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 :)

R3D sample: Folks with more RED experince than I please weigh in on noise performance

Hi.

WIll look at the sample and haven't read the full thread, but...

Give me a single good reason to expose with a camerasetting below 320 ISO... :)

Cheers

Gunleik
 
Having read a bit more...

William is getting the concept about right. Even though I don't see any reason to fiddle a lot with the ISO from 800, unless you need to rise it on set, as 800 has the headroom to accept tungsten WB...


William Robinette said:
Timur, I'm very interested in having a discussion about this.

As far as I can tell, everything with regards to RED and ISO in relation to exposure is backwards from what we assume. I've started shooting at 320 indoors and 800 outdoors, exactly opposite from what you think you would do. Bringing it down to 250 ISO isn't cleaning up the image at all in terms of noise from the sensor, it just affects your exposure decisions because it changes the monitoring exposure path. I've started living and dying by the histogram and the goalposts. If you bring your R3D into REDCINE, your histogram is still crammed against the lower end of the scale. I would imagine exposing even a stop over what is "correct" for this image and bringing it down in post is going to get rid of a lot of that noise.

I'm interested in hearing more about this...
Click to expand...
 
Something that I thought I share: when prepping for a film, I was doing some camera tests and was viewing the raw footage on my crappy laptop in my hotel room. The noise on that screen was horrendous! That totally freaked me out. Short of going to the rental house and demanding a change of camera, I decided to bring the footage into the post house which we were collaborating with and run it through their system, watching it projected on a 20ft screen. And the noise was no longer there. AFAIK, the post house at that time did not have a red rocket yet so I eliminated that as a variable. It was confirmed there's no noise reduction in the chain. So apparently, something in those computers were probably influencing the way the image are shown on the screens. Graphic cards or whatever, I don't really know.

This reminds me of the time a couple of years back when Panasonic HVX200 was announced and there were heaps of complains about how noisy that camera is. So Barry Green from DVXUSER decided to run some tests and seems to, IIRC, conclude that the noise issue is partly due to variables in graphic cards of computers used to view the footage. So perhaps there's some link here?

By the way, regarding Scarlet's exposure overlay, I do notice that even when I exposed till all the red portions disappear and no traffic light is lit up (ie no clipping), looking at the image in RC-X, I can still see some clipping, albeit very small areas. Could this be due to the low resolution of the touch screen monitor that resulted in failure to show the red portions on very small areas?
 
Derrick Lu said:
It was confirmed there's no noise reduction in the chain. So apparently, something in those computers were probably influencing the way the image are shown on the screens. Graphic cards or whatever, I don't really know.
Click to expand...
My guess is that you were previewing the footage on your laptop at 1/8 or 1/16 resolution vs the post house running from the original 4k images. Previewing footage at lower resolutions in RedCine-X is going to exagerate the noise you see because it's extracting that resolution from the 4k file - not downscaling 4k to a lower resolution.

Come to think of it, could that explain the majority of the confusion in this thread?
 
Last edited:
Previewing footage at lower resolutions in RedCine-X is going to exagerate the noise you see because it's extracting that resolution from the 4k file - not downscaling 4k to a lower resolution.
Come to think of it, could that explain the majority of the confusion in this thread?
Click to expand...
<< Interesting point Brad I will definitely keep that in mind.
 
So I am not a scarlet user, but have had my red one for a few years now and thought i would have a look. With a rocket card installed on my computer i viewed your footage at full res and set it to 800 iso with only a slight tweak on the shadow control... and maybe its cause my eyes are tired, but i am seeing very very little noise and something that would be indiscernible more than a few inches from my monitor. again very tired, and contacts a bit dry, but it looks fine to me.
 
Gunleik Groven said:
Give me a single good reason to expose with a camerasetting below 320 ISO... :)

Cheers

Gunleik
Click to expand...

You only need one. 50% gray is half as noisy if you expose for 320 ISO. The tradeoff is you get 1 less stop of headroom.
 
Gavin Greenwalt said:
You only need one. 50% gray is half as noisy if you expose for 320 ISO. The tradeoff is you get 1 less stop of shadow detail.
Click to expand...


@ 320 - -I agree.

For a wider DR image, the risk of unwillingly overexposing one or more channel when you expose bellow 320, is rather imminent.

Of course, if you have full controll of the cam and understand it, you could set the iso to 50... :) For a lowcontrast image and really expose to the right. But usually that isn't how people intuitively use the ISO. I have seen it used more frequently as a composation for either lack of NDs or lights, which in both cases is futile.

* bellow 800, I have seen quite a bit of nasties under tunngsten.

But of course... If you know what you do, you know what you do.

If not... Shoot 800+ :)

(starting to sound like Mark Toia here... hahahahaha!
 
This little exposure chart for the ALEXA kind of explains it.
RED response may vary but the theory is similar.
Grayscale+Mapping+at+different+EIs.jpg
 
Michael Romano said:
This little exposure chart for the ALEXA kind of explains it.
RED response may vary but the theory is similar.
Grayscale+Mapping+at+different+EIs.jpg
Click to expand...

