Issues

[Miguel "Macgregor" De Olaso]

That makes sense. Lifting exposure simply shifts up the whole exposure range. Watch the histogram. You won't have blacks any more unless you use curves. Curves are the way to go. Crush those shadows.

I was once told that ISO change was the first and best thing to do because it was performed before the debayering of the sensor data.

 

[Adrian T.]

I was once told that ISO change was the first and best thing to do because it was performed before the debayering of the sensor data.

If the whole picture is underexposed, yes, of course, do it first. But be careful not to clip the whites. But afterwards you need to bring down the shadows again. Set an appropriate black point and a nice S curve.

 

[Warren Kommers]

Here are my thoughts after playing with these stills and perhaps things will be better soon.

-Yes on curves for sure.

-Hello 270 degree shutter for exposure help. (Perhaps the sensor is not as sensitive as we had hoped).

-compression looks too heavy as opposed to sensor noise being the culprit.

I am worried about things like this thread despite the useful shooting information being generated and feedbck to RED. I have a feeling it's going to spread many bad rumors here in LA when really the camera and image are still beta. People like to make quick judgements here. I can't tell you how many times on set we hear an AC say something like, "I worked on a RED feature last month. You can't even monitor in two different places!" Implying of course that this will never change, which it has already. I like that these first 100 cameras are being tested by the public and the results will get better from their feedback but this kind of stuff spreads like wildfire round here and may slow the redvolutions immediate impact at least.

 

[Casey Green]

David,

I just took one Foster's beer with pizza before I will start with LAPHROAIG whisky later.

By the way if ever you pass Vienna give a call and I will show you the best of "Wien".

can we stay on topic here please. constructive comments and the like?

 

[Jim Arthurs]

Back in July, Mike Brennen questioned whether a filter to correct to tungsten might be the way to go... I joined in with the question, hoping to see some testing on this, possibly at LART, and I was asked by Evin Grant (who was becoming frustrated at my insistance that this might be neccessary) this question...

http://www.reduser.net/forum/showpost.php?p=58396&postcount=9

I had no good answer, as I'm not an expert at RAW techniques. I think we can answer his question now, and say that it's increased grain/noise/whatever in the blue channel as a result of not pre-filtering with CC in front of the lens, and to look for that in the tests.

Now, up until yesterday, there has been no talk _anywhere_ that shooting under tungsten is potentially more noisy in blue than shooting under daylight balanced sources. Now we have some statements by both J.J. and Deanan indicate it might be so...

http://www.reduser.net/forum/showpost.php?p=116287&postcount=57

This is the nature of learning and experimenting, and no one's "fault" either for not knowing the answer, or for asking the question in the first place!

Knowledge is good, and if I know this thing, I can ignore or worry about it on a per situation basis. For instance, if I want the best green or bluescreen plate, I would sure take a hit on the light levels and use proper filtration, or half filtration as Stephen Williams suggests, or go HMI on set if it means less noise which means cleaner matte edges and smoother keyed shadow areas. Other times, the minute quality gain of the correction might not be worth the bother due to the subject matter or any other reason.

Now I suggest something really, really radical!

Someone, with access to a camera do a test and end the debate one way or the other. Just set up a scene or chart and light it with both tungsten and then daylight sources, shoot with RED un-filtered and see if there is any more noise in the tungsten lit scene. If so, then filter the camera to correct and shoot another clip. Share the results, end of discussion.

Nothing about the situation "bothers" me one way or the other... the only goal I have is a better understanding of the best methods for using this wonderful tool to its full potential.

 

[Casey Green]

Here are my thoughts after playing with these stills and perhaps things will be better soon.

-Yes on curves for sure.

-Hello 270 degree shutter for exposure help. (Perhaps the sensor is not as sensitive as we had hoped).

-compression looks too heavy as opposed to sensor noise being the culprit.

I am worried about things like this thread despite the useful shooting information being generated and feedbck to RED. I have a feeling it's going to spread many bad rumors here in LA when really the camera and image are still beta. People like to make quick judgements here. I can't tell you how many times on set we hear an AC say something like, "I worked on a RED feature last month. You can't even monitor in two different places!" Implying of course that this will never change, which it has already. I like that these first 100 cameras are being tested by the public and the results will get better from their feedback but this kind of stuff spreads like wildfire round here and may slow the redvolutions immediate impact at least.

Good points, but at some point you have to have faith that the Camera will stand on it's own. There will always be people out there believing what they will and I don't think we should worry too much about that. Just keep on going, making the Camera the best it can be with valuable feedback.

LART should definitely help a lot with tons of new footage in a test environment.

Keep up the great constructive feedback everyone.

 

[Warren Kommers]

Back in July, Mike Brennen questioned whether a filter to correct to tungsten might be the way to go... I joined in with the question, hoping to see some testing on this, possibly at LART, and I was asked by Evin Grant (who was becoming frustrated at my insistance that this might be neccessary) this question...

http://www.reduser.net/forum/showpost.php?p=58396&postcount=9

I had no good answer, as I'm not an expert at RAW techniques. I think we can answer his question now, and say that it's increased grain/noise/whatever in the blue channel as a result of not pre-filtering with CC in front of the lens, and to look for that in the tests.

