Hi @bitman Thanks. Up to now we’ve compared similar frame rates, see the link that BruceUSA gives in the very first post of this thread. (29.97fps). So its interesting that the issue exists with a different source frame rate clip, in your case.

I would appreciate it if you can confirm on your system that a same frame rate source clip, 29.97fps clip also causes a duplicate frame, thanks, no hurry. Happy Xmas.

I don't know if I did the right test, because I did not follow this subject because I cannot compare both GPU's in the OP.
I did a test with files from my Gopro 5 (2,7K 50p), Panasonic (FHD 50p) and my smartphone(FHD framerate variable ~30p) on the laptop from signature..
All frames of the files do have movement to better see a possible duplicated frame.
All are rendered in a FHD 29,970 project to that same FHD 29,970p.
Putting the files back on a FHD 29,970 timeline and I could not detect any duplicated frame.

Maybe you can

Hi @j-v thanks for chipping in. I found that my smartphone vfr was ok.

The methology is to create a source file longer than say 10 seconds, in 29.97fps, @bitman found the error could also occur in source material of 23.98, but for simplicity maybe keep source at 29.97fps. If you cannot set your cams to that fps just use BruceUSA's clip that he makes available in the very first post in this thread.

Render out a 29.97fps clip the same resolution as the source, so source and rendered clip are both uhd or both fhd.

Put the source and rendered clip on the timeline, one directly above the other.

Set the compositing mode to “difference” on both tracks to see if there is anything in the preview window other than a black screen after 2 seconds and the 27th/28th frame. If it’s not completely black you may find that the 2s;27th frame and the 2s;28th. frame are identical.

The render must be the Magix Nvenc codec. I see you have an Nvidia gtx in your laptop and PC, so thats great to confirm or not.

@Former user
I followed precisely your instructions with a file from my Gopro set to NTSC 2,7 K 29,970 fps.
This is the result

Thanks @j-v so no issues.

@j-v There may be a minimum clip length before it happens, I know I did say longer than 3 seconds, your clip is about 3 seconds long.

I just did a quick test of only 3 seconds of the Original BruceUSA clip, it showed no issues.

I then took 10 seconds, from the start also, of the original clip and the issue is there.

So perhaps you can use at least 10 seconds or longer test clip? Maybe you can try the Original clip also.

I cannot download the original Vimeo clip, while in the same instance of Vimeo in get the possibility to download my own files there and some from others but not this one.
Sorry I can't help you any further. I tried to help 2 times, the last time following your instructions with an original 29,970 file I normally not use and I did not see what I had to see and more I'm not going to try because the whole subject is not affecting me and my way of working.

aok @j-v it doesn’t affect me either, I live in Pal land and it’s extremely unlikely that I would ever need to render out to 29.97. For me its simply a bug that should be addressed.

Thanks for all your help.

Hi @j-v thanks for chipping in. I found that my smartphone vfr was ok.

The methology is to create a source file longer than say 10 seconds, in 29.97fps, @bitman found the error could also occur in source material of 23.98,

@Former user Just to clarify: the source material at 23.98 on the timeline was OK, no duplicates, but after render to 29.97 via NVENC, and putting the rendered result on the timeline, then you see the duplicate frame in the rendered result.

There are quite often I hear that GPU render is bad and at time its true. But since then GPU render has improved. I made a short video and want to ask any one of your eagle eyes out there can see tell the different in quality?

https://vimeo.com/305223909

@BruceUSA I sent you a PM. Would you mind having a look at it?

@Former user I now used a 29.97 source clip, and rendered this to the same 29.97, there are no duplicates here. I used the method described by JN_ (with the compositing method to "difference") So the new NVENC hw in the Turing NVIDIA cards may be OK?

But come to think about it, isn't it normal to add duplicates or extra frames to convert from 23.98 to 29.97 like in my original test?

Oh and by the way, I also live in a PAL country, so I would normally never render in 29.x/30/60...

Oh and by the way, I also live in a PAL country, so I would normally never render in 29.x/30/60...

I saw that @bitman when I was confirming you had an Nvidia card in your signature, thats why I asked you to confirm, or not my test results. I don’t know much about adding frames going from 23.97 to 29.97 but all the tests I’ve done are 29.97 to 29.97. I used BruceUSA's original test piece, see his very first post in this thread. To confirm I created 3 new test pieces from 2 different cameras, two 4k and one fhd, all at 29.97fps.

All 4 pieces exhibit the same duplicate frame issue. Waag also appears to confirm although I’m not sure of the details, which Nvidia card, was he using Bruceyusa's test piece etc.

I suppose what i’m attempting to do is seperate out whether its a Nvidia card issue or a Magix software issue, so the need for just more than my own Gtx 1080 as a test card.

@Former user I now used a 29.97 source clip, and rendered this to the same 29.97, there are no duplicates here. I used the method described by JN_ (with the compositing method to "difference") So the new NVENC hw in the Turing NVIDIA cards may be OK?

 

This maybe points to an issue with the cards rather than the software, probably need a bit more input to confirm.

