VMs won’t do for long, because you won’t have proper acceleration as it’s required by gfx apps like Lightroom. Sure, they’ll work, but you’ll experience slowdowns. You can run accelerated VMs, but I find them buggy.
If you’re going to dual boot, you should install Linux on a separate DRIVE, not just a partition, and install the bootloader on that second drive. You force Linux to do that by disabling in the BIOS the Windows drive first, before installation. Then, you re-enable it again. Then you can choose what to boot at using F12 during boot time. If you put them on the same drive, Windows will eventually overwrite the bootloader.
GPU pass through with VFIO is literally how services like GeForce Now, PSN Streaming work. It’s not too buggy if your system is set up properly and it’s far better than dual boot.
Yes, it does not have ML denoise, but there are very good reasons why you don’t want to have that in your raw pipeline. Sure, after raw development is fine, but denoise in a raw pipeline needs to maximise the signal-to-noise ratio. Machine learning denoising would introduce hallucinations, which are not real signal, and that’s why it’s best kept out of raw files.
Well, yes, some specific camera support features are missing, such as Fujifilm look-up tables, it still is the best raw editor I have used in my entire life and I can highly recommend it.
Are you sure the pipeline works that way? I know what you mean and it would seem like a huge oversight on their part to apply the denoise before the other edits. I would assume that increasing exposure, for example, would ignore the applied denoise rather than apply it overtop? If that wording makes any sense
Regardless I’ve used it to rescue photos I’ve taken on a nexus 4 over a decade ago, making them look like proper photos, and I find the feature so useful that it’s irreplaceable to me
The other feature is the AI content aware fill. In darktable, can you circle a piece of garbage on the ground and effortlessly remove it? Or do you have to do some manual clone stamping etc etc?
In a recent instance, a friend requested an album cover from a 3x2 image that I needed to expand to be 1x1. Can you tell that the left and right edges of this are not real? I don’t think I would have been skilled enough to pull this off without AI tools. https://f4.bcbits.com/img/a1356058193_10.jpg
VMs won’t do for long, because you won’t have proper acceleration as it’s required by gfx apps like Lightroom. Sure, they’ll work, but you’ll experience slowdowns. You can run accelerated VMs, but I find them buggy.
If you’re going to dual boot, you should install Linux on a separate DRIVE, not just a partition, and install the bootloader on that second drive. You force Linux to do that by disabling in the BIOS the Windows drive first, before installation. Then, you re-enable it again. Then you can choose what to boot at using F12 during boot time. If you put them on the same drive, Windows will eventually overwrite the bootloader.
The ideal thing is to actually move to Darktable. https://mathiashueber.com/migrate-from-lightroom-to-open-source-alternative/
GPU pass through with VFIO is literally how services like GeForce Now, PSN Streaming work. It’s not too buggy if your system is set up properly and it’s far better than dual boot.
Darktable doesn’t have AI denoise, and also doesn’t have camera profiles for fuji RAF files, just off the top of my head.
Yes, it does not have ML denoise, but there are very good reasons why you don’t want to have that in your raw pipeline. Sure, after raw development is fine, but denoise in a raw pipeline needs to maximise the signal-to-noise ratio. Machine learning denoising would introduce hallucinations, which are not real signal, and that’s why it’s best kept out of raw files.
Well, yes, some specific camera support features are missing, such as Fujifilm look-up tables, it still is the best raw editor I have used in my entire life and I can highly recommend it.
I will try it based on your second paragraph.
Are you sure the pipeline works that way? I know what you mean and it would seem like a huge oversight on their part to apply the denoise before the other edits. I would assume that increasing exposure, for example, would ignore the applied denoise rather than apply it overtop? If that wording makes any sense
Regardless I’ve used it to rescue photos I’ve taken on a nexus 4 over a decade ago, making them look like proper photos, and I find the feature so useful that it’s irreplaceable to me
The other feature is the AI content aware fill. In darktable, can you circle a piece of garbage on the ground and effortlessly remove it? Or do you have to do some manual clone stamping etc etc?
In a recent instance, a friend requested an album cover from a 3x2 image that I needed to expand to be 1x1. Can you tell that the left and right edges of this are not real? I don’t think I would have been skilled enough to pull this off without AI tools. https://f4.bcbits.com/img/a1356058193_10.jpg
password protect your bios. that’s the only way i found for windows not to mess it up on my machine, seriously.
Single gpu passthrough vm works flawlessly if you can take the time to set it up