Help: Vulcan error when I try to launch vkdt.exe

Yogansh_Bhatt · December 9, 2025, 12:57pm

I couldn’t fix it with the solution provided on the site/github by making a config.rc file with “intqvk/device_id:0” in “C:\Users\yogan.config\vkdt”

i tried almost all the permutations and combinations till even claude, chatgpt and my braincells gave up. New to everything but I wanted to use AgX-Emulsion in vkdt because the one from the site is great and slow

@hanatos I would love to tell you how much I have loved Darktable and I know I would love filmsim the moment I can get it to work!

C:\Program Files\vkdt>vkdt -d qvk
[gui] vkdt 0.9.99-989-gbc9955aa (c) 2020--2025 johannes hanika
[gui] glfwGetVersionString() : 3.4.0 Win32 WGL Null EGL OSMesa MinGW-w64
[gui] monitor [0] Generic PnP Monitor at 0 0
[gui] vk extension required by GLFW:
[gui]   VK_KHR_surface
[gui]   VK_KHR_win32_surface
[qvk] dev 0: vendorid 0x10de
[qvk] dev 0: NVIDIA GeForce RTX 4050 Laptop GPU
[qvk] max number of allocations -1
[qvk] max image allocation size 32768 x 32768
[qvk] max uniform buffer range 65536
[qvk] dev 1: vendorid 0x10de
[qvk] dev 1: Microsoft Direct3D12 (NVIDIA GeForce RTX 4050 Laptop GPU)
[qvk] max number of allocations 4096
[qvk] max image allocation size 16384 x 16384
[qvk] max uniform buffer range 65536
[qvk] dev 2: vendorid 0x1414
[qvk] dev 2: Microsoft Direct3D12 (Microsoft Basic Render Driver)
[qvk] max number of allocations 4096
[qvk] max image allocation size 16384 x 16384
[qvk] max uniform buffer range 65536
[qvk] num queue families: 2
[qvk] picked device 1 with ray tracing and without float atomics and with coopmat support
[qvk] validation layer: loader_validate_device_extensions: Device extension VK_KHR_deferred_host_operations not supported by selected physical device or enabled layers.
[qvk] validation layer: vkCreateDevice: Failed to validate extensions in list
[qvk] error VK_ERROR_EXTENSION_NOT_PRESENT executing vkCreateDevice(qvk.physical_device, &dev_create_info, NULL, &qvk.device)!
[ERR] init vulkan failed
[ERR] failed to init gui/swapchain

Yogansh_Bhatt · December 9, 2025, 3:06pm

I rookie mistake. The directory it looks for is not user/.config/vkdt/config/ but user/vkdt/config/ on windows. Well it took forever but here’s to

and I think adding a specifc “For windows users make it in this directory…” section would save some newbies like me.

hanatos · December 10, 2025, 10:39am

ah, glad you figured it out. i really have no idea about windows. might have to look into how they organise files (do they?).

Yogansh_Bhatt · December 10, 2025, 10:59am

It was a pain to waste so much time on something so silly but hey it works and oh dear it works so freaking well! the Napari method just for Agx-emulsion was SLOW even with the latest commits…with vkdt I can literally run mcraws through it with no sweat even on my less than mediocre pc.

I do have one more request though, Some mcraws had wrong levels (initially I thought the ones I shot on RAW_SENSOR were the only ones but soon I found the same issue on a RAW10 file.)

If you could add a functionality to be able to change levels on the fly it would not only be able to fix this issue but also enable me to use DCG videos with 12bit or 14bit levels.

Rest is beautiful. I thought I will be overwhelmed by the complexity but I am overwhelmed by the intuitiveness AND gorgeous results!

I may have other requests but this is the most I can ask of you right now @hanatos

hanatos · December 10, 2025, 11:33am

hm, no root on my phone do you have a pointer to an example DCG file so i can see/reproduce the issue with levels? not sure i understood it from the text.

hanatos · December 10, 2025, 12:24pm

hm. works for me (mi13u), both RAW_SENSOR and RAW10 (i don’t have DCG). can you share sample files somewhere?

Yogansh_Bhatt · December 10, 2025, 12:25pm

Thank You for the quick response!
I shot: one RAW_SENSOR mcraw, one RAW_SENSOR but with DCG enabled, and one RAW10. To my surprise the DCG actually has correct levels (it needs to have 4095/256 for my device) but both the other tests have wrong levels.
https://drive.google.com/drive/folders/17V9dPAaB_F55uFnz65iustu2RFzIK4oF?usp=sharing
here are the mcraws

Hope that is helpful. And Idk why I have weird issue with links that I had to repost again and again huh @hanatos

hanatos · December 10, 2025, 12:49pm

interesting. the metadata says 256/4095, but on these files 64/1024 looks correct. new version will expose black/white as parameters in i-mcraw

Yogansh_Bhatt · December 10, 2025, 12:49pm

Also you could use dynamic black / white level in per frame metadata as its the one used in mc. static data levels are still included but only matter for mlv usage for now.
I am sure you can make it work. (also noticed when I open a timelapse folder it crashes vkdt, Maybe mcraw in general would benefit a ton from your precious time

hanatos · December 10, 2025, 1:02pm

ugh really? i don’t want to parse json for every frame. that sounds slow.

Yogansh_Bhatt · December 10, 2025, 1:53pm

I talked to sebastian and other devs,
: if he wants to include shading map in processing i guess there is no way around that anyway
:wl cannot be 1024. that would not fit in 10b.
last comment also makes me wonder how even 1023 whitelevel be correct for the DCG when it is 12bit and with exiftool requires 4095/256 to be correct. (unless it is an mc side issue which they are gonna fix so that every device with 10b, 12b or 14b levels has correct levels in post)

Also I didn’t use any json or anything so this is what you get straight out of it.

I thought DCG would have incorrect levels (assuming it would have 10b levels) but here instead 10b footage has incorrect levels no matter what mcraw container I use (10b or 16b)

Thanks for being quick and helpful so far @hanatos

Yogansh_Bhatt · December 11, 2025, 9:17am

thank you very much!