Blackmagic Forum

So I was editing this movie and I had to relocate the clips from my harddrive to an external SSD.
But the project with the relocated clips now has all kinds of weird artifacts all over the screen. In the timeline, viewer, media pool, everywhere...

I've tried rebooting my pc and Davinci resolve.

Anyone know what caused this? And how to fix this?

Best,
Thijs

Maybe just coincidence, but this looks like graphic card/driver? issues.

What happens if you redo the cache?
And I assume these artifacts are only visible in DVR regarding video playback?
what happens if you create a new timeline, and add some of the videos in it. Is the result the same?

just thinking out loud here

redoing the cache didnt help, but good idea!
It's the same for every project, timeline and menu in DVR. It also occurs when not playing the video. Very strange..

Wow. How strange. Bad memory in the PC?

Looks like a hardware issue, could be drivers too.

Dan Anon wrote:Wow. How strange. Bad memory in the PC?

I have 32gb memory I believe. I'm no expert but I believe that's plenty for DVR.

Never had issues with drivers or hardware before, especially now since I always edit off of my external SSD.

First guess is failing hardware. Are temperatures OK?

Thijs van Daalen wrote:So I was editing this movie and I had to relocate the clips from my harddrive to an external SSD.
But the project with the relocated clips now has all kinds of weird artifacts all over the screen.

Here is your answer. Either your clips got corrupted during relocation or your external SSD / connection to it is not working correctly. If you still have original clips try comparing them to one on SSD by using something like WinMerge. Also try by using internal drive for clips again. That could give you an idea of what is happening.

4EvrYng wrote:Here is your answer. Either your clips got corrupted during relocation or your external SSD / connection to it is not working correctly. If you still have original clips try comparing them to one on SSD by using something like WinMerge. Also try by using internal drive for clips again. That could give you an idea of what is happening.

Thanks for the reply, the thingies did occur after the relocating but it wasn't just that project, when the artifacts start it translates over to every single project in DR. Even the loading menu.
The strange thing is, it was gone for a couple hours and then it came back.
Sadly do not have the original clips, I ctrl+x them to the SSD.

The temperatures of the drive were okay.

I’d still suspect a hardware issue.

could be, I'll run a few checks to see if there's anything unusual about the disk

Thijs van Daalen wrote:... but it wasn't just that project, when the artifacts start it translates over to every single project in DR. Even the loading menu.
The strange thing is, it was gone for a couple hours and then it came back.

I have seen attempts to display corrupted / incorrectly read file trigger daisy chain artifacts until system is restarted. It is rare but I have seen it happen so you still need to check is it external SSD that affected you because move of clips to it coincides with start of your problems. If you want to check stability of your system (which, I agree with others, you should) I heartily recommend combination of USB bootable MemTest86 and OCCT.

I’ve seen this screen corruption a few times over the years when a video card is experiencing issues or has been overclocked too much. Check the temps of the GPU’s memory & cores. Try rolling back the driver a few versions. Swap out with an older GPU if you have one laying around.

It looks like corrupted VRAM. Strange rebooting didn't solve it.

If your temps etc are fine I'd think about rolling back your driver to an older one. Or updating to the latest if you're not on the latest.

If you have a second computer you can play the SSD files on that and see what happens. But I think it would be strange for bad files to corrupt the whole screen.

Most here suggest a hardware issue. I concur.

You might get lucky by re-seating any cards (memory/graphic) and reconnecting/re-tightening cables.

Update: the strange artifacts are gone after two days of them popping up everywhere. Don't know what caused it to appear and also disappear.

My VRAM is 8gb which isn't a lot for effect-heavy edits. Maybe that could be the culprit. Heat was okay.
Drivers are up-to-date.
I did have troubles with my SSD card ever since buying it. Every now and then it peaks to 100% while usually sitting at around 20%. When that happens DVR freezes for 3 seconds and works fine after. Maybe the artifacts could be a product of the SSD peaking.

I can't see how the SSD could be the culprit.

Unless you mean the CPU also peaked at 100% what you're seeing is just the video part of the computer.

Such a strange series of artifact. My own experience with bad memory is inconsistant behaviour and odd shutdowns. I had a stick of bad 16gb a couple of weeks ago. Kind of a pain to test for. Memtest86. You can test two sticks at a time then narrow it down to 1 stick. This is with 4 sticks but I doubt this is the problem here.

My wild guess in the dark is a poorly seated GPU which was knocked when installing the SSD drive. If external then doesn't make sense. Dust can cause minor havoc on the gold pins on memory and PCIe slots / and the opposite of slots whatever they are called PCIe sticks?

Only do it if you are confident in grounding yourself and the PC a the same time. And do not wipe memory sticks. But yes I would reseat everything (if you can) and/or check your temps on that graphics card too.

Thijs van Daalen wrote:Update: the strange artifacts are gone after two days of them popping up everywhere. Don't know what caused it to appear and also disappear.
...
I did have troubles with my SSD card ever since buying it. Every now and then it peaks to 100% while usually sitting at around 20%. When that happens DVR freezes for 3 seconds and works fine after.

