR4

Geforce RTX 3090

Windows 10

Laptop:

ASUS Zenbook Pro Duo 32GB (9980HK CPU, RTX 2060 GPU, dual 4K touch screens, main one OLED HDR)

RealityStudio wrote on 8/28/2018, 11:19 AM

Just got Vegas 16, for me gpu render time is the same as with Vegas 15. I rendered one minute of 4k 100mbps footage that had FilmConvert applied (it supports gpu) and a watermark, in both cases it took 1:30 to render with my 1080ti gpu and 8 core Amd 1700 cpu.

 

Peter_P wrote on 8/28/2018, 11:32 AM

When you compare render times, please add information about the framerate of your footage and output (and format/codec). The rendertime is different for 25p and 30p and surely for 50/60p.

On my i7-8700k rendering 60 seconds of UHDp30 XAVC-S 100Mbps footage with no FX to Vp16 MAGIX HEVC 'Internet 4k UHD 2160p 29.97 (QSV) take 68 seconds.

Rendering the same footage in Vp15 to MAGIX AVC (H.264) 2160p 29.97 (QSV) takes 51s and to Intel HEVC 30-40Mbps takes 70s.

fr0sty wrote on 8/28/2018, 11:39 AM

I've updated my benchmark to include the fact that it was 24p (not 23.97) footage I was using.

Reality, in addition to frame rate, what formats did you render to?

Systems:

Desktop

AMD Ryzen 7 1800x 8 core 16 thread at stock speed

64GB 3000mhz DDR4

Geforce RTX 3090

Windows 10

Laptop:

ASUS Zenbook Pro Duo 32GB (9980HK CPU, RTX 2060 GPU, dual 4K touch screens, main one OLED HDR)

Trensharo wrote on 8/29/2018, 12:22 AM

I did a test to see if it was the same, and it was not. I did a side by side test using a 24p 4K 10 bit clip from my GH5, untouched (no processing), using NVENC on a file that was 1 minute long. The total render time was 1:16 in Vegas 15, 56 seconds (about 25% faster) in Vegas 16.

My system:

Ryzen 7 1800x CPU

GTX 970 GPU (4GB)

Windows 10

Video of the test:

Windows seems to incorrectly report GPU usage, when I use GPUz the numbers are far higher for GPU utilization.

The speeds are about the same. You really have to test this with longer clips. It's not much different - not enough to matter, anyways; and if you use QSV/NVENC/VCE then the render speeds are fast enough that it shouldn't be that much of an issue, anyways...

Secondly, GPUz may be throwing NVENC usage in with everything, and reporting high usage due to that. NVENC is an ASIC block on the GPU that does only Video Encoding (there is also NVDEC). It's analogous to QSV. It's not like CUDA-based Encoders (i.e. the H.264 Encoder in Resolve Studio). It has a completely different usage pattern, even if some simple tools are reporting the "GPU usage" as high when they are activated.

Windows is likely discriminating the NVENC usage because the driver knows that NVENC is not a GPU, so why attribute its usage to the GPU. NVENC/DEC are distributed with the GPU, but they are no more GPU usage than usage of the Image Signal Processor in an SoC is CPU usage. It's just in the same package.

Use the software on a machine with an Intel CPU and render with QSV, then check to see how the Nvidia GPU is being used ;-P

Task Manager has different categories listed under GPU that allow you to see exactly how your "GPU" is being used: 3D, Copy, Video Encode, Video Decode, Compute, etc.

Peter_P wrote on 8/29/2018, 12:45 AM

in addition to frame rate, what formats did you render to?

@fr0sty

It is listed in my post by the name of the used encoder.

Vp16 MAGIX HEVC 'Internet 4k UHD 2160p 29.97 (QSV)

Vp15 MAGIX AVC (H.264) 2160p 29.97 (QSV)
Vp15 Intel HEVC 30-40Mbps

OldaC wrote on 8/30/2018, 4:27 PM

 

Windows seems to incorrectly report GPU usage, when I use GPUz the numbers are far higher for GPU utilization.

May I ask, how did you activate the progress dialog windows with all the details?? Thank you.

fr0sty wrote on 8/30/2018, 5:41 PM

https://www.semw-software.com/en/extensions/

usbsound wrote on 8/30/2018, 5:47 PM

Why does not anyone use the example project that comes in Vegas 16 to do render tests?

fr0sty wrote on 8/30/2018, 8:56 PM

Because it is not backwards compatible with Vegas 15, so there is no way to do side by side comparisons.

Systems:

Desktop

AMD Ryzen 7 1800x 8 core 16 thread at stock speed

64GB 3000mhz DDR4

Geforce RTX 3090

Windows 10

Laptop:

ASUS Zenbook Pro Duo 32GB (9980HK CPU, RTX 2060 GPU, dual 4K touch screens, main one OLED HDR)

Trensharo wrote on 8/30/2018, 11:45 PM

Take the source media and reproduce with the same effects.