Yup... This is very good. Mix it with this:
http://www.gunleik.com/tests/alexmx/RED1_MX_and_Arri_Alexa/ISO_&_Kelvin_-_an_introduction.html

And start contemplating that WB is doing basically the same with the channels, when you adjust it. (Thus you have less headroom in red at Tungsten than at daylight)

And all in all the world should start to make sense. At least when you start thinking of linear mapping of light and what 18% grey really means for that.

OM OM OM OM
 
Brad Allen said:
My guess is that you were previewing the footage on your laptop at 1/8 or 1/16 resolution vs the post house running from the original 4k images. Previewing footage at lower resolutions in RedCine-X is going to exagerate the noise you see because it's extracting that resolution from the 4k file - not downscaling 4k to a lower resolution.

Come to think of it, could that explain the majority of the confusion in this thread?
Click to expand...

Interesting point on the lower rez setting in RC-X! But I was sure I was reviewing them at full rez, albeit as stills. My crappy laptop can't play back smoothly at 1/2 rez so full rez is totally impossible. So does it make a significant difference looking at a still image vs a moving one?
 
This was a really good read and I feel I have a good grasp on everything that has been discussed, but I still fight noise on the Scarlet and question if something is wrong with it. I am a still photographer, I use Phase One digital backs and am accustomed to absolutely no noise even when pushed or pulled all over the place manipulating the raw data. So maybe my expectations are too high and/or irrelevant for something in motion. But still, I would not expect to see as much blue channel noise as I do when I have clean goal posts and a full histogram. (yes, under tungsten lighting)

And sometimes there is just not enough light when balancing ratios to practicals. Sometimes the lighting ratio is perfect, and your brightest source can't go any brighter, and you are still lacking in overall quantity of light, so you have to overfill which then changes the balance and sometimes leads to a shot that looks over lit. That was a long sentence. :) With a still you can just go to a longer shutter speed and keep that lighting ratio the same, you don't have that luxury when shooting motion. These are things I am struggling with.

I would expect to see noise in shadows on the Scarlet if I had info crushed up a bit on the low end and a little red in the raw goal post. Especially if trying to lift it in post. But not when I have a full histogram and no raw clipping. I see it in 1/4 tones even. Maybe tungsten is the issue, I have not tried balancing the lens or lights back towards daylight, because the overall quantity of light goes down. But maybe it is better to have uniform noise than just a bunch of blue channel noise. Perceptually it might NOT stand out as much....?

I would expect the Scarlet to be close to 5DmkII, in average to low light, but it seems that the 5DmkII is cleaner, albeit noise is smudged away in the camera's algorithm, perceptually it looks goods.

How many people are running footage through Neat Video or equivalent when encountering noise? I find very very little visual difference when adjusting "denoise" in RCX.

Any and all thoughts are welcome!
 
Jeremy,

I was Shooting a doc today on Scarlet (which is waaaay easier than i thought it would be), and i've been exposing far to the right. Overexposing without clipping, the images are "questionable" on set but a bringing it down in RCX makes it sing. Grainless and rich. Its a different way of working. But the rewards are great. Follow the advice of others. It seems unnatural and against your instincts, but it works.

Also, got to try out hdrx today. Its the shit. Love it. Want it in 4k more than anything.....
 
right on!

Timur Civan said:
Jeremy,

I was Shooting a doc today on Scarlet (which is waaaay easier than i thought it would be), and i've been exposing far to the right. Overexposing without clipping, the images are "questionable" on set but a bringing it down in RCX makes it sing. Grainless and rich. Its a different way of working. But the rewards are great. Follow the advice of others. It seems unnatural and against your instincts, but it works.

Also, got to try out hdrx today. Its the shit. Love it. Want it in 4k more than anything.....
Click to expand...
 
Timur Civan said:
Jeremy,

I was Shooting a doc today on Scarlet (which is waaaay easier than i thought it would be), and i've been exposing far to the right. Overexposing without clipping, the images are "questionable" on set but a bringing it down in RCX makes it sing. Grainless and rich. Its a different way of working. But the rewards are great. Follow the advice of others. It seems unnatural and against your instincts, but it works.

Also, got to try out hdrx today. Its the shit. Love it. Want it in 4k more than anything.....
Click to expand...

Glad it's working out! Ditto on HDRx in 4k, mmmm.

Jeremy, I'm curious as to how you are viewing your footage. I find that makes a huge difference in perceived noise levels. Are you watching it rendered out of your NLE/finishing system or are you judging off of something like 1/4 playback in Premiere Pro? If you are in a proper grading suite with a calibrated monitor you can ignore this question of course
 
How you monitor is so important. Try this, view a really noisy image at 100 percent at full quality, then scroll through the different viewing percentages and notice how different the noise patterns looks at each size.
 
You must log in or register to reply here.
Back
Top