Welcome to our community

Be a part of something great, join today!

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 :)

Dragon and low light noise

Brian Broz said:
I recently noticed a huge reduction in noise (on my Dragon) switching from a manual to the "factory" black calibration. It was very surprising...and noticeable.
I've been too busy to test or read through numerous threads but it certainly worth investigating further IMHO.
Click to expand...

Very, very interesting! Can someone from RED confirm this? Does RED do the same process end users do, except they get to save it in the 'factory' setting?

I suppose its possible to assume the manual black calibrations were performed incorrectly in one way or another, but have any of the experts on here had such an experience with a perfectly performed manual black shade?

Since black shading is so temperature dependent, should the fan setting be at a certain specific setting to do the best possible black shade? Will the fan changing the sensor temp during the black shade potentially affect it?
 
guys I've got the solution worked out....just wait til you see my refrigerator module, you will never black shade again! seriously bro!
 
I'm getting the same problem. Just got my Dragon last week with the Low Light OLPF. The level of noise I'm getting seems much higher than with my Epic. I've contacted Red and will do some tests with the sensor calibration and report back what I find.
 
Michael Gyori said:
I'm getting the same problem. Just got my Dragon last week with the Low Light OLPF. The level of noise I'm getting seems much higher than with my Epic. I've contacted Red and will do some tests with the sensor calibration and report back what I find.
Click to expand...

This kind of stuff scared me of the upgrade . I have no problems with noise on the epic. At the same time most seams to be cool with the low light olpf up to iso 2000. On the other hand, if it still is some quirks left, red will iron them out soon.
 
Hi All,

Last night, I shot with some low light scene and although I apply all the experiences that you guys share I still see noise in the frame. Would like to check with you guy whether this could be denoise in post production ...

15765499359_ec256eafd8_c.jpg
 
I think Brandon was meaning an actual R3D snaphshot (single frame) so people can have a look at it in RCX and offer advice. :)

Edit: Lol - Brian posted the same thing.
 
How is the noise after grading?

Redlog film is the starting point for grading and will always, always appear noiser than a normally graded image. Just want to be sure that part is understood because log images seem to be finding their way into many "final" images.

noise tends to fall off when a contrast curve is added.

Battistella
 
David Battistella said:
Just want to be sure that part is understood because log images seem to be finding their way into many "final" images.
Click to expand...

Log is the new shallow DOF.

It used to be "look at my large sensor and the shallow depth of field I got—even if it's totally inappropriate!"

Now it's "look at my sweet over exposed, washed out log images! You can bet these came from one of 'em sweet cinema cameras you hear so much about!"

Sony users seem particularly guilty with "14" stops of DR.
 
Like the term "low light" seems to be substituted for "no light", as in "no metered light", as in, if I spot metered that scene there might not be a reading, or that meter reading might be below f 0.9 or something.

generally stuff can be noisy even if it is actually lit to 1.4 and the electronic camera noise floor is pretty incredible at that.

Battistella
 
My issues with Dragon noise seems to be related to relying on tungsten lighting, which is starving the blue channel. And since the Dragon sensor is a daylight calibrated sensor (as I've been reminded by people at Red), it's more vulnerable to deficiencies in the blue channel. Therefore, noise is introduced with low exposed tungsten lit scenes. It would seem then, that if you can't have much light, at least make sure you have some blue light (and change the temp in metadata or post if you want a warmer look), right? All this talk about blackshading seems to be mostly a red herring compared to exposure. Maybe it's possible to get a clean 2000 ISO image with daylight lighting, but I don't see how that's possible with tungsten. Are there blue flame candles? :)

Also, don't expect the Low Light OLPF to enable one to shoot clean low light images. I bought into the hype and my expectations were too high. The Dragon, even with a LL OLPF is not a great low light camera. Other than this, I love most everything about this camera, and I know I'll be capturing some amazing footage with it... just not in many low light situations.