Takes some effort, but if you think it's worth it :-P

dream wrote on 8/31/2018, 1:16 AM

i dont find any difference ,but how gpu utilization is 15-30% only ,why ? even cpu is 60%

i am asking not 100% but 80% is good

bitman wrote on 8/31/2018, 3:28 AM

@fr0sty can you share your material to try out the 25% speed increase you claim?

From the few quick tests I have done on my stuff and my PC (4K video without any editing nor plugins), I do not see any rendering speed difference 15 to 16, in fact it was never faster, mostly the same, occasionally even a fraction slower (but that could be coincidence or just minute render spread, I have not done any statistically meaningful trials).

The only significant difference is using dynamic ram 0 or 200, but that applies to both VP15 and VP16.

APPS: VIDEO: VP 365 (21 - build 187), VP 365 20, VP 19 post (latest build -651), (uninstalled VP 12,13,14,15,16 Suite,17, VP18 post), Vegasaur, a lot of NEWBLUE plugins, Mercalli 6.0, Respeedr, Vasco Da Gamma 16 HDpro XXL, Boris Continuum 2024, Davinci Resolve Studio 18, SOUND: RX 10 advanced Audio Editor, Sound Forge Pro 17, Spectral Layers Pro 10, Audacity, FOTO: Zoner, DXO, Luminar, Topaz...

  • OS: Windows 11 Pro 64, version 23H2 (November 2023)
  • CPU: i9-13900K (upgraded my former CPU i9-12900K), Air Cooler: Noctua NH-D15s
  • RAM: DDR5 Corsair 64GB (5600-40 Vengeance)
  • Graphics card: ASUS GeForce RTX 3090 TUF OC GAMING (24GB) 
  • Monitor: LG 38 inch ultra-wide (21x9) - Resolution: 3840x1600
  • C-drive: Corsair MP600 PRO XT NVMe SSD 4TB (PCIe Gen. 4)
  • Video drives: Samsung NVMe SSD 2TB (980 pro and 970 EVO plus) each 2TB
  • Mass Data storage & Backup: WD gold 6TB + WD Yellow 4TB
  • MOBO: Gigabyte Z690 AORUS MASTER
  • PSU: Corsair HX1500i, Case: Fractal Design Define 7 (PCGH edition)
  • Misc.: Logitech G915, Evoluent Vertical Mouse, shuttlePROv2

 

 

OldaC wrote on 8/31/2018, 4:14 AM

Please, could someone with Ryzen 2700 or 1700 CPU render SampleProject in Vegas16 using "Internet HD 1080p 25 fps" not NVENC and tell me the time it tooks and how the utilization of CPU was? Thank you!

Path to the project is: C:\ProgramData\VEGAS\VEGAS Pro\16.0\SampleProject

 

usbsound wrote on 8/31/2018, 10:48 AM

Because it is not backwards compatible with Vegas 15, so there is no way to do side by side comparisons.

Example project of vegas 16 for vegas 15

Sample Project 16 to 15

 

bitman wrote on 9/4/2018, 1:50 AM

@fr0sty can you share the clip, I would like to repeat the test on my system, thanks.

APPS: VIDEO: VP 365 (21 - build 187), VP 365 20, VP 19 post (latest build -651), (uninstalled VP 12,13,14,15,16 Suite,17, VP18 post), Vegasaur, a lot of NEWBLUE plugins, Mercalli 6.0, Respeedr, Vasco Da Gamma 16 HDpro XXL, Boris Continuum 2024, Davinci Resolve Studio 18, SOUND: RX 10 advanced Audio Editor, Sound Forge Pro 17, Spectral Layers Pro 10, Audacity, FOTO: Zoner, DXO, Luminar, Topaz...

  • OS: Windows 11 Pro 64, version 23H2 (November 2023)
  • CPU: i9-13900K (upgraded my former CPU i9-12900K), Air Cooler: Noctua NH-D15s
  • RAM: DDR5 Corsair 64GB (5600-40 Vengeance)
  • Graphics card: ASUS GeForce RTX 3090 TUF OC GAMING (24GB) 
  • Monitor: LG 38 inch ultra-wide (21x9) - Resolution: 3840x1600
  • C-drive: Corsair MP600 PRO XT NVMe SSD 4TB (PCIe Gen. 4)
  • Video drives: Samsung NVMe SSD 2TB (980 pro and 970 EVO plus) each 2TB
  • Mass Data storage & Backup: WD gold 6TB + WD Yellow 4TB
  • MOBO: Gigabyte Z690 AORUS MASTER
  • PSU: Corsair HX1500i, Case: Fractal Design Define 7 (PCGH edition)
  • Misc.: Logitech G915, Evoluent Vertical Mouse, shuttlePROv2