Is the Monitor Useless? It NEVER Matches the Output Look

[Jason Honeycutt]

I just got back from set and am looking at this scene of two people in a car. It was at night, lit well, everything looked really nice on the monitor... looked fine in the RAW view, all the detail is there... then I get it into RedCine and it's GOD-AWFUL; totally noisy (I just black-shaded) and you can barely see anything!!

[/IMG]

This happens way too much so should I basically be overlighting and assuming that what I see isn't what I get? My frustration, also, is how pristine it looks in the monitor -- with NO noise -- then you get it back and it's brutal.

(The crazy thing is that it doesn't look so bad when I export a still... but in RedCine, there's a ton of noise.)

 

[jordan levy]

the monitor has never been accurate for the look or brightness(exposure) of what you are shooting, it's way off, that's why i insist to people that are new to RED that they dont judge any of that by the RED monitors, its only for settings and information. The best and only ways to expose using that monitor is with the waveform and false color. if you have been judging lighting and exposure by just looking at the monitor you have most likely been severely under exposing, hence why you get grainy images.

so the answer is yes, it is useless for judging the output look.
but it is useful in judging exposure by the tools it provides, im sure if you can post your files there are people here that can analyze them and most likely tell you that you are under exposing, probably since you've been shooting.

 

[Jake Wilganowski]

2 things, you have to post a histogram and cam settings for anyone to know what's going on, also, if you are viewing in redcine using anything other than a FULL DEBAYER you will get a significantly noisier image

You can shoot dark scenes with red no problem, but you have to do it correctly

 

[Nick Gardner]

If the brightness on your monitor is set right (i think three clicks down from full bright) I find that it matches what you see on a computer pretty much perfectly.

You can get fooled when doing a low light exterior as your eyes are all opened up from the dark, and the monitor seems really bright. I made that mistake once. Now I double check my brightness on my monitor on every set up. The histogram and false color can be usefull as well as the wave form if you find yourself in a difficult exposure scenario, suck as a flat overcast day.

Nick

 

[Terry VerHaar]

Never, ever set exposure from looking at the monitor. RED cameras have a lot of good tools; goal posts, histogram, GIOscope, stoplights, just to name a few. Use them.

 

[Jason Honeycutt]

Thanks for these tips, guys. I'm going to just look at the scopes, click down the brightness of the monitor, etc.

 

[Jason Honeycutt]

the monitor has never been accurate for the look or brightness(exposure) of what you are shooting, it's way off, that's why i insist to people that are new to RED that they dont judge any of that by the RED monitors, its only for settings and information. The best and only ways to expose using that monitor is with the waveform and false color. if you have been judging lighting and exposure by just looking at the monitor you have most likely been severely under exposing, hence why you get grainy images.

so the answer is yes, it is useless for judging the output look.
but it is useful in judging exposure by the tools it provides, im sure if you can post your files there are people here that can analyze them and most likely tell you that you are under exposing, probably since you've been shooting.

I'm at the office now but I shot it wide open at 1.4 (shot it 180 degrees), plenty of light should have been getting in. I had an array of 1K LEDs around the car. We had to dim them because it was too bright -- or so we thought. Here's the RedLog version of it... can see everything but super noisy;

I'm just going to solely look at the scopes and false color now, not trusting this monitor. I'm still curious though, even in the raw settings, that much light and wide open lens... it's still super noisy, even though you can see everything. The crazy thing is that we used a DSLR for BTS footage and everything was super bright. Like, here's even a pic from just a cellphone, no grain at all (just compression);

But, fortunately, I have NeatVideo to clean this up... but tonight, my concern is that I don't know what else to do to lose that noise again with the same set-up. Super hot light is just going to take care of that? I'll be shooting 1.4 50mm again, maybe I'll just go 1/24 on the shutter.

 

[Jake Wilganowski]

Please post your histogram and settings and we can tell you what's up

 

[David Mullen ASC]

I use monitors all the time to judge exposures... BUT

(1) I try to balance the monitors to backwards-match what I get in tests and later in dailies

(2) I try to remember that in a dark environment like a night exterior or very dark interior, the monitor is like a light source and the image will feel brighter than it really is, so unless you want to take the time to darken the monitor way down in order to not be fooled, you have to know that if you expose on the bright side, it will look normal in a normal environment. Same goes for a monitor in a bright environment, people typically overexpose the image trying to get it to look bright enough on the monitor, only to find that they overexposed everything when they get back to a normal viewing environment. So you have to expose so that the image feels dim if viewing in a bright environment.

(3) Of course you should always double-check your exposures with other tools, at least until you get used to how things look on the monitor in that setting.

(4) Log images always look a little noisy because the blacks are lifted so don't freak out about noise in the shadows if it looks fine when viewed with the correct display gamma applied.

(5) Unless you are balancing with some element in the frame that is dim and uncontrollable, like a city skyline at night, if you are lighting and things look a little bright, try dropping to a lower ISO setting rather than stopping down or cutting the light levels, i.e. don't rob the sensor of more light just to get a darker image, just process the signal at a lower ISO setting. Like I said, the caveat here is that you may be lighting at low levels to balance with something else that has a set brightness level.

 

[Terry VerHaar]

Slight modification to my post above - never, ever judge exposure from looking at the monitor, unless you have the amount of talent and experience that David Mullen does.

 

[David Mullen ASC]

The other thing I want to add is that this is just a learning experience -- you never really stop learning about exposing, there is a lot of trial and error, that's how we learn. Maybe you got burned this time, though the image seems fine to me, but next time you'll know better.

