Welcome to our community

Be a part of something great, join today!

Register Log in
  • 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 :)

DEB Comparison

Frank Hague said:
I think I'm doing something wrong because when I exported some recent clips at 1600 iso and compared them with and without the DEB box checked I could not really tell the difference, yet these two comparisons look awesomely different. Anybody know what I could be doing wrong?
Click to expand...

Deb shows its effect most when there is lots noise to begin with. It has no effect on MX footage (it just isn't made for that) only Dragon.
 
Frank Hague said:
I think I'm doing something wrong because when I exported some recent clips at 1600 iso and compared them with and without the DEB box checked I could not really tell the difference, yet these two comparisons look awesomely different. Anybody know what I could be doing wrong?
Click to expand...

The debayer has to be set to full or half Rez premium in order for deb to have an effect on it. If you make a 1/4 Rez debayer, then you will not see the rendered effects of DEB.

Battistella
 
David Battistella said:
The debayer has to be set to full or half Rez premium in order for deb to have an effect on it. If you make a 1/4 Rez debayer, then you will not see the rendered effects of DEB.

Battistella
Click to expand...

Thanks for your help, David. I actually exported the sample clips to 1080p though and that's where I could barely tell the difference. Does the playback resolution for the redcine monitor have an effect on my export?
 
Frank Hague said:
Thanks for your help, David. I actually exported the sample clips to 1080p though and that's where I could barely tell the difference. Does the playback resolution for the redcine monitor have an effect on my export?
Click to expand...

Nope, that won't have an effect on export.

A fun thing you can do is make a "Still Snapshot" and jump into Full Debayer and view it at 100%. Toggle D.E.B. on and off and you will be able to see the difference. Or just export a couple still snapshots and toggle between the two.

At 1080p, especially if your source material is 6K, it might be tricky to see fully. At 4K however you can see it.
 
Frank Hague said:
Thanks for your help, David. I actually exported the sample clips to 1080p though and that's where I could barely tell the difference. Does the playback resolution for the redcine monitor have an effect on my export?
Click to expand...


Phil is correct. remember 1080P is technically 1/4 res and in the case of Dragon footage less than 1/4 res. For any 1080P finish you would be less likely needing to engage DEB (depends on the footage and situation of course) For a 4K finish DEB is your friend.

Just double check also that your export preset is set to FULL debayer or 1/2 res premium debayer.

Cheers,
Battistella
 
Phil Holland said:
Nope, that won't have an effect on export.

A fun thing you can do is make a "Still Snapshot" and jump into Full Debayer and view it at 100%. Toggle D.E.B. on and off and you will be able to see the difference. Or just export a couple still snapshots and toggle between the two.

At 1080p, especially if your source material is 6K, it might be tricky to see fully. At 4K however you can see it.
Click to expand...

David Battistella said:
Phil is correct. remember 1080P is technically 1/4 res and in the case of Dragon footage less than 1/4 res. For any 1080P finish you would be less likely needing to engage DEB (depends on the footage and situation of course) For a 4K finish DEB is your friend.

Just double check also that your export preset is set to FULL debayer or 1/2 res premium debayer.

Cheers,
Battistella
Click to expand...

Does this mean that DEB is only applicable to stills then? (considering 99% of everything we shoot is viewed in 1080 at best)

I just can't wrap my head around how good these shots look in this thread in comparison to mine even at 1600
 
I think stills at 100% crop is the best way to see what is going on. You will see the effect of it less when you are scrunching all of those pixels down into a 1080 frame.

DEB has it's place and on material it is really needed you will see the effects in 1080. You see it much more when pixel peeping than when you are looking at full frames. It's really only affecting and correcting specific parts of the image.

Battistella
 
Frank Hague said:
Does this mean that DEB is only applicable to stills then? (considering 99% of everything we shoot is viewed in 1080 at best)

I just can't wrap my head around how good these shots look in this thread in comparison to mine even at 1600
Click to expand...

What David said.

Also, some of us (me) don't work in 1080p and haven't for a long, long time :)

If you have an .R3D to talk a look at I can scope out one of your shots. Just an .R3D snapshot.

Nearly everybody I see having difficulty with real deep noise is due to their shots being fairly underexposed and attempting to lift it up in post.
 
Point is basically that you need "exposure" even at 6400 ISO

And if the camera isn't heated up to blackshade temp, you'll have nasties happening all over.

Is that in the metadata somewhere BTW (the blackshade temp and if the cam was within range) ?
 
Phil Holland said:
What David said.

Also, some of us (me) don't work in 1080p and haven't for a long, long time :)

If you have an .R3D to talk a look at I can scope out one of your shots. Just an .R3D snapshot.

Nearly everybody I see having difficulty with real deep noise is due to their shots being fairly underexposed and attempting to lift it up in post.
Click to expand...

Those of us who work in advertising work exclusively in 1080p. The two features I shot were mastered in 2K (off 5K R3Ds). What kind of jobs are you doing exactly? They sound cool!
 
Phil Holland said:
What David said.

Also, some of us (me) don't work in 1080p and haven't for a long, long time :)

If you have an .R3D to talk a look at I can scope out one of your shots. Just an .R3D snapshot.

Nearly everybody I see having difficulty with real deep noise is due to their shots being fairly underexposed and attempting to lift it up in post.
Click to expand...

