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 :)

Scarlet Dragon ISO Test - LLO OLPF

A

Akin A

Well-known member
Joined
May 3, 2014
Messages
868
Reaction score
0
Points
16
http://youtu.be/ENOQTvKXRs0

This is an ISO test, using the RED Scarlet Dragon with the Low Light Optimized OLPF.
Each clip is a separate shot, exposed for a different ISO, by closing down the lens aperture. Nothing else was changed.
(I removed the audio, but my lips are slating the ISO and talking about the key/fill lights)

Resolution: 5K WS (2.37:1) 5120x2160
- downscaled to a 1080p timeline. The zoomed area is a 5K crop
Lens: Sigma Art 50mm F/1.4
Shutter: 1/48s
REDCODE: 7:1
Color Space: DRAGONcolor
Gamma Curve: REDgamma4
Color Temp: 5500K
Tint: -15

NLE: Davinci Resolve
- Export: DNxHD 1080p 220/185/175 8-bit

Notes:
- I don't know the exact apertures used, because I'm using the Nikon G mount version of the Sigma Art 50/1.4, and a cheap Nikon to Canon mount adapter. I used RED's exposure tools to keep the exposure consistent as I closed the aperture down.
- The downscaling from 5K to 1080p, plus the compression of the export (DNxHD 220/185/175 8-bit), plus YouTube's compression, hides a lot of the noise that is visible in the .R3Ds.

Issues:
- I noticed that the tint got more green as a closed down the aperture. Has anybody else experienced that? I used the same white balance settings (that looked good at ISO 250) for every shot.

Let me know if my test has any flaws.
 
Great work Akin, but as you say, the YouTube compression is killing the noise.
 
Thanks a lot for your test.

I know now more than ever that my OLPF will be the LLO!!!
I hope it will arrive soon!


I wonder why you use REDGAMMA4 as it's IMHO a really crushed gamma curve.

I usually use RG3 and now DragonColor2 or RC4.

Don't you feel that RG4 isn't giving you a nice starting point?

Pat
 
David Battistella said:
Good reading...

http://www.reduser.net/forum/showthread.php?126194-Native-Dragon-ISO%85once-and-for-all!
Click to expand...
Yup, that thread inspired this test.

Martin Stevens said:
Great work Akin, but as you say, the YouTube compression is killing the noise.
Click to expand...
Thanks.
Nick Morrison said:
Pretty darn impressive. Everything seemed totally usable till 12k ISO!
Click to expand...
I just checked the R3Ds and I would probably use up to 4000. 2500 is where the noise started to show but wasn't a distraction to me up through 4000. After that, noise reduction (none used in this test), downscaling, compression and gamma curve would help a lot in making it look usable.
Patrick Tresch said:
Thanks a lot for your test.

I know now more than ever that my OLPF will be the LLO!!!
I hope it will arrive soon!


I wonder why you use REDGAMMA4 as it's IMHO a really crushed gamma curve.

I usually use RG3 and now DragonColor2 or RC4.

Don't you feel that RG4 isn't giving you a nice starting point?

Pat
Click to expand...
You're welcome. I only used RG4 because it's the latest. I'm still new to RED, so I'm experimenting to find what works for me. You're right though, it's crushing the shadows more than other curves. And Resolve doesn't have DC2 in the current version as of this test, so I went with DC.
 
Is it normal for Dragon to get more green as less light hits the sensor? Notice how there's more of a green tint as I closed down the aperture at higher ISO levels.
It's a very gradual change, but if you skip around in the video from beginning to middle to end, you'll notice it more.
 
Akin A said:
Is it normal for Dragon to get more green as less light hits the sensor? Notice how there's more of a green tint as I closed down the aperture at higher ISO levels.
It's a very gradual change, but if you skip around in the video from beginning to middle to end, you'll notice it more.
Click to expand...

What was your light source for this test? It might have a a bit of green in it which gets more pronounced as you dig into the noise.
 
Martin Stevens said:
What was your light source for this test? It might have a a bit of green in it which gets more pronounced as you dig into the noise.
Click to expand...
I used 2 soft boxes with cheap 5500K CFLs. This is my current kit: http://www.amazon.com/dp/B009NUBJGM
The bulbs do have a slight green tint and I usually fix that in post. Setting tint to -15 looked good at ISO 250. I didn't want to throw in more variables on the test by giving different tint values to different clips.
 
Akin A said:
Is it normal for Dragon to get more green as less light hits the sensor? Notice how there's more of a green tint as I closed down the aperture at higher ISO levels.
It's a very gradual change, but if you skip around in the video from beginning to middle to end, you'll notice it more.
Click to expand...

I'm curious to hear what you find after DC2 comes around in its final form. I sometimes notice that my blue-channel seems slightly compressed (from top to bottom of the color parade) when reviewing current RED footage on both Scarlet MX and Dragon - giving green a little more visual presence. Nothing RCX can't remedy real quick, and something that may be a thing of the past soon enough. =]
 
I sometimes see a color shift stopping my lenses down, but I remember that Ryan Walters (?) did a test over / under exposing the dragon and the higher ISO shots had a lot more green as well.
 
Akin A said:
Is it normal for Dragon to get more green as less light hits the sensor? Notice how there's more of a green tint as I closed down the aperture at higher ISO levels.
It's a very gradual change, but if you skip around in the video from beginning to middle to end, you'll notice it more.
Click to expand...


The Short answer is yes, longer answer not really. The green is because of the high iso, but high iso doesn't necessarily produce green, it depends no what you are shooting. It could easily create a little magenta and green in the same skin tone. The way I have been looking ISO, noise is a non issue if you expose properly, but the accuracy of the color is another story. Rating 250 has far more nuanced accurate color reproduction that 2000 regardless of exposure.
 
very cool test, matches great exposure wise! also surprising how usable it looks at higher ISOs (granted your doing it right by putting light on the sensor vs others doing it wrong by not having enough light and cranking it for help and ending up disappointed with results when they have to crank too far)
 
Brian Merlen said:
very cool test, matches great exposure wise! also surprising how usable it looks at higher ISOs (granted your doing it right by putting light on the sensor vs others doing it wrong by not having enough light and cranking it for help and ending up disappointed with results when they have to crank too far)
Click to expand...

Well going higher ISO is usually done when you don't have enough light like shooting city at night, I don't understand your point on putting more light on the sensor... If you put more fill light is like if you would use less iso (or working in a lower key image) and therefore having less noise.


Pat
 
Very nice test. Helpful for new Dragon players like myself. Thanks.
 
You must log in or register to reply here.
Back
Top