That is odd. Here are few thoughts in no particular order:

1. Things starting and going away later could be triggered by behind the screen update processes. Check your Windows update logs and installation logs to see if that happened. Also, Microsoft tends to push their security definition updates on regular basis and I have seen it more than once impact things.

2. If you are using 3rd party anti-virus that is another potential culprit. Check its logs.

3. What do you mean "SSD card"? What exactly are you using for your external disk?

4. Disk at idling shouldn't be at 20%, it should be at 0%, unless you have behind the scenes processes that are using it, like indexing, anti-virus scanning, swapping, etc. Take a look at task manager which processes are using your disk.

5. SSD disk appearing "freezing", especially after lots of activity, can indicate disk is emptying its buffer. Also, Windows once in a while performs a TRIM on SSDs in order to restore their performance. Which SSD you are exactly using?

6. Last, but not least, fast boot option is known to cause instability issues. Based on your description that doesn't seem to be the case here but I consider disabling it great pro-active practice.

I have also seen this in version 20.0B3 - I have never had this before. The entire UI glitched out. Even when there was no load on the GPU or no video was playing. Simple mouse movements triggered it. I can only assume it has something to do with a fairly complex Fusion Comp I was working on before it happened. When I restarted DaVinci, the problem was gone, and I can't reproduce it.
I ran some stress tests with my GPU and it works fine - also other demanding software like After Effects and Blender do not trigger this. I'm pretty sure this isn't a hardware issue.

Thanks for the reply's and ideas

There was indeed a behind the screens windows update around the first time I saw those glitches. Second day the artifacts popped up there was no behind the scenes update going on.
SSD card was me confusing SD card and SSD, sorry. I use a reformatted SanDisk PRO G-40 external SSD with a lightning 3 cable.
Disk is sitting at 0% normally but 20% while editing.
The freezing seems totally randomized though, I have found no trigger yet. Sometimes it freezes twice in 5 minutes and sometimes it can go all day without "peaking" to 100, and thus freezing.

Luckily the strange glitches have not made a comeback yet but I'm very curious to see what caused them.

Last week my laptop one day started pegging the CPU at 100%. When I opened task manager the system interrupts had grabbed 100%. It was so bad the clock couldn't keep up . After two reboots it finally cleared.

This was after a Windows update

Thijs van Daalen wrote:I use a reformatted SanDisk PRO G-40 external SSD with a lightning 3 cable.
Disk is sitting at 0% normally but 20% while editing.

SanDisk PRO G-40 should be more than fast enough but I don't know what you mean by "reformatted". Also, do you mean Thunderbolt 3 cable? If yes beware that Thunderbolt on Windows machines can be the issue. If issues reoccur try using local disk and if that resolves it try USB-C _with high quality cable_.

4EvrYng wrote:SanDisk PRO G-40 should be more than fast enough but I don't know what you mean by "reformatted". Also, do you mean Thunderbolt 3 cable? If yes beware that Thunderbolt on Windows machines can be the issue. If issues reoccur try using local disk and if that resolves it try USB-C _with high quality cable_.

Hi,
Yes I meant thunderbolt 3 cable, I was sleepy and started confusing some things .
Is the apple iPhone usb-c cable high quality enough to replace the thunderbolt 3 cable?

The external SSD is reformatted as in, the PRO-G40 is preformatted in APFS for MacOS. You can also format the SSD for use with Windows, but you need to reformat it. Maybe this has nothing to do with the sudden peaking but I thought it could be useful information.

You never know with USB cables. If you own it test it.

If you want to be safe get a TB4 cable. An USB 4 cable should also be safe.

Or if you want to check specs make sure whatever you're buying is rated for fast data transfer. Many of the USB C cable being sold can barely hit USB 2 speeds Nobody notices because they're only being used for charging.

Thijs van Daalen wrote:Is the apple iPhone usb-c cable high quality enough to replace the thunderbolt 3 cable?

I thread very carefully when it comes to Thunderbolt vs. USB-C connections as it is so easy to make a mistake by not knowing details of the gear person is using so I don't want to answer with simple 'yes/no'. Instead, if issues reoccur, I encourage you to do your own research for your particular gear and try to answer question is it external SSD that is triggering issue and if yes what about it (type/port/cable) could be causing it.

Thijs van Daalen wrote:The external SSD is reformatted as in, the PRO-G40 is preformatted in APFS for MacOS. You can also format the SSD for use with Windows, but you need to reformat it.

Got it, thank you for the clarification.

The artifacts are very unlikely to be related to media or some I/O issue. They occur on the UI part of Resolve, not just the program monitor.

The tiny rectangles appear to be about 12 x 18 pixels, regardless of the audio waveform shape, the clip image, or the zoom level of the timeline. They stop at tile boundaries, which implies a fault in the GPU memory path (driver or VRAM), not inside Resolve’s UI rendering logic.

A 12 × 18-pixel rectangle at first seems odd since NVIDIA drivers store data in 16 × 16 or 32 x 32-pixel tiles. However, if the display uses scaling for a Dell 4k panel or similar, a 16 x 16 pixel tile can become 12 x 18 after scaling.