The duplicated and overwritten frame problem is clearly a bug in the Vegas implementation of NVENC rendering. Using Happy Otter Scripts, there is no duplicated frame and the preview window remains dark throughout the test clip. The easiest way to test is to add a Timecode Fx (e.g. absolute frames), then render and add back to the timeline which HOS can do automatically. Here are a couple of screenshots at 2:29 showing the frames of original and rendered are the same. Also a screenshot of the preview window and scopes when Difference is chosen.

 

Hi @wwaag Just curious, in case its HW instead of software related, when you did your test which Nvidia card did you use, thanks. Also, can I take it that your source clip is BruceUsa's vce test, but just renamed on track.

There are quite often I hear that GPU render is bad and at time its true. But since then GPU render has improved. I made a short video and want to ask any one of your eagle eyes out there can see tell the different in quality?

https://vimeo.com/305223909

@BruceUSA I sent you a PM. Would you mind having a look at it?

Oldsmoke. Sorry I just read your post and I did not see your PM. You can try to send again?

There are quite often I hear that GPU render is bad and at time its true. But since then GPU render has improved. I made a short video and want to ask any one of your eagle eyes out there can see tell the different in quality?

https://vimeo.com/305223909

@BruceUSA I sent you a PM. Would you mind having a look at it?

Oldsmoke. Sorry I just read your post and I did not see your PM. You can try to send again?

It’s in your messages within the forum.

@Former user

Hi @wwaag Just curious, in case its HW instead of software related, when you did your test which Nvidia card did you use, thanks. Also, can I take it that your source clip is BruceUsa's vce test, but just renamed on track.

I am using a 1050ti. It is my understanding that Nvenc renders are the same regardless of which card (1060, 1070 ,etc) is being used. I purchased the card early this year when prices were outrageous. And Yes, I did use BruceUSA's clip.

One thing I might add which I overlooked when I first read an earlier post. Getting duplicate frames when the frame rate changes from 23.976 to 29.97 is to be expected unless you use some type of frame interpolation (which Vegas doesn't do too well). If you disable resampling, you will always gets duplicate frames regardless of how it is rendered.

wwaag

@wwaag Ok, thanks for confirming. So at the moment the only outlier is @bitman 's result, although he didn’t use the original BruceUsa clip, maybe he might like to test using it.

Good to know that about 23.976 to 29.97.

I think its only muddying the waters doing anything other than 29.97 to 29.97 with a view to confirming the duplicate frame using Nvenc.

I've left a query with Magix about this, got ticket no. and automated reply. As you know the more info. the better.

I would be surprised if it was a hardware issue, i.e. GTX vs RTX generational issue/fix, but anything is possible.

But in a sense, although I don’t fully understand it all, you managed to avoid the duplicate frame issue in software, using HOS, so realistically that should eliminate the hardware possibility.

@Former user

For grins, I rendered the BruceUSA clip using Vegas 16 NVENC. I added a timecode to easily check for duplicate or missing frames. I couldn't find any duplicate frames although I admit I didn't look at each of 855 frames. I did look at 2:27 and it was OK. All in all, it seems to be OK, except that the Vegas render dropped the last frame and also cut off the audio starting with the previous frame. I also added an HOS NVENC render which worked OK. Here are the results.

wwaag

@Former user I am not able to download BruceUsa clip that's why I used some 29.97 test clip I had kept for such purposes.

@wwaag

I am using a 1050ti. It is my understanding that Nvenc renders are the same regardless of which card (1060, 1070 ,etc) is being used.

I am not sure the Nvenc renders are the same for all cards:

The Turing NVIDEA cards (RTX 20xx series) have a different NVENC HW versus the Pascal cards like yours. The current consumer (although not cheap!) range of Turing cards have only 1 NVENC encoder unit whereas the higher range of Pascal cards have 2 encoder units. According to Nvidea this should be not a problem with the new architecture (!)... Also weather Vegas 16 is up to date to make full use of the new features and optimizations with the new Turing Nvenc hardware regarding driver/dll/api or whatever is unclear.

Here is a list of what is new to the Turing NVENC encoder unit (and the second HW unit NVDEC for decoding).

• Up to 3x Decode throughput with multiple decoders on professional cards (Quadro & Tesla)
• Higher Quality Encoding - H.264 & H.265
• Higher encoding efficiency (15% lower bitrate than Pascal)
• HEVC B-frames support
• HEVC 4:4:4 decoding support

@Former user

For grins, I rendered the BruceUSA clip using Vegas 16 NVENC. I added a timecode to easily check for duplicate or missing frames. I couldn't find any duplicate frames although I admit I didn't look at each of 855 frames. I did look at 2:27 and it was OK. All in all, it seems to be OK, except that the Vegas render dropped the last frame and also cut off the audio starting with the previous frame. I also added an HOS NVENC render which worked OK. Here are the results.

wwaag

“The duplicated and overwritten frame problem is clearly a bug in the Vegas implementation of NVENC rendering. Using Happy Otter Scripts, there is no duplicated frame and the preview window remains dark throughout the test clip.”

Hi @wwaag I'm confused, I assumed from your previous posting that you were able to “fix” this issue by using HOS. Perhaps you hadn’t yet done the test that you just now did?

I also found the last frame was truncated, I assume to make both file lengths the same because of the inserted duplicate frame at 2s;28f.

I know you have your own preferred method of testing using timecode, is there any chance you could humor me one last time and check using the compositing difference mode test method?