3D-Daz Daz3d Art - Show Us Your DazSkill

5.00 star(s) 12 Votes

Night Hacker

Forum Fanatic
Jul 3, 2021
4,416
21,835
56 second render! Genesis 8.1 Elly
View attachment 1689534
It took longer to clothe and pose than to render. It also took longer to load all the textures in VRAM than to render.

Please do not say denoising gives you the same quality as a full render, that simply is not true. It may be acceptable quality for posting on F95 or for relatively low resolution images but denoising is a destructive process and removes information from an image.

Fotsirk, save your pennies and when you can afford it get a better video card, it will make a world of difference. There are a lot of tricks you can learn to speed up render times without resorting to cutting off renders with a lot of noise thinking you can save it with a denoiser.
I rendered the following to images using a GTX1050TI with only 4GB of VRAM (not the best card for this). One of these images took me 45mins using standard settings, no denoising. The other image only took me 5 mins using denoising. There's is ABSOLUTELY a HUGE difference! You cannot tell the difference between the two at all, yet the denoised version was MUCH MUCH faster without any quality loss. This is especially important on a low end video card like mine as you cannot denoise at all when you run out of VRAM and that makes renders MUCH MUCH slower without denoising on low end cards (or any card for that matter)...

Now which of these images was 5min and denoised, and which one took 45 mins? Look close, you may be surprised at the answer...
A:
1700806_Wanda_at_the_Beach_-_45m.png

B:
1700807_Wanda_at_the_Beach_-5m.png
 
Last edited:

Deleted member 2741424

Well-Known Member
Aug 17, 2020
1,784
26,533
I rendered the following to images using a GTX1050TI with only 4GB of VRAM (not the best card for this). One of these images took me 45mins using standard settings, no denoising. The other image only took me 5 mins using denoising. There's is ABSOLUTELY a HUGE difference! You cannot tell the difference between the two at all, yet the denoised version was MUCH MUCH faster without any quality loss. This is especially important on a low end video card like mine as you cannot denoise at all when you run out of VRAM and that makes renders MUCH MUCH slower without denoising on low end cards (or any card for that matter)...

Now which of these images was 5min and denoised, and which one took 45 mins? Look close, you may be surprised at the answer...
A:
View attachment 1689570

B:
View attachment 1689571
I may be old and half blind, but I'm not seeing a difference.
 
  • Like
Reactions: Night Hacker

Fotsirk

Newbie
Jan 2, 2019
42
250
I rendered the following to images using a GTX1050TI with only 4GB of VRAM (not the best card for this). One of these images took me 45mins using standard settings, no denoising. The other image only took me 5 mins using denoising. There's is ABSOLUTELY a HUGE difference! You cannot tell the difference between the two at all, yet the denoised version was MUCH MUCH faster without any quality loss. This is especially important on a low end video card like mine as you cannot denoise at all when you run out of VRAM and that makes renders MUCH MUCH slower without denoising on low end cards (or any card for that matter)...

Now which of these images was 5min and denoised, and which one took 45 mins? Look close, you may be surprised at the answer...
A:
View attachment 1689570

B:
View attachment 1689571
Picture A was rendered in 5mins and denoised. You can tell when you zoom in that image A has lost some quality compared to Picture B (it's blurrier and has lost some details). Without zooming in, however, you are correct in that they are almost identical and certainly acceptable for posting here.

EDIT - I was mistaken! Picture B is actually 45 mins. So, as Night Hacker says, it really is hard to tell sometimes when something is denoised or not lol!

I appreciate your trying to help me out. I am actually using denoiser for now, but I personally think it's better to not use it if possible. And the only way to not use it is to have better hardware. I actually have a powerful desktop that I use for video editing and even started on Blender work. BUT, it has an AMD card (as a long time NVIDIA guy, I wanted to try out the competition), and unfortunately AMD GPUs are not compatible at all with Daz. Ugh.

