@MarcinKrk

Hi Marc.

Thank you for your posting your results along with some screen shots. I will be in contact with you at some point tomorrow.

If you wish to discuss anything please make additional comments on this thread.

There are a couple of things I noticed from the screenshots and data given so I have a couple of questions.

Are you working with just one monitor? If not have you plugged both monitors into your nvidia card?

Are all of those drives SATA? Internal or external? What was the drive (Make / model) you put the project on to? Did you export to a second drive? If so, again, the make and model please.

Ray.

Hi, i´m interested in participating and would kindly ask for the download-link. Thx in Advance.

@RichardeRicharde

Hi Richard.

Please visit this page first and read through the terms and conditions for participation in the test.

I will give a link to the files for anyone that complies fully. I am going to be strict about this as I want to be able to verify results as best I can so the eventual list of entries data can be as relied upon as much as possible.

Thank you.

Ray.

@CubeAce

Hi Ray,

VPX14 - I noticed that in the Advanced screen of the export settings, you have version 2.0.5.3999 and I have 2.0.5.4071. Also, for Hardware encoding, you have GPU and I have multiple GPU. I don't know if this makes any difference.

Which version of VPX14 do you have? Mine is 20.0.3.176.

I am getting disastrous results. With VPX, MP4, the best was 11:05, the worst 14:47. I tried different combinations of the Intel/NVIDIA for processing and export. I also put the files on my main working drive SATA, not an external USB drive and the results were the same. Exports go to an SSD.

HEVC - 1 test only, gave me 11:20.

Something is wrong because I should not be getting significantly worse results than others with my hot machine.

Test with MS2023 were about the same - the best so far was 10:59.

John CB

@browj2

Hi John.

You are correct. Under the multiplexer it does say version 2.0.5.4071. That could be down to when I had problems with seeing the MP4 files on my machine and downloaded the KLite codec pack. I can't think of another reason but my exports seem to be in line with the others published so far, and the others seem to be on a predictable curve of performance when looking at individual specs. You do seem to be experiencing a problem. Have you checked all the drives in use for fragmentation? What is the power draw with all those drives? What temp is your CPU reaching? You could be thermal throttling somewhere. Try running CrystalDisk Info to see if your drive temperatures are OK.

The project was originally created in MEP 2019 I think. My version of VPX 14 is 20.0.3.176 (UPD3)

So the project will probably open with old selections present. To that extent I haven't tampered with the project so those that used it before could see if there was any improvement over the earlier versions of the programs. Everyone has downloaded the same file set.

Feel free to change any program performance settings. Only the export settings have to be the same for everyone. Did you read my brief?

There I did say, "You May alter sound buffers or optimise the Import, Processing, and Export settings if that helps but you must include these settings anyway in your reported findings."

Check the folder contains only 1.59GBs of files.

I'm sorry John, but that is all I can think of.

Ray.

@CubeAce

Hi Ray,

I have 2.92GB of files:

Do you see anything out of order there?

My first tests were with NVIDIA Studio driver. I just tested with Game-ready driver and my basic test with NVIDIA for both processing and export was 10:57.

As for drives, they never were stressed at all, mostly 0. That is why I tried with the files on two vastly different drives. No impact.

With NVIDIA for both processing and export, the load was shared by both Intel and NVIDIA - a lot.

I also tried with VPX on my UHD screen rather than running it on a FHD screen. No difference.

I have 3 screens connected.

John CB

@MarcinKrk

Hi Marc.

Thank you for your posting your results along with some screen shots. I will be in contact with you at some point tomorrow.

If you wish to discuss anything please make additional comments on this thread.

There are a couple of things I noticed from the screenshots and data given so I have a couple of questions.

Are you working with just one monitor? If not have you plugged both monitors into your nvidia card?

Are all of those drives SATA? Internal or external? What was the drive (Make / model) you put the project on to? Did you export to a second drive? If so, again, the make and model please.

Ray.

I have 2 monitors (both plugged into Nvidia), but the 2nd one was switched off during that time. My primary one (active at the time) is running at 3440x1440.

Also, project files were located on .M2 SSD disk (Kingston SKC1000240G), and the target file was exported to another HDD (not SSD) - WD RED (WD30EFRX)

@browj2, @CubeAce

Hi John, Ray

. . . . I am getting disastrous results. With VPX, MP4, the best was 11:05, the worst 14:47 . . . .

I am getting similar figures best was 12:27 and 12.33 (RTX/RTX/RTX and UHD/RTX/RTX device options respectively).

I do have some thing odd going on with my OS drive that is causing VPX to intermittently hang for up to 15 secs when exporting, this also affect other programs intermittently - new SSD drive arrives tomorrow.

John EB

Hello @all

I uninstalled all Intel I-GPU drivers after many tests.
Then turn off the Intel UHD in the bios.
At least on laptops there is too much speed loss when the Intel keeps intervening.

This applies not only to Magix but also to other programs.
Especially if OpenCL is used.
Then programs simply use the Intel I-GPU instead of the NVidia card. Although I have set the NVidia in all Windows and NVidia settings.