Now, up until yesterday, there has been no talk _anywhere_ that shooting under tungsten is potentially more noisy in blue than shooting under daylight balanced sources. Now we have some statements by both J.J. and Deanan indicate it might be so...

http://www.reduser.net/forum/showpost.php?p=116287&postcount=57

This is the nature of learning and experimenting, and no one's "fault" either for not knowing the answer, or for asking the question in the first place!

Knowledge is good, and if I know this thing, I can ignore or worry about it on a per situation basis. For instance, if I want the best green or bluescreen plate, I would sure take a hit on the light levels and use proper filtration, or half filtration as Stephen Williams suggests, or go HMI on set if it means less noise which means cleaner matte edges and smoother keyed shadow areas. Other times, the minute quality gain of the correction might not be worth the bother due to the subject matter or any other reason.

Now I suggest something really, really radical!

Someone, with access to a camera do a test and end the debate one way or the other. Just set up a scene or chart and light it with both tungsten and then daylight sources, shoot with RED un-filtered and see if there is any more noise in the tungsten lit scene. If so, then filter the camera to correct and shoot another clip. Share the results, end of discussion.

Nothing about the situation "bothers" me one way or the other... the only goal I have is a better understanding of the best methods for using this wonderful tool to its full potential.

For sure. Some useful information is being generated here. It's a beautiful thing

However, being to expected to shoot at 160 asa in interiors, using a 80A, or using finicky expensive HMI's as opposed to tungsten are not small adjustments in the professional cinematography world. I'm optimistic though that the RED team will find a solution.

 

[chuck colburn]

Hi Casey,

I see from your bio that you have done tons of video playback operating. While doing that did you ever notice if color correction filters were being used on the cameras? Or did they have you chained up to the video village. lol
I'm wondering if this is done with other video cameras as a standard practice when shooting in day light.

thanks,

 

[Dalibor Fencl]

Maybe we can point out a result, valid not only for the RED but all digital cameras. Assuming that all sensors are daylight balanced, It's always wise to use HMIs rather than tungsten to achieve less grain. (At least less grain in blue channel)

 

[Rudi Herbert]

I have a feeling filtration will definitely help, and so will going with HMI's...

However, although we have touched on the subject, there's no real consensus on how much of a difference the REDCODE codec makes. A codec is a VERY tricky thing and the lower its data rate, the lower the final quality image will be, this is a fundamental fact. Now, squeezing a 4 K image into such low rates always felt to me like a very optimistic thing to do, however innovative and revolutionary the concept of recording 4 K to CF cards is. All the solutions regarding filtration, tint and color temperature are VERY valid, but I suspect (and a suspicion is nothing more than a half informed conjecture) that the RED image will not be "all it can be" until a higher rate codec is implemented and/or a compression of no more than 5:1 is used. 10:1 is way too high, again, just a personal feeling.

For whatever is worth, might it not be something interesting to get our hands onto a few shots from an SI 2K camera and take a look at the noise levels? I know the SI is only 2K and I don't know the specs of the Altasens sensor, but the SI is the only other camera besides the RED that does onboard compression (via Cineform codec) and sends out a compressed RAW image. Both cameras are vastly different animals, but for the sake of enlightment, I would love to see some SI/Cineform images shot under similar circumstances as the frames Mc Gregor presented to see how that combination deals with the noise relative to the RED.

 

[Warren Kommers]

Now, squeezing a 4 K image into such low rates always felt to me like a very optimistic thing to do, however innovative and revolutionary the concept of recording 4 K to CF cards is. All the solutions regarding filtration, tint and color temperature are VERY valid, but I suspect (and a suspicion is nothing more than a half informed conjecture) that the RED image will not be "all it can be" until a higher rate codec is implemented and/or a compression of no more than 5:1 is used. 10:1 is way too high, again, just a personal feeling.

That's what I feel like I'm seeing as well. It's just too much perhaps. I'd be completely willing to lose the convience CF cards and shoot 40-45 Mbps if it meant a better image that needs to be ready for a 50' screen. Even 100Mbps would be so acceptable. The camera would still be so friggin breakthrough.

 

[Dalibor Fencl]

The SI2K with it's 2/3 sensor has pretty same density of pixels as RED's 35mm at 4K so, you're right the noise should be comparable. The rest is Redcode versus Cineform. ?????

 

[David Mullen ASC]

Having just finished reading another book on 3-strip Technicolor, which was daylight-balanced for much of its early life, I just wanted to point out what DP's back then were saying, that the switch to a tungsten-balance made life a lot easier for doing night interior scenes because of the practicality of smaller tungsten fresnel lamps for portrait work and for matching tungsten practical lamps.

