Clipping Highlights

[Sanjin Jukic]

"Waiting for Tropic Thunder" is a quick & dirty test about clipping highlights that became one of our favorite topic from the beginning.

It is an example that has range from extreme to acceptable clipping highlights and also shows sometimes where should be place for a "headroom".

Maybe it is not the best example of clipping highlights at all and it doesn't have intention to be but shows something that often happened shooting digital or film.

Aperture has range in this test from f/2.8 to f/16 and is variable as you can see it later.

About "exposure technique" using RED1 I always follow and also would not have anything to say more that was written by Jim Jannard at >>Exposure 101...>>>.


Clipping can often occur in image highlights as a result of an incorrect exposure when initially photographing ... LINK>>>


Before you can see the clip could be very good to show a couple of screen shots from other CH examples

that were shot on RED1, Genesis and film (of course all were shot with a certain artistic decision about that exposure level):

CH example 1 from "CHE" shot on RED1 and directed by Steven Soderbergh.

CH example 2 from "Before the Devil Knows You're Dead" shot on Genesis by Ron Fortunato and directed by Sidney Lumet .

CH example 3 from "Saving Private Ryan" shot on film by Janusz Kaminski (1999 won Oscar for Best Cinematography with that film)
and directed by Steven Spielberg (also Oscar for Best Director).

Clipping points are happening in general where the exposure exceeds the dynamic range of the sensor or film.

Clipping highlights can happen so often in the movies, mostly in available light situations when you have a source of natural lighting

(that mostly you cannot control) that comes from background and you want to get

(your actor/actress/object that is in the front of that light) as better visible as you can.

CH example 4 from "Waiting for Tropic Thunder" quick & dirty test shot on RED1 with Leica Summilux R 50mm f/1.4 @ f/2.8,
ISO 320, K5600, 4K 2:1, 25 fps, shutter 1/50, filters: 55mm Hoya Circ Pola + Leica IR-UV 55mm.



This footage doesn't have any sort of artistic or creative intention at all and it is a pure or simple technical example.

Shot in Vienna May 26, 2009 about 05.15 PM, on my small balcony, audio came from another room

and was live recorded on RED1 with Rode NTG3 mic and line out from Avid/M-Audio MicroTrack II recorder to RED1 Channel 1 XLR audio in.

This test is also an extension from my previous test about a comparison between vintage still glass (Dogs) and RPPs.

Footage processed in RA output as RED Log, Camera RGB, edited and graded in FCP using

Sapphire Plug-In Film Effects, scaled CC value at 2.341 that scales the amount of color correction performed due to the film types, gamma values, and exposure values.

(For example set to 0 to disable color correction. If you increase this above 1.0 it exaggerates the color correction, which normally increases the contrast.)

Output: Effect Negative Film Kodak_5245: Eastman EXR 50D, low speed, daylight balanced, very fine grain.

and

Print Film Kodak_2383: Kodak VISION Color Print Film, rich blacks.



Download "Waiting for Tropic Thunder" Quick & Dirty Test footage 169.3 MB 1280x640>>>

*


*

 

[Sanjin Jukic]

Something very interesting, more than 200 visitors of the tread and nobody is a brave to discuss the topic: CLIPPING HIGHLIGHTS!

Also I can see a lot of "foreign intelligence service work" involved here in that sort of mood or behavior.

Sorry but I couldn't resist!!!

 

[Steve Sherrick]

You'll need the original files/film transfers to do a proper comparison. How do you know what the captured file looks like without seeing it pre-graded? I think it's an interesting topic nonetheless, so take this not as a criticism but merely as a guide for doing real analysis between capturing devices. I'd love to see a comparison between things shot same time and place on different formats. That might be a bit more telling.

 

[Sanjin Jukic]

You'll need the original files/film transfers to do a proper comparison. How do you know what the captured file looks like without seeing it pre-graded? I think it's an interesting topic nonetheless, so take this not as a criticism but merely as a guide for doing real analysis between capturing devices. I'd love to see a comparison between things shot same time and place on different formats. That might be a bit more telling.

I didn't say that somebody could directly compare my files with other examples showed here taken from the

works of the respectful artists quoted above but what I said is that clipping occur as usual as we know well.

Also nobody now could pay for that sort of comparison test between RED1, Genesis, F35, film camera,...etc...

 

[Steve Sherrick]

Okay, got ya. I think I understand where you're coming from on this now. Sorry for the pointless interjection.

 

[Sanjin Jukic]

Okay, got ya. I think I understand where you're coming from on this now. Sorry for the pointless interjection.

Also I think is good to show sometimes a sort of TRASH (over-clipped highlights) that everybody making it daily just because nobody's perfect.

We are all "fishing" for better daily as we can.

BTW, it is good to watch one more time "Fishing with John" a film by John Lurie to get what I would like to say "metaphorically" .