This is the one I tested from a recent Santana music video I shot, with the DEB and without:

https://www.dropbox.com/sh/r0g1gqyema4luwe/AAArsrbcRqgFac7iTINR92Gba?dl=0
 
Just taking a quick peek.

at 1600ISO it looks to me by the histogram that the shot is about 2 stops underexposed.
The shutter was at 144.4 Degrees, at 1/60 of a second when shooting 23.98 so you lost a bit of light there as well. If it had been at 1/48th 180 then maybe you would have been able to squeeze a bit more out of the shot. 10:1 compression might be biting you a bit here as well.

To process this shot you have to really push it to 3200 ISO, this means one would have to push a 2 stops under image one stop as a baseline grading point. Still remarkable the texture is not that bad.

The DEB correction mostly shows up in the shadow area's of "the man" on in left frame, but it does a pretty nice job in those lighting conditions.

little_test.jpg
 
Last edited:
David Battistella said:
Just taking a quick peek.

at 1600ISO it looks to me by the histogram that the shot is about 2 stops underexposed.
The shutter was at 144.4 Degrees, at 1/60 of a second when shooting 23.98 so you lost a bit of light there as well. If it had been at 1/48th 180 then maybe you would have been able to squeeze a bit more out of the shot. 10:1 compression might be biting you a bit here as well.

To process this shot you have to really push it to 3200 ISO, this means one would have to push a 2 stops under image one stop as a baseline grading point. Still remarkable the texture is not that bad.

The DEB correction mostly shows up in the shadow area's of "the man" on in left frame, but it does a pretty nice job in those lighting conditions.

little_test.jpg
Click to expand...

I'm well aware the shot is underexposed, that was the DP's decision for whatever reason and not relevant to the question at hand because, even at your proposed two stops up of 6400, I still can't see a discernible difference between the image with and without DEB (but I admittedly don't have the best eye for these things). Here's the two images at 6400-full res: https://www.dropbox.com/sh/c4uc8gyo7nb48nr/AADMBghb__qJeY76UgUS0rSLa?dl=0

Is there another step that I'm missing? How did you guys get those images at the beginning of the thread to look so clean at higher ISOs? Was there additional NR done?
 
Frank Hague said:
I'm well aware the shot is underexposed, that was the DP's decision for whatever reason and not relevant to the question at hand because, even at your proposed two stops up of 6400, I still can't see a discernible difference between the image with and without DEB (but I admittedly don't have the best eye for these things). Here's the two images at 6400-full res: https://www.dropbox.com/sh/c4uc8gyo7nb48nr/AADMBghb__qJeY76UgUS0rSLa?dl=0

Is there another step that I'm missing? How did you guys get those images at the beginning of the thread to look so clean at higher ISOs? Was there additional NR done?
Click to expand...


Its not a dis that I mention the underexposure, it's a contributing factor and I mention it for other people reading the
thread. You have to monitor full Rez to see it. Set your monitor to 100% and click the switch, you will see a difference. Since this shot leans toward the Reds, with all the red light, you might see it less where deb was designed to clean up red channel noise that was showing up in grey or grey green or certain parts of skin tine.

With Jans examples, he is always showing post processed images which could be a combination of add export, cc and some NR to clean things up a bit. As far as I have seen man has not posted r3d's and takes great care to present his images (which every person should).

DEB acts on very specific tonal range, if you want more NR than this then you need to run it through resolve or maybe neat video.

Battistella
 
Last edited:
Frank Hague said:
I can definitely notice the difference here when I turned the denoise all the way up as well. Still not as dramatic a difference as the others in this thread though unfortunately, but definitely looks better. Here's at 3200 without DR and DEB and with both: https://www.dropbox.com/sh/y5yhr0icv8csi41/AACPvOyQB_XxNOhlheHRy-6Da?dl=0
Click to expand...

Got home late tonight.

If you are referencing Jan's examples there are images that utilize DEB, ADD, and other Noise Reduction in there. An additional thing to note is you are looking at Jan's 6K FF versus 6K 5:4 2X ANA, which is a notable resolution difference when comparing them side by side.

I downloaded the image posted and ran DEB on it. It's doing what it's supposed to be doing. Cleaning up the slight "red speckles" across the tonal range:

phfx_santanaDEB.jpg
 
I dunno. If you are happy, stick with what you got. I can't describe why certain modes have effects on certain shots at 100%crop compared to the past versions. They are both very solid cameras with excellent sensors. Some images are just going to come out relatively the same.

I can cut in my 4K RED ONE footage from 7 or 8 years ago without much issue. That's sort of a first in the digital era.

Sometimes I think of myself as an investor in RED and the camera is my stock in the company. I've made a lot of good stuff with their technology and I like what they have done to disrupt the industry. I like to support that. One way of supporting that is to continue on the very generous upgrade path and be part of the obsolescence obsolete model. This means that I have had FOUR different camera from RED over the years, each one better than the last, (RED ONE, REDONE MX, EPIC X, DRAGON X) without having to buy a whole new camera but paying to upgrade it and then being able to stay current and competitive in the marketplace.

I like the model.

So do I blindly go into every upgrade? No. Am I loyal? They haven't let me down yet. But admittedly I don't look around much either because I have not seen anything on par with the resolution or REDcode RAW.

Battistella




sam karr said:
I was playing with some footages posted by Werner Jauch comparing scarlet dragon and scarlet mx (http://www.reduser.net/forum/showthread.php?117974-Any-Scarlet-Dragons-out-there/page10). And I noticed these red artifacts on the scarlet dragon image (polluting the blacks). I'm new to dragon technology, is it something already experienced? DEB has no effect.
scarlet dragon
on MX:
scarlet mx
Click to expand...
 
You must log in or register to reply here.
Back
Top