I saw a RedUser thread where people are speculating on what Red is working on next. I hope it's a sensor that does well in low light. The dual ISO Panasonic Varicam 35 produces some amazing low light shots. I'd much rather have great low light performance than 8k. I also hope that they don't announce their next sensor until they are ABSOLUTELY sure it's 6 months away (or less). :)
 
Last edited:
Dragon at 2000ISO @f16. Sure there is some "to be expected" noise. but not a disaster at all. Meter reading left for reference.

iso2000dragon.jpg
 
Michael Gyori said:
My issues with Dragon noise seems to be related to relying on tungsten lighting, which is starving the blue channel. And since the Dragon sensor is a daylight calibrated sensor (as I've been reminded by people at Red), it's more vulnerable to deficiencies in the blue channel. Therefore, noise is introduced with low exposed tungsten lit scenes. It would seem then, that if you can't have much light, at least make sure you have some blue light (and change the temp in metadata or post if you want a warmer look), right? All this talk about blackshading seems to be mostly a red herring compared to exposure. Maybe it's possible to get a clean 2000 ISO image with daylight lighting, but I don't see how that's possible with tungsten. Are there blue flame candles? :)

Also, don't expect the Low Light OLPF to enable one to shoot clean low light images. I bought into the hype and my expectations were too high. The Dragon, even with a LL OLPF is not a great low light camera. Other than this, I love most everything about this camera, and I know I'll be capturing some amazing footage with it... just not in many low light situations.

I saw a RedUser thread where people are speculating on what Red is working on next. I hope it's a sensor that does well in low light. The dual ISO Panasonic Varicam 35 produces some amazing low light shots. I'd much rather have great low light performance than 8k. I also hope that they don't announce their next sensor until they are ABSOLUTELY sure it's 6 months away (or less). :)
Click to expand...

From my testing and what I have seen on this site, the red channel is where the Dragon struggles not blue, so in fact, tungsten light should reduce noise, and thats what I have seen on Dragon. Tungsten desaturates colors, but is cleaner underexposed, than daylight underexposed.

Also I think RED is doing low light the right way. For image fidelity, lower ISO is better, If you can add more light and lower ISO, you will have a beefier negative. To me RED influences you to shoot that way, "Low Light" cameras advocate 3200 ISO or higher, but realistically speaking you aren't getting the highest image fidelity out of the camera.

Also if you need high isos RED can accommodate, on either OLPF, as long as you expose for your ISO and can do noise reduction. I like that with red the control of noise is in post with all the information there rather than destroying potential information in camera. If you run Noise Reduction on Red, you know even when its very noisy to the eye, it doesn't take much to restore the image.
 
David Battistella said:
Dragon at 2000ISO @f16. Sure there is some "to be expected" noise. but not a disaster at all. Meter reading left for reference.

iso2000dragon.jpg
Click to expand...


Pardon my ignorance, but what does the 5000k on your info chart indicate?

While the noise is not terrible, it's not great... not clean.
 
Ryan Emanuel said:
From my testing and what I have seen on this site, the red channel is where the Dragon struggles not blue, so in fact, tungsten light should reduce noise, and thats what I have seen on Dragon. Tungsten desaturates colors, but is cleaner underexposed, than daylight underexposed.

Also I think RED is doing low light the right way. For image fidelity, lower ISO is better, If you can add more light and lower ISO, you will have a beefier negative. To me RED influences you to shoot that way, "Low Light" cameras advocate 3200 ISO or higher, but realistically speaking you aren't getting the highest image fidelity out of the camera.

Also if you need high isos RED can accommodate, on either OLPF, as long as you expose for your ISO and can do noise reduction. I like that with red the control of noise is in post with all the information there rather than destroying potential information in camera. If you run Noise Reduction on Red, you know even when its very noisy to the eye, it doesn't take much to restore the image.
Click to expand...

Perhaps my issue then is just having one of the color channels starved... blue, red, or green. Balanced light would then be the best, no? I'll have to experiment further.
 
I always leave the metadata at 5000k. That is what 2000iso exposed at f16 looks like. You can expect noise. You don't want noise, use lights. Starve the sensor of light, expect noise.

This was shot with the sth OLPF and the WB was taken on the middle grey patch. Source was tungsten.

You our theory on blue channel noise is a bit far fetched as dragoncolor is producing some of the most accurate color I have seen on digital sensors under tungsten light.

Battistella
 
You must log in or register to reply here.
Back
Top