10bit 422 may be less likely to show issues as AMD GPUs can't decode it. Wonder if that is a clue.

RedRob, I’m in the same boat. I do get better results with with iGPU off and GPU set to AMD or Nvidia (depending on the computer I’m using) but there’s still issues and I suspect just having an Intel component brings those issues to life. Again, I could be wrong but I don’t think I’ve seen the memory issue with all non-Intel parts.

@RogerS 

Interesting! That'd imply his GPU isn't being used for the decode and only helping with other processing and possibly the render.

My source is 4:2:0 8-bit (no judging!)

Similar disclaimer -- the samples demoing the problem (above) -- we had 7 cameras rolling so no judging the footage please ;-) By design, there are moments we wouldn't be using that camera's take.

@VEGASDerek Any chance this could be prioritized in the dev/fix queue? I'm happy to help however you need me to. As you well know, "memory issues" have long been a sore spot, and thankfully your team has squished many over the years! This one seems to be... well.. still there.

Yes, and thus may use less VRAM so less likely to demonstrate such issue is my guess.

I wish we had a sample project with media that runs out of ram on 8GB VRAM systems we could share with the developers.

No judgment here, it's part of the deal with live events.

I wish we had a sample project with media that runs out of ram on 8GB VRAM systems we could share with the developers.

Is that a not-so-subtle hint? ;-)

When you say 8GB VRAM, are you talking about GPU RAM? My GPU isn't running out.. it's overall system RAM (demonstrated in the video)

I can find a BTS clip that's short that probably, played on loop, would demonstrate the problem. If you watched my vids above, you can't get much simpler. One clip on timeline with an FX. Play. It happens.

@RedRob-CandlelightProdctns Yes, for GPU Vram as in the other thread we had the same media tested with two AMD systems, one an 8GB GPU and one 12GB and the 12 wouldn't crash. Modest amounts of system ram would also be helpful as high amounts may mask the issue. If these are two separate issues with separate causes that's also important to know (VRAM vs overall memory issue).

We really do need something to replicate and get to the bottom of this. I can't directly help not having an AMD GPU but others have a variety of systems.

gpu stuck on 95% happens to me at least constantly with BCC light leaks OBSOLETE when rendering on 4k
also vram gets topped i have 64 system ram and 8 vram (rtx 3070)
i'm sure it can be an incompatibility with BCC obsolete plugins, but just saying in case someone is having the same issue

@RedRob-CandlelightProdctns to see the amount of GPU RAM is being used, you have to go in Task Manager to the Performance tab, then click on the GPU. It will tell you how much dedicated GPU memory is being used from the GPU, and there should be GPU Memory which would include system RAM as a virtual GPU memory.

I will test later on. I too experienced that issue with rising system RAM where it would go all the way to the my system limit, causing everything else to screech to a halt in the past. I think @Gid was able to get his system RAM to be used on a simple clip to get over 100GB on his massive 256GB computer, but he would have to comment on it.

I'll do the same test as you did.

And out of curiosity, does this happen with only NeatVideo applied?

@RedRob-CandlelightProdctns, after looking at your video, I am seeing the same thing with the continually rising of system RAM and a higher GPU RAM usage compared to legacy AVC decoding.

When using legacy AVC, system RAM usage levels out and doesn't increase. Without using legacy AVC, system RAM usage does NOT stop increasing. The more you play back your timeline, the RAM continues to grow.

With the clip I was working with (no effects), GPU RAM with legacy AVC was 5.1GB. With legacy AVC unchecked, 8.8GB. This is with only Vegas 22 open. No browser.

Unrelated to the RAM issue but related to the "Use Legacy Decoder" --

Does it not seem obvious to all that the non-legacy decoder has an issue which rears its ugly head every now and then, related to RAM and/or stability?

I'd love to squash this, and would be willing to enable debugging (if there was such a thing) to help try and track things down and get that improved.

Today on my main editing laptop, we have an 8-up (Multi-View), all clips have been proxied to be 1080p or 720p MP4s.. some or 60p some are 30p. Performance has been fine with full-frame playback, but as soon as we start editing, V22 was crashing -- every time. Checking Windows Performance Monitor, memory CPU and GPU all seemed within reasonable levels. Enabling Legacy Decode solved the problem and things are slightly slower on the edit, but very stable.

Thoughts?

Share a sample project with media so others can replicate and report the issue is one way to debug.

I was having some out of memory issues with my large project with 4K sources (mix AVC + HEVC) recently on my AMD CPU + AMD GPU system (full HW specs in my signature).

I've found that enabling "File I/O | Enable experimental HEVC decoding" dramatically decreases RAM usage but increases VRAM usage to max (I have not found any details on what this option actually does). I noted that it adds some delays during editing that are very unpleasant.
In my case I left the option turned OFF and added more RAM (16 => 32). Smooth editing is "the must" for me.

As for the problematic project -- I've just split it in parts for editing (planning to combine before final render by putting a few VP projects on timeline).

'non-legacy decoder', VP is pretty much at default settings, the 2 boxes are not ticked in File IO.

 I quite often see high RAM use & VP hanging on to it long after that project is closed, what the exact cause is is hard to say as I'm not watching my RAM all the time. Luckily I've got enough to ignore it most of the time but something weird just happened, so I've gotta share.

I went for a kip for a couple of hrs, the PC is quite often just put to sleep, so when I woke up I woke it up, made a cuppa & read your latest comment, looked at my RAM & it was 60GB 😒, I've got a system care app that cleans the RAM so I ran that which reduced it down to 45GB, that's still high. I looked at Vegas which was open, nothing on the timeline (also open is GIMP, Google with several tabs, a couple of Windows folders + Taskmanager) , . As I say I'm used to ignoring this high RAM but just as a test I opened a project I've been playing with recently to see if that was the cause for the high RAM. 🤔 On opening of that project RAM immediately dropped to 14GB, that's as low as it ever gets, even on Restart of the PC with nothing open it's about that amount 🤷‍♂️😂😂

Normally I close VP to clear the RAM, but I'm now wondering if opening this project is an easier solution as I don't lose the vfx preset cache that finds presets with the search box 😂

Share a sample project with media so others can replicate and report the issue is one way to debug.

The one which kept crashing today has 664 GB of media on the timeline. 8 cameras, 3 acts, 8 channels of 32-bit float audio. Sharing the project is easy.. sharing the media -- not happening. (would do so privately with Vegas dev team if it helped, but not the general public)

The crashing today didn't seem memory related (was not anywhere near exhausted), but was resolved all the same by enabling the legacy decoder.

@RedRob-CandlelightProdctns In that case I'd file a support request vs this public forum and offer them the files.

If you want to share things privately you are welcome to (feel free to message me) but I am not the dev team.

Here's an example of what I see often with just average editing, MP4's, MOV's .. It doesn't really matter & can arise from numerous different scenarios, as in this vid the simplest of scenarios, a couple of MP4's & a BCC Linear wipe. As I wrote above I'm not aware of high RAM usage without looking at Taskmanager but I do have Taskmanager open so I can have a quick look, esp with the way VP is behaving.