Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Half Res. Good vs. Full Res. Premium

Stephen, that's not what I see here. I just ran Andrae's test shot through the three options and half-good looked worst, with Full (post reduced 50%) and half-high looking best. Half high looked a bit sharper, Full (50%) looked more natural to my eyes, a bit smoother.

Half-high is very much like doing a full demosaic then a 50% scale reduction in one step, so no wonder it looks similar to doing full.

Graeme

Well when I test it "half good" not "high/premium" looks more like "full" in regards to sharpness.

Algorithms how do they come into the picture? Ive never been able to read a post about them. So I usually just use the sharpest.
 
I think what Andrae is responding to is the impression that a clearly noisier image appears sharper to the viewer.

Take a slightly out of focus image and add some grain to it and you'll be suprised to see that the image will appear sharper.

Basically I think the half-res debayer with the extra noise hides the fact that the 4K image isn't perfectly sharp or that there's some motion blur.
 
Hi, just for reference, the OLPF and Image Detail settings are adjustable in Resolve on a project level in the config screen and on a clip level in the color screen. Right click on the thumbnail to select 'Edit Red Decoder Settings'
Peter
 
One is fractions of resolution, the other is the balance between effort and speed in processing.
But I agree that "Premium" vs. "Good" is marketing speech – I'd call it "Best" and "Medium"…
 
RED Rocket only does a full resolution, full quality debayer. Any output you receive that is lower than that, is a down-scaled version of the full debayer. Rocket does not support OLPF compensation or chroma noise reduction as part of the debayer process.

I have not run Resolve without the Rocket, so can't comment on if there's any visual differences between using it and not using it. I'll do some tests in regards to that this coming week.

I'm going to be re-configuring my Resolve system, I'm putting the Quadro4000 GPU back in and adding a second Quadro4000. The GT160 plus GTX285 works fine, but it just doesn't cut it for uses outside of Resolve. At this point I can't dedicate a workstation to only Resolve as I really need to use CS5 and a few other apps that need a good GPU as the primary. It's just too much hassle with the GT160. And with the GTX285 swallowing both PCIe power leads, I couldn't run the Quadro as well. This system is just too full and I was having power issues when I tried to pull power from a drive connector.
 
ill try and ask about downsample algorithms/ filters again.

When using the RED rocket which always does a full debayer, which downsample algorithm does it use? (mitchel, CatMul, Gauss...etc)

and why would you chose a LESS sharp algorithm?
 
Hi, just for reference, the OLPF and Image Detail settings are adjustable in Resolve on a project level in the config screen and on a clip level in the color screen. Right click on the thumbnail to select 'Edit Red Decoder Settings'
Peter

Thanks Peter didn't notice those settings before.

I have not run Resolve without the Rocket, so can't comment on if there's any visual differences between using it and not using it. I'll do some tests in regards to that this coming week.

Interested in seeing your results. I really don't want to buy a RED Rocket card (again) unless quality demands it. Getting great realtime playback and renders out with just the CPUs from the Mac Pro 12 core at 1/2 res Good. Especially interested in RR renders potentially looking soft. I don't know what process Davinci Resolve uses at Full Res. Premium CPU debayer... but in my testing it looks softer than 1/2 Res. Good. So definitely would like to see an RR debayer in comparison to 1/2 Res. Good or Premium. Also 1/2 Res. Premium CPU debayer looks really really blurry.

Can the downsampling algorithm in Davinci Resolve be switched? I have a lot more frame grabs showing the softness issues:

1. http://farm6.static.flickr.com/5292/5434017447_7af6e1ac71_o.png "Half Res. Good"
2. http://farm6.static.flickr.com/5096/5434019617_d09a267d85_o.png "Half. Res. Premium"
3. http://farm6.static.flickr.com/5260/5434021487_17de281944_o.png "Full Res. Premium"

BTW... what I see in the Davinci Viewer... will that be what the footage looks like when rendered to disk?
 
Andrae - are you able to try it again with the sharpness on and see how it looks?

Graeme

http://farm6.static.flickr.com/5059/5435306014_1df70d3cc6_o.png (Full Res Premium - Radius Sharpening 40% in Davinci Resolve)

http://farm5.static.flickr.com/4104/5435304602_e6a4629172_o.png (Half Res Good - Radius Sharpening 40% in Davinci Resolve)

http://farm6.static.flickr.com/5293/5434692043_b6eeb19d1e_o.png (Half Res Good - No Sharpening in Davinci Resolve)

In my opinion if you are going out to a 2K or 1080p timeline... 1/2 Res Good debayer is where it's at.
 
How do the OLPF settings affect the image output versus using a Red Rocket card? Which is better to use for the best image quality (not performance)?
 
I'd avoid OLPF compensation at the start, even if it's mild.

Sharpening should be the very last step, since both it's technical effect and perception can be changed by your grading decisions.

So, it rather comes down to the question if you wan't to keep an older Mac and speed it up with the Rocket or spend you money on a 12-core.
 
So the gist is that OLPF is bad when used immediately after debayering? It should be disabled, right?
 
OLPF comp is not recommended on full demosaic. It is recommended on half-rez premium.

Graeme
 
Is it disabled for full demosaic by default? Or does it need to be turned off in REDCINE-X?
 
I guess all is in the eye of the beholder.

When I saw the first images before reading anything, I really found the "one to the right"pleasent, while the other looked quite steppy, less gradients in the lows and over all more brutal and video-harsh - even in the jpg translation.

Then I started reading...

We are essentially looking for different things in our images, I guess...

Good thing is both can have our cake...
 
We are essentially looking for different things in our images, I guess...

Good thing is both can have our cake...

Yea we are... without a scientific way to analyze the images we are all going by what pleases our own aesthetics. I personally don't like blurry images. Would you agree that the one to the right looks more blurred? Especially with the details in the eye.
 
Back
Top