The artifacts probably have a consistent shape and pattern because the GPU's memory controller accesses VRAM in specific block patterns. It is conceivable that underclocking the GPU VRAM might help.

This is consistent with a hardware or driver issue. In such cases, there is no software code in Resolve that can be changed to fix this.

joema4 wrote:The artifacts are very unlikely to be related to media or some I/O issue. They occur on the UI part of Resolve, not just the program monitor.

The tiny rectangles appear to be about 12 x 18 pixels, regardless of the audio waveform shape, the clip image, or the zoom level of the timeline. They stop at tile boundaries, which implies a fault in the GPU memory path (driver or VRAM), not inside Resolve’s UI rendering logic.

A 12 × 18-pixel rectangle at first seems odd since NVIDIA drivers store data in 16 × 16 or 32 x 32-pixel tiles. However, if the display uses scaling for a Dell 4k panel or similar, a 16 x 16 pixel tile can become 12 x 18 after scaling.

The artifacts probably have a consistent shape and pattern because the GPU's memory controller accesses VRAM in specific block patterns. It is conceivable that underclocking the GPU VRAM might help.

This is consistent with a hardware or driver issue. In such cases, there is no software code in Resolve that can be changed to fix this.

I agree that practically 100% of the time such symptoms indicate video card hardware/driver issue and that is why I agreed that warrants thorough system test. However, I couldn't / can't discount possibility of I/O issues / corrupted media because:

a) Start of his issues coincides with moving of media to external SSD, issues were persistent, and now they mysteriously went away. If there were no changes to the system after issues started and they were caused by faulty hardware/drivers they would not have went away on their own.

b) My current system has been put through extreme stress tests when it was built, it has passed all of them with flying colors, and has been overall very stable, yet I have had few occurrences of whole screen artifacts, system freezes and BSODs when playing certain videos. When that happened I repeated stress tests and they passed with flying colors again. Rolling back drivers didn't help either. Only thing I could put finger on is that it would happen when using hardware accelerated decoding with them while others didn't have issue being used in same way. I don't know what was "special" about few culprits but I remember it as an example that under "right" set of circumstances playing certain video file in the certain manner can throw whole system a curved ball in turn raising a question mark about that file.

Hey
I had this problem, and also anything using the Nvidia RTX Scaling inside DaVinci rendered with errors, flickers and artifacts. But only happened inside DaVinci, nowhere else.

EVGA RTX 3080 Ti XC3 ULTRA GAMING

I fixed it by going back to october 2024 Studio drivers (had to try going back month by month until the artifacts stopped).
Nvidia Studio Drivers 565.90 DHC

Now I can use RTX scaling without problems and I don't have artifacts on the UI anymore.

I tried multiple other things, but I could not repeat the errors outside DaVinci so I had to rule out hardware problems. Stress tests, memory tests, bios updates, etc, nothing helped.

Hope this helps anyone else having issues like this.

I'm very glad I didnt have to replace my card

Gonzalo Moreno wrote:I had this problem, and also anything using the Nvidia RTX Scaling inside DaVinci rendered with errors, flickers and artifacts. But only happened inside DaVinci, nowhere else.
...
I fixed it by going back to october 2024 Studio drivers ... Nvidia Studio Drivers 565.90 DHC

Thank you for sharing this!

Problems came back after a few days.

Now trying fixes found in this post:
https://forum.blackmagicdesign.com/viewtopic.php?t=202729&p=1059029

problem seems fixed, even in latest drivers, but can't tell if it's because a fix in the 10bit displays settings or the vsync fix (i changed from 10bit to 8 bit in nvidia control panel)

update: UI doesn't flicker anymore, but rendering artifacts still appear sometimes.
I was monitoring GPU VRAM usage: I had almost nothing free during the render, but at some point, 7GB where suddenly available and glitches in the render started happening.

For what it's worth, I had a similar problem once in LibreOffice which turned out to be a problem with the "SKIA" engine (Which draws the GUI in LibreOffice) and the Nvidia driver.

It looked exactly like a typical hardware-fault, but was in fact driver-related (I've experienced 2 GPUs that ended up failing over the past 25 years, and was afraid I was about to lose a 3rd GPU when it turned out to be related to the SKIA software)

A later update to the Nvidia driver made the problem go away.

Skia is an open-source library which apparently is used widely in a lot of software. I don't know if it's used in Davinci though.

But perhaps they're using parts of it, or code that behaves like it, which may then run into problems with certain Nvidia-drivers? (If yes, then you can only wait the problem out, until either Davinci is updated or the Nvidia driver is updated)

Nvidia has had a bunch of trouble lately with drivers, since the RTX50 GPU-series release, so I wouldn't be surprised if this was driver-related.

I suppose an update to the SKIA software could also fix it, but I'm not sure which order these things go in (An internet-search says Skia is made by Google and uses the GPU in Vulkan and OpenGL)

Long story short; don't sell your GPU just yet, as it may be fine