 

[Stacey Spears]

Also, the monitor covers less than 80% of the 709 color space.

Dont try and judge noise on the LCD. The noise is really too small to see on the display. The image is also being scaled down, so that will filter out the noise. The left goal post will give you a good indecator of noise. Same for the exposure tool.

i use the display for framing and focus.

 

[Jason Honeycutt]

I use monitors all the time to judge exposures... BUT

(1) I try to balance the monitors to backwards-match what I get in tests and later in dailies

(2) I try to remember that in a dark environment like a night exterior or very dark interior, the monitor is like a light source and the image will feel brighter than it really is, so unless you want to take the time to darken the monitor way down in order to not be fooled, you have to know that if you expose on the bright side, it will look normal in a normal environment. Same goes for a monitor in a bright environment, people typically overexpose the image trying to get it to look bright enough on the monitor, only to find that they overexposed everything when they get back to a normal viewing environment. So you have to expose so that the image feels dim if viewing in a bright environment.

(3) Of course you should always double-check your exposures with other tools, at least until you get used to how things look on the monitor in that setting.

(4) Log images always look a little noisy because the blacks are lifted so don't freak out about noise in the shadows if it looks fine when viewed with the correct display gamma applied.

(5) Unless you are balancing with some element in the frame that is dim and uncontrollable, like a city skyline at night, if you are lighting and things look a little bright, try dropping to a lower ISO setting rather than stopping down or cutting the light levels, i.e. don't rob the sensor of more light just to get a darker image, just process the signal at a lower ISO setting. Like I said, the caveat here is that you may be lighting at low levels to balance with something else that has a set brightness level.

Whoa... David Mullen just jumped in on a post of mine... (I'm looking for the 'bowing down' emoji)

Thank you, great advice!

 

[Jason Honeycutt]

balancing[/I] with some element in the frame that is dim and uncontrollable, like a city skyline at night, if you are lighting and things look a little bright, try dropping to a lower ISO setting rather than stopping down or cutting the light levels, i.e. don't rob the sensor of more light just to get a darker image, just process the signal at a lower ISO setting. Like I said, the caveat here is that you may be lighting at low levels to balance with something else that has a set brightness level.

This is going to be key I think. I think I fell into the trap of trying to light for a "grade" it in camera because I personally hate overlit car interiors because if it's just the moon outside on a mountain road, there'd hardly be any light source. So I'll just light for 320 ISO, watch my scopes, then pull it back later in the grade.

That's again!! I'll post the results after tonight's shoot.

 

[Daniel Reed]

Remember you can always darken (pull down) after the fact and get great results,
but often not so much the other way around (push up).


If you're able, make a custom LUT and load it in your non-red preview monitor;
Ensure saturation is adjusted to match the change in "brightness".


If you look at dailies for big production "night scenes" you'll notice how bright they are lit.


Another caveman tip for night scenes:
Look at the scene through a 2 stop ND filter to get a closer approximation of what you want.

OR

Try wearing good sunglasses, I'm talking densities that are color neutral and that you've tested/metered in advance.

 

[Mike P.]

Are you sure your playback resolution in RCXp is set to full? If it's at anything less - even 1/2 - it'll look grainy even though it's not.

Also, there should a forum rule; "post an R3D Frame or it didn't happen."

 

[Patrick Tresch]

I use the monitor to judge exposure but I always put a -0.2 in shadow look, zoom in picture to judge the noise on darker areas, give more light on the Dragon as on the m-x (as it handles better high light roll off), shoot 800 iso and with LLO and be careful when going up or down of this level. Develop full Res as this changes a lot the noise.
Pat

 

[David Battistella]

This is a great thread with great info.

All i would add is trust all of your instruments and use them together. It's like basic pilot training, you learn to fly by wire because when you can't see you have to trust you instruments. Here, I think you probably place too much trust in one instrument (the monitor) which now can have LUTs and all sorts of other things that do not represent the light your negative/sensor (r3d) file.

the only thing not asked here was you compression setting, which in a scene like this might be best at 5:1 or 6:1.

In dark and high detail scenes folks can sometime forget that compressing the image less is better and you might not end up with this kind of surprise.

Other than that, you had enough light to probably shoot the scene at f4 but you opted for wide open, which will always soften your image and degrade the image a bit while limiting your options in post.

Take a a look at a bunch of scenes that "look" like night scenes, but you will discover most of them are very well lit while still giving the illusion of nighttime.

David

 

[Hrvoje Simic]

Monitor is useful if it is calibrated. Also, LCD monitor used for judging noise floor on set on gamma corrected footage is not the best idea because it cannot display shadow detail. This shot is underexposed.

DSLRs and cell phones apply noise reduction. Raw shooting cinema cameras don't.
That cell phone pic shows not just compression but NR mud as well. If you saw it unprocessed it would look much worse.

 

[Jason Honeycutt]

Thanks everyone... everyone's tips made such a huge difference of just flooding the light, letting the sensor soak it up. I guess I was in that mentality of hating the "video look" which usually results in way too much light and maybe I was being too subtle with it... and you guys were right -- chop it in post. The key thing for me was going from 1/48 to 1/24 and bumping up the lights a notch. I was already shooting a wide open 1.4 prime. The scopes, before, were pushed to about 1/4ish on the left (flat lined after that), now they were about to the half... even though the ones before looked okay in the monitor (which I now know can't be trusted). And, the main thing, was using the false color.

It made all the difference, you guys rock... here's a still (there's hardly any noise in the actual shot, but the compression on here kinda looks like it);