In general, I have the feeling that the drivers have become worse since version 5XX.XX than version 497.29
It all started when NVidia inserted the ANTI-Mining program code.

All my Movie and Sound files are on a Samsung T5 connected via USB 3.1 Gen1 (5GB).
It hardly ever makes a difference for me whether the files are on the USB SSD or internal SSD, although the internal SSD is 3.5 times faster.

I only had to copy the files to the internal SSD for files with MOV ProRes 10 BIT 4:2:0 800 MBit.

But one thing remains strange for me too.
All typical resolutions up to 4096X2160 run just fine.
Also the typical 8K resolution up to 30 FPS 8Bit 4:2:0 200 MBIT

Only 5.7K and 5.3K and 5.9K always causes problems.

@browj2

Hi John.

There appears to be two files extra. I don't know if your program created them if your backup was turned on.

I have shown where they are on your image, crossed through.

The rest appear to be OK.

@MarcinKrk

Hi Marc.

To have both GPUs active one monitor must be plugged into the motherboards graphics port or use a headless ghost as a monitor dummy plug or the Intel GPU will not be active. Whether it is selected to be used or not in the program, the program will use it anyway. It doesn't show a lot of activity, mainly a little 3D work but it seems to help a lot. I can't find anything wrong with your drive arrangement. 👍

@johnebaker

Hi John.

Maybe @Marc-Goder has a point about the newer nvidia drivers but the only flaw in that is I am running the nvidia 528.49 Studio driver. Unless it is only effecting RTX graphics cards. I personally have never found disabling the Intel GPU to be advantageous. It used to be that in Windows Settings / System / Graphics settings that you could set the programs to run in three different modes but now that is best left alone for Windows to control IMO.

As far as 5.7K and 5.3K and 5.9K resolution goes for exporting into a viewable file. They are not monitor display panel resolutions except within one or two digital camera back screens. They are not even sensor sizes but constructed by individual companies trying to do something different with their 8K sensors. Mainly it's marketing trying to convince a market of something they hope to convince they need. If a sensor can record 8K and you want more resolution then record at 8K. There is no pixel binning in such systems. No noise reduction in lower light to be gained. Just a mess of pixels trying to be shifted around to fit another eventual screen resolution. One advantage for such files would be for video cropping down to centralise a subject within a frame, and another is an alternative to digital zooming.

There are no plans at present within Europe to do any digital broadcasting at 8K. Even 8K streaming content outside of Japan is limited. My concern with such resolutions is when real 8K becomes the norm, these early attempts with smart phones and action cameras with such small sensors are really going to show the processing involved. One can always find a use for such things but why export to a non native screen resolution?

My view so far is that the output times posted are not far off the mark if you look at the individual components and allow a slight deviation for potential things like motherboard layouts and ram timing differences, amount of CPU processing cores / speeds etc. There is a slight upward curve on the more capable component layouts. I'm at the bottom so far from the posted specs 👍

If nothing else it has shown up some anomalies.

If it's any consolation to anyone I am still having intermittent mouse freezes. Sometimes none at all and then problems for two hours. Then it may clear up for another day.

Ray.

@Marc-Goder

Hi

. . . . Only 5.7K and 5.3K and 5.9K always causes problems . . . .

The 5.3K format we have discussed in another topic - it has an AR of 8:7 and is a compromise AR for creative content users which enables recording a video once, and from the result create both 16:9 and 9:16 AR videos to fit the various Social media sites and devices.

The 5.7K is 2:1 AR may also be a compromise format to give an approximate widescreen cinema AR video.- technically the AR should 1.9:1 for DCI video.

The 5.9K is 16:9 AR however the cameras I know of that have this record in 10 or 12 bit depth and/or RAW video format.

John EB

The 5.3K format we have discussed in another topic - it has an AR of 8:7 and is a compromise AR for creative content users which enables recording a video once, and from the result create both 16:9 and 9:16 AR videos to fit the various Social media sites and devices.

@johnebaker

Actually an intelligent solution from GoPro. Satisfy your entire clientele and target group.

Something is wrong because I should not be getting significantly worse results than others with my hot machine.

@browj2 🤣 I hear ya there bro, @Marc-Goder has a machine 1/3 - 1/4 the power of mine yet has posted times almost a minute faster,😣😖🙈🤣🤔

Perhaps tho, for my case anyway, it's as @CubeAce says the intention of this isn't a race but to help users with poss purchases in reg to working with MS/VPX. My PC doesn't have an on board GPU, yours has but you're getting crap export times? don't know what to make of that 🤷‍♂️ All I know is my PC does sweet FA in the first & second parts of this project where the graphics are playing, it just bumbles through until it reaches the video part.

I'm going to fiddle with this & see why it doesn't work too well, I've already tried Muting parts of the graphic start & finish but that didn't improve playback at all.... I'm going to try turning off n on effects to see which is the rotten apple,

But yep Ray I'll keep the project as it was downloaded, the Exports settings as requested & I'll post asap 👍

​​​​​@Former user

​​ what does that mean

"PC does sweet FA"

If you had asked me what kind of video editing software the user uses with the hardware specified in your signature, my answer would have been "DaVinci Resolve".

