[Resolved] How to decrease the render time

4sol-n wrote on 10/17/2020, 3:12 PM

1. My end files are 1-2h long

2. The size of 1h file to be upload to SV is about ~13 GB [ 28k Kbits 1920x1080p 60 fps ]

3. Rendering with Main Concept AVC, values all the same, except resolution, which is needed to be 2560x1440p

4. It takes about 7:30 h to render a 1:40h long file ...... (kill me)

5. My PC Specs:

 

Any tips?

(I can't lower the resolution and the bit rate)

Thank you.

Comments

j-v wrote on 10/17/2020, 3:15 PM

Upgrade the vegas version to the latest VPro 18 of Magix Vegas.

 

met vriendelijke groet
Marten

Camera : Pan X900, GoPro Hero7 Hero Black, DJI Osmo Pocket, Samsung Galaxy A8
Desktop :MB Gigabyte Z390M, W11 home version 24H2, i7 9700 4.7Ghz,16 DDR4 GB RAM, Gef. GTX 1660 Ti with driver
566.14 Studiodriver and Intel HD graphics 630 with driver 31.0.101.2130
Laptop  :Asus ROG Str G712L, W11 home version 23H2, CPU i7-10875H, 16 GB RAM, NVIDIA GeForce RTX 2070 with Studiodriver 576.02 and Intel UHD Graphics 630 with driver 31.0.101.2130
Vegas software: VP 10 to 22 and VMS(pl) 10,12 to 17.
TV      :LG 4K 55EG960V

My slogan is: BE OR BECOME A STEM CELL DONOR!!! (because it saved my life in 2016)

 

4sol-n wrote on 10/17/2020, 3:17 PM

Upgrade the vegas version to the latest VPro 18 of Magix Vegas.

 

And what next? Main Concept AVC will be rendering faster there or you are talking about other method?

4sol-n wrote on 10/17/2020, 3:26 PM

I just read comments about 334 version, it renders twice as long as the previous one.

michael-harrison wrote on 10/17/2020, 3:30 PM

Do you need a constant bit rate?

Is this destined for Youtube?

System 1:

Windows 10
i9-10850K 10 Core
128.0G RAM
Nvidia RTX 3060 Studio driver [most likely latest]
Resolution        3840 x 2160 x 60 hertz
Video Memory 12G GDDR5

 

System 2:

Lenovo Yoga 720
Core i7-7700 2.8Ghz quad core, 8 logical
16G ram
Intel HD 630 gpu 1G vram
Nvidia GTX 1050 gpu 2G vram

 

j-v wrote on 10/17/2020, 3:30 PM

Not with me , on laptop from signature QSV rendering is fastest , followed by NVENC .
In your case with the Magix AVC codec. Those are a lot newer and better than the old Mainconcept which work only on CPU.

met vriendelijke groet
Marten

Camera : Pan X900, GoPro Hero7 Hero Black, DJI Osmo Pocket, Samsung Galaxy A8
Desktop :MB Gigabyte Z390M, W11 home version 24H2, i7 9700 4.7Ghz,16 DDR4 GB RAM, Gef. GTX 1660 Ti with driver
566.14 Studiodriver and Intel HD graphics 630 with driver 31.0.101.2130
Laptop  :Asus ROG Str G712L, W11 home version 23H2, CPU i7-10875H, 16 GB RAM, NVIDIA GeForce RTX 2070 with Studiodriver 576.02 and Intel UHD Graphics 630 with driver 31.0.101.2130
Vegas software: VP 10 to 22 and VMS(pl) 10,12 to 17.
TV      :LG 4K 55EG960V

My slogan is: BE OR BECOME A STEM CELL DONOR!!! (because it saved my life in 2016)

 

4sol-n wrote on 10/17/2020, 3:37 PM

Do you need a constant bit rate?

Is this destined for Youtube?

yes and yes

I'm recording a game

Not with me , on laptop from signature QSV rendering is fastest , followed by NVENC .
In your case with the Magix AVC codec. Those are a lot newer and better than the old Mainconcept which work only on CPU.

OK, I will brb with the answer after I finish rendering this file and update to magix 18 and will try to render the same file with magix avc

 

Musicvid wrote on 10/17/2020, 5:36 PM

Rule #1: Don't upscale! It does nothing except to make your files bigger and slower to encode. It does not make them "UHD."

Rule #2: When in doubt, refer to rule #1.

As for upscaling and machine encoding, you will never see me go there. It makes no sense, twice.

As for Mainconcept software encoder, Sony AVC is better quality at a given bitrate.

Former user wrote on 10/17/2020, 7:09 PM

The game he is playing may be in 1440p, and he wants to preserve the detail. If he is uploading at 1440P to trigger the VP9 codec, which people say give better quality, in my last testing a couple of months ago, it is not necessary for 1080P60 footage, it will get VP9 encoding, 1080P30 however won't

john_dennis wrote on 10/17/2020, 8:30 PM

The listed display is 1920x1200.

Optimal resolution 1920 x 1200 @ 60 Hz; WUXGA

4sol-n wrote on 10/18/2020, 1:35 AM

1440p is for YouTube VP9, for the better quality, I do not want to upscale, but the f YouTube forces me to do that to get the best quality after uploading my video and yes, I'm recording in 1080p.

You can google this issue and see quality comparison 1080p and upscaled 1440p with 0.05 sharpen effect.

4sol-n wrote on 10/18/2020, 5:14 AM

Not with me , on laptop from signature QSV rendering is fastest , followed by NVENC .
In your case with the Magix AVC codec. Those are a lot newer and better than the old Mainconcept which work only on CPU.