So, all of my Daz rendering is currently being done on my laptop - which has a GTX 1070.

Currently, I do cut off renders when they reach around 90% progress, and then put it through the intel denoiser. It does depend on the scene, since some scenes are obviously way more complex and thus take longer. But there is a loss of quality when denoising because, as m4dsk1llz said, it is a destructive process. Even if it might not be noticeable at first glance. Though I think how noticeable it might be does depend on the actual model/scene/lighting too.
 
Last edited:

DitaVonTease

Active Member
Jul 25, 2021
582
1,189
Honestly I have better things to do than pixel peep your images. I even said that low resolution images and forum posted images are acceptable with denoiser, but rerender these at 4K/6K/8K and cut off at 5 minutes then denoise. Or add more than a single character and single HDRI background and do the same test. That would really show you what a denoiser can and can't do. Also I would never advise someone not to spend their money on a better GPU if that is what they want to do. I didn't ask you to dump your GTX1050Ti, nor would I ever. My biggest objection to your post is not your artwork (which I like, by the way) it is your insistence on saying "without any quality loss."

Heck when I started with DAZ I only had an AMD RX570 so you could image that nearly every render was an overnight job, especially since it was CPU only. Every render was 1280P at the most. It wasn't until I had a 1080Ti that I started doing HD renders, and then it wasn't until a second 1080Ti, did I do any 4K renders.

I am also not saying everyone should go run out and buy a RTX3090/3080/3070, because that is ridiculous if you aren't making any money to recoup your investment, or you have some discretionary income, but I don't discourage others from buying one either.

In the end I am not trying to pick a fight with you, like I said, I like you images and I am still learning DAZ, so a lot of my renders might be faster but that doesn't mean mine are any good.

I totally agree with you, my problem with the ''Great God Denoiser Brigade'', or is it the ''Peoples Front for Denoiser'', is the total 'THOU MUST DO'' attitude, as if it's the only way to render, & that doing something different make you some sort of ''Abomination unto Nuggan''. I've tried several things over the years, including going back to the abacus, oddly that didn't work. What does work is finding what works for you & your machine, at the end of the day it's the finished image that matters. Using the ''GGDB/PFfD'' option is like the average poster paint mess of a 3 year old, compared to the 'Mona Lisa' of an Artist who has the patience to learn their tools, even if they don't use every trick they know to 'improve' the image, because it's often not needed. Often the two most important tools available, are the least used, the Human Brain & Patience. At the end of the day is it better to have two pictures your really happy with, that could be printed to A0 size without a loss of quality, or 100 renders that barely look good on a computer screen, & are only good enough to print up as small postcards.......I'll leave you all the time you need to work out the best way for you, & to the ''Great God Denoiser Brigade/Peoples Front for Denoiser'', remember what happened to all the 'Flat Earthers'.....
 

44Tom

Active Member
Oct 19, 2018
700
807
Picture A was rendered in 5mins and denoised. You can tell when you zoom in that image A has lost some quality compared to Picture B (it's blurrier and has lost some details). Without zooming in, however, you are correct in that they are almost identical and certainly acceptable for posting here.
I think that Night Hacker has made his point. Picture A was NOT denoised. The title of the picture was "Wanda_at_the_Beach_-_45m.png". So, it took 45 min. to render. Picture B title was "Wanda_at_the_Beach_-5m.png". My assumption was that Picture B WAS denoised, thus took 5 minutes to render, and picture A, NOT denoised and took 45 min. to render.
 

atheran

Member
Feb 3, 2020
355
2,753
This time, I didn't even bother trying to upload the 4k.
Also, I might want to stay away from DoF blur during render. Takes forever and a half and I'm not convinced it's as good as Nik's collection's filters that are lightning fast while still trying to emulate a proper camera/lens combo.


Waiting_low.png

But I DO need to find a better skin/freckles combination. I really don't like how the skin looks here and the freckles? That's just like someone spray painted them on her.
 
5.00 star(s) 12 Votes