Perhaps tho, for my case anyway, it's as @CubeAce says the intention of this isn't a race but to help users with poss purchases in reg to working with MS/VPX.

@Former user

That's right. An example is this post

https://www.magix.info/de/forum/idee-benchmark-erfahrungsaustausch--1303866/

At the time, @RichardeRicharde were also looking for new hardware.

@Former user

Hi Gid.

In MEP 2022 the project acts exactly as you describe in both playback and rendering. VPX I get smoother playback and slightly faster rendering. The Intel GPU only shows less than 1% of any video decoding and varies between 4-7% 3D graphics during playback, No encoding. It doesn't show any activity worth reporting when rendering and yet it does better than if I set Windows to ignore the Intel GPU. To quote you...🤷‍♂️.

There are so many things that can go wrong and difficult to track down which is why I prefer to build as much of a system for myself as I can. Why I always have a power supply rated well above what should be needed. My CPU cooler is so large I had to cut parts of my full tower case for it to fit. Four 50mm fans just keeping my hard drives cool. CPU never gets above 56℃. At max I haven't seem my nvidia card over 78℃. So far in the last 3 years I've had one SSD failure and one SATA drive failure. I am slowly running out of space on my C: drive and will need to get a larger one. I also have one internal WD black SATA drive showing signs of slowing down but it's seven years old and in semi constant use. I have all folders and files backed up twice and not connected to the PC (Plus one backup off site, not cloud based). I run check-ups on the backups every eight months or so. Otherwise they are just storage. Internally I only have about 5TB of storage across four disks including the operating system and its clone drive, and an external USB WD RED 2TB drive.

The only thing about these types of tests I'm not keen about is they never represent any individuals workload.

I seldom can do a project without resorting to pre-rendering but for this I didn't want some people who didn't need to competing against those that did. That would have skewered the results of playback capabilities.

It's just as well I didn't put forward one of my other projects for this 😅. I felt I had to try to find a balanced project. I had no idea of any consequences that may occur.

It's just as well I didn't put forward a more complex one.

As for the animation sections. Take a real close look at the timeline expanded to see what is causing that to slow down. You won't believe how many individual objects make up a few seconds, all with slightly different settings.

Try soloing one track and slowly scroll across it.

Ray.

 

​​ what does that mean

"PC does sweet FA"

@Marc-Goder Sweet Fanny Adams but you're prob better off Googling that

If you had asked me what kind of video editing software the user uses with the hardware specified in your signature, my answer would have been "DaVinci Resolve".

I have no idea why you would write that ? I use Adobe AE for effects & Magix MS for basic video editing because I've used it for many yrs so I'm comfortable with it, it rarely crashes & it has a fast export

I had no idea of any consequences that may occur.

@CubeAce Yep. look at the shit you've stirred up 🤣 Watching Lucy on TV, I'll get back to this 👍

@Former user

@CubeAce Yep. look at the shit you've stirred up 🤣 .

Sounds accurate 👍👍👍🤣

Ray.

@Former user

I have no idea why you would write that ?

I read again and again for DaVinci that you need a CPU with as many cores as possible, strong graphics cards and a lot of RAM.

256GB is a lot of memory.

@Marc-Goder I bought what was top of the list, biggest CPU, most amount of RAM & biggest GPU, there were a few bigger items I could have got, there was a 48GB GPU, for another £2k I could have got a Threadripper 64core CPU & my £1k motherboard says it can take 'Max. 2048GB', not sure if that's right or how that works.. but my purchase might be a good example of doing your research before purchase, although my PC has the most of this & that, your render times on this benchmark are better, on the Vegas Pro benchmark my time was about 14th but has dropped to 29th position now due to new times posted, In this Magix benchmark @CubeAce's time, 7.17mins is only about 20secs slower than mine, he has a decent CPU, 8core 5ghz compared to my 32core 3.7ghz, his 4GB 1650 GPU is less then the 6GB 1660 in my old PC & quite a bit less than the 24GB of RTX 3090 I have now, yet my performance doesn't beat his hands down, far from it,

My PC Is a bit different tho, It has a history of being a server that's why it has no on board GPU, it's designed to be stable not speedy, I can & often leave it on for days without any heating problems or noticeable lag, & I can have several programs open & working no problem. it's swings n roundabouts, power & speed over stability, can you have all options 🤷‍♂️..

I think this benchmark will be good, It will help others in the future make a choice about what to buy or vice versa, it'll give them more info to take with them when they do go shopping, Intel CPU's generally seem to be a better choice but I don't know why some others on here have huge render times when they have Intel CPU's, ,

@Former user @RichardeRicharde @johnebaker @browj2 @MarcinKrk @Marc-Goder

It was brought to my attention last night that HEVC at 50 fps and the '4K' setting was not possible. It was remiss of me not to add that information. Simply use the 25 fps '4K' HEVC setting but manually change the frame and GOP settings to 50.

I will add this information to the topic.

Thank you to our German users who brought this to my attention. I am guilty of not being that provincial and not visiting or adding the German site who are more active than I had imagined.

Ray.