It worked and even better than I thought.

Before in v.13 I was using Mainconcept AVC and to render a 1:38m footage it was needed more than 8 hours

Now in v.334, using Magix AVC it needs 4 hours

And if using Magix HEVC only 1:15h! This is ridiculous!

Thank you for the intel. 🙂

 

Musicvid wrote on 10/18/2020, 6:25 AM

@4sol-n

I am well aware of the internet hype surrounding upscaling and VP9; however, I have seen no quantitative tests using SSIM, PSNR, and VMAF to support such claims of alchemy, a pet peeve of mine. Nor am I able to draw such a conclusion based on my experience (signature below). Also, using machine encoders runs exactly counter to your stated goal of maximizing quality, meaning you have completely different criteria than I have chosen to address.

It is time for me to step out of the discussion; you seem to favor speed over quality, I won't be able to address that, but keep testing as you develop a more critical eye and refined techniques. Glad to have been of "some" help. Welcome to the forum, and best of luck!

dream wrote on 10/19/2020, 2:10 AM

with that hardware and proper vegas(what i was asking) render time should be max.10min

Former user wrote on 10/19/2020, 3:21 AM

It's a 1h38m video, that he's upscaling to 1440p, AND adding sharpening. The sharpening is the important bit. In Resolve the estimate is 2hours, so longer then it takes with Vegas. The reason for this is because Resolve hightly utalises GPU but the GPU this guy and I have uses 80% GPU with without the sharpening, and that would take 38mins, but add sharpening, and the GPU is maxed out and becomes the bottleneck. Yes you could get a better GPU and probably have an encode time not that much longer than 38mins, but not with current GPU.

In this situation with his PC, Vegas actually is faster than Resolve

4sol-n wrote on 10/19/2020, 3:38 AM

Actually not only sharpening effect, but with Sony saturation and Sony color correction.

EricLNZ wrote on 10/19/2020, 4:34 AM

So dream's 10 minute comment was just that - "a dream".

michael-harrison wrote on 10/19/2020, 11:18 AM

@EricLNZ @dream With even the best hardware available today, there's no way a 2hr video at 1080p 60hz will *ever* render in 10 minutes from VP. That's not a "dream" that's a drug induced "fog" :-)

System 1:

Windows 10
i9-10850K 10 Core
128.0G RAM
Nvidia RTX 3060 Studio driver [most likely latest]
Resolution        3840 x 2160 x 60 hertz
Video Memory 12G GDDR5

 

System 2:

Lenovo Yoga 720
Core i7-7700 2.8Ghz quad core, 8 logical
16G ram
Intel HD 630 gpu 1G vram
Nvidia GTX 1050 gpu 2G vram

 

dream wrote on 10/19/2020, 1:40 PM

So dream's 10 minute comment was just that - "a dream".

maybe for you not for me, you need to wake up.

@EricLNZ @dream With even the best hardware available today, there's no way a 2hr video at 1080p 60hz will *ever* render in 10 minutes from VP. That's not a "dream" that's a drug induced "fog" :-)

clearly vegas cant do this which is why im talking about vegas for improvements. .

i did 2 hours long video in 10 min at different software.

michael-harrison wrote on 10/19/2020, 1:59 PM

@dream "different software"

Perhaps you haven't noticed. This is the Vegas forum :-)

System 1:

Windows 10
i9-10850K 10 Core
128.0G RAM
Nvidia RTX 3060 Studio driver [most likely latest]
Resolution        3840 x 2160 x 60 hertz
Video Memory 12G GDDR5

 

System 2:

Lenovo Yoga 720
Core i7-7700 2.8Ghz quad core, 8 logical
16G ram
Intel HD 630 gpu 1G vram
Nvidia GTX 1050 gpu 2G vram

 

4sol-n wrote on 10/19/2020, 2:10 PM

i did 2 hours long video in 10 min at different software.

What software?

walter-i. wrote on 10/19/2020, 2:45 PM
 

i did 2 hours long video in 10 min at different software.

@dream

It's easy for you to prove, dreamer.
Take a comparable video file upscaling to 1440p, AND adding sharpening, throw your software on the timeline and record the rendering with OBS Studio.

Then we can all see this here - at least I'm very curious.

Former user wrote on 10/22/2020, 11:34 PM

@4sol-n

I am well aware of the internet hype surrounding upscaling and VP9; however, I have seen no quantitative tests using SSIM, PSNR, and VMAF to support such claims of alchemy, a pet peeve of mine. Nor am I able to draw such a conclusion based on my experience (signature below).

I uploaded the 4K stress test video to youtube as 2160, 1440p, 1152P, and 1080P, then compared the 1080P resolution for each. That is the order they are shown in, a clockwise direction from top left. Remarkable difference. You can go back years ago and find people talking about the 1440P upscale advantage on this forum, although as another user suggested the 2160P upload gives an even better 1080P re-encode than the 1440P upload, but even so 1440P makes a big difference

4sol-n wrote on 10/23/2020, 2:10 AM

 2160P upload gives an even better 1080P re-encode than the 1440P upload

1440p is already triggering VP9, what's the point in up scaling the footage even more? But I haven't tested it yet, need media data and frames in motion to compare.

Your 2160 just have less sharpening, so it's looking blurry.

Former user wrote on 10/23/2020, 2:32 AM

1080P60 will trigger VP9, the samples are all VP9, but there's a clear difference in quality. The thing is though, the 4K with my test was real 4K, real 1440P etc, not sharpened upscaled 1080P so I think you're right, if source video is 1080P probably not worth uploading at 4K unless you do a real upscale and as that's such a slow heavy computational process not really feasible for many videos