HMI's are fine for big lighting jobs, sort of brute force lighting of large spaces or for large soft sources, but for smaller spaces, you don't really have a good HMI equivalent to a Tweenie, Midget, Inky, Pepper. Of course, you could use blue gel on those units, but at a considerable loss of light output. And you'd be putting in blue photofloods and daylight compact flos into your table lamps and whatnot. And then there are the issues if you want to put your lights all on dimmers.

I think the notion that movies are going to start lighting small night interior scenes supposedly lit by low wattage tungsten tamp lamps, candles, and whatnot, with blue lighting is highly unlikely. Nor are people going to start using blue filters and working at low ASA values as a result.

 

[Casey Green]

Hi Casey,

I see from your bio that you have done tons of video playback operating. While doing that did you ever notice if color correction filters were being used on the cameras? Or did they have you chained up to the video village. lol
I'm wondering if this is done with other video cameras as a standard practice when shooting in day light.

thanks,

Hi Chuck,

The Video Playback work I primarily have done was for on-set displays (CRT, LCD, Plasma TVs, Computer Screens, Projection, etc.) within the scene.

When the Film or Video Cameras shoot our screens, weather they are just in the background or are a story specific hero, we do color correction to match both the lighting and Stock on a Film Camera or color balance on the Digital Camera.

If for instance, there is an interior scene using Tungsten lights, we use our color correction processor box to change the video temp to 3200, otherwise the TV monitor will come out looking very blue. (which sometimes Directors ask for because of years of stereotyping the TV light is blue. - We all know it isn't, it's just in years past, noone was correcting for this, so whenever TVs were shot (or even just seeing the light cast on the walls of the set), the Video usually was in the wrong temp and came out looking blue.

We on occasion use CTO gels over screens if for some reason we lack a color corrector or the monitor is limited in it's ability (such as a Black and White monitor).

 

[Adrian T.]

I have a feeling filtration will definitely help, and so will going with HMI's...

Filtration won't help. Noise is apparently present in underexposed areas. Using a filter you would underexpose them even more, and the perfectly exposed areas shift down into underexposure.

HMI will help, in the sense that it's a powerful light. Use it or any powerful tungsten to fill your shadows. Bring them out of underexposure. And then in post, crush them all to get a nice noise free picture.

 

[Warren Kommers]

I think the notion that movies are going to start lighting small night interior scenes supposedly lit by low wattage tungsten tamp lamps, candles, and whatnot, with blue lighting is highly unlikely. Nor are people going to start using blue filters and working at low ASA values as a result.

totally. although constructive it's not realistic nor compatible with modern lighting trends.

 

[chuck colburn]

Casey,

I got it. I love this site, it's an everyday learning exsperience.

thanks,

 

[Rudi Herbert]

Filtration won't help. Noise is apparently present in underexposed areas. Using a filter you would underexpose them even more, and the perfectly exposed areas shift down into underexposure.

Well then, if we add to that the logic and common sense of David Mullen's assesment that this realistically won't happen any time soon, then for many of us who are basically placing our hopes on RED, we better start looking at whether the REDCODE is the main culprit or not. I, for one, would sooooooo not care if I have to record to a hard drive instead of a CF, however amazingly convenient the latter one is.

 

[Jim Arthurs]

Hi Casey,

I see from your bio that you have done tons of video playback operating. While doing that did you ever notice if color correction filters were being used on the cameras? Or did they have you chained up to the video village. lol
I'm wondering if this is done with other video cameras as a standard practice when shooting in day light.

thanks,

Just an FYI from NAB back in 2002 or so... I was in the Sony booth talking to the guy involved in the "Panafication" (if that's a word) of the f900's and he told me that you shouldn't white balance that camera, the sensor is 3200 native and just filter it. Of course, easier done going that CC direction, and I wonder, like you, how many people do it?

 

[Dalibor Fencl]

Having just finished reading another book on 3-strip Technicolor, which was daylight-balanced for much of its early life, I just wanted to point out what DP's back then were saying, that the switch to a tungsten-balance made life a lot easier for doing night interior scenes because of the practicality of smaller tungsten fresnel lamps for portrait work and for matching tungsten practical lamps.

HMI's are fine for big lighting jobs, sort of brute force lighting of large spaces or for large soft sources, but for smaller spaces, you don't really have a good HMI equivalent to a Tweenie, Midget, Inky, Pepper. Of course, you could use blue gel on those units, but at a considerable loss of light output. And you'd be putting in blue photofloods and daylight compact flos into your table lamps and whatnot. And then there are the issues if you want to put your lights all on dimmers.

I think the notion that movies are going to start lighting small night interior scenes supposedly lit by low wattage tungsten tamp lamps, candles, and whatnot, with blue lighting is highly unlikely. Nor are people going to start using blue filters and working at low ASA values as a result.

I'd say in small interiors there will always be enough of light even with small set of 200W Dedolight, so the underexpositions - hence noise issues, won't be problem, but in larger interiors and exteriors with large dark parts the HMIs might be the right choice.