 

[Dominic Cochran]

Not sure why the OP posted those grabs since they obviously can't be compared objectively.

Here is an interesting thread about one way(and certainly not definitive) to deal with maximizing dynamic range on the RED and minimizing clipping:

http://reduser.net/forum/showthread.php?t=28705

Hopefully even more people will post examples of how to deal with clipping instead of just examples of clipping. I also like the Recon thread the OP linked to. So what exactly is this thread trying to accomplish other than just acknowledging that clipping happens?

 

[Pietro Impagliazzo]

Hopefully even more people will post examples of how to deal with clipping instead of just examples of clipping. I also like the Recon thread the OP linked to. So what exactly is this thread trying to accomplish other than just acknowledging that clipping happens?

I must admit I'm a little lost when it comes to Sanjin's intention on creating this thread...

:rofl:

 

[Ruairi Robinson]

I didn't say that somebody could directly compare my files with other examples showed here taken from the

works of the respectful artists quoted above but what I said is that clipping occur as usual as we know well.

Ok so clipping occurs, as we all know well.

And...?

 

[Sanjin Jukic]

Not sure why the OP posted those grabs since they obviously can't be compared objectively.

Here is an interesting thread about one way(and certainly not definitive) to deal with maximizing dynamic range on the RED and minimizing clipping:

http://reduser.net/forum/showthread.php?t=28705

Hopefully even more people will post examples of how to deal with clipping instead of just examples of clipping. I also like the Recon thread the OP linked to. So what exactly is this thread trying to accomplish other than just acknowledging that clipping happens?

Dominic,

this tread never had intention to tell you how you should "deal with clipping"...

In certain available light situations to get right shot it's more similar to the "fishing metaphor" as I said above...

And those example above are mostly about that...

Also I'm not working for Sekonic and try "to sell" some profiles for RED1.

 

[Sanjin Jukic]

Ok so clipping occurs, as we all know well.

And...?

For sure Ruairi,

it's more in available light (uncontrollable lighting conditions) that not any "profile" can "fight" against it.

 

[Meryem Ersoz]

I must admit I'm a little lost when it comes to Sanjin's intention on creating this thread...

:rofl:

He's basically trying to not hijack MacGregor's thread any more, thank goodness, because that was becoming a somewhat insane conversation and killing what promised to be a valuable contribution and discussion.

There is a major difference between clipping as an artistic choice and clipping/underexposing, because you don't understand the full complement of tools available in the camera and their uses/limitations.

Sanjin, you need to take a second look at Jim's post. Note that it was called "Exposure 101" and was directed toward "new RED shooters...that are still struggling with exposure."

JJ wasn't serving up orthodoxy...he was serving up basic instructions on how to expose properly using a simple and streamlined set of guidelines.

What MacGregor was offering us was more like "Exposure 301" - a grad level course in obtaining more range between highlights and shadows, based on serious experimentation, with a clear thesis and supporting evidence.

There's no conspiracy or bad intent. Just a profile which has been thoughtfully and generously offered up to the community, and we are all free to test it. Or not. No need for veiled accusations and innuendo.

 

[Sanjin Jukic]

There is a major difference between clipping as an artistic choice and clipping/underexposing, because you don't understand the full complement of tools available in the camera and their uses/limitations.

Lol... :rofl:

Sanjin, you need to take a second look at Jim's post. Note that it was called "Exposure 101" and was directed toward "new RED shooters...that are still struggling with exposure."

JJ wasn't serving up orthodoxy...he was serving up basic instructions on how to expose properly using a simple and streamlined set of guidelines.

As I remember well JJ never suggested in his "streamlined set of guidelines" use of any sort of "instant" profiles to deal with RED1 exposure...

What MacGregor was offering us was more like "Exposure 301" - a grad level course in obtaining more range between highlights and shadows, based on serious experimentation, with a clear thesis and supporting evidence.

There's no conspiracy or bad intent. Just a profile which has been thoughtfully and generously offered up to the community, and we are all free to test it. Or not. No need for veiled accusations and innuendo.

Miguel MG and Co. openly claimed that "native" ISO of RED1 is 250 or even lower...

and that quasi fact put my rental couple of times in a "danger" when I met similar guys that

claimed the same and then we split because of that "claim" they wanated to lower the price and left in a hunt for more ISO sensitive camera...

 

[David Mullen ASC]

250 ASA is only one-third of a stop different from 320 ASA -- it's within an exposure margin of error, so it's silly to make a big deal over this. You can't say that the RED is definitely 320 ASA... but definitely not 250 ASA when I could shoot a face twice, one-third of a stop different in exposure for each --- and they would barely look different.

 

[Miguel "Macgregor" De Olaso]

Perhaps someone at RED should explain what they understand by "native ISO". To me, it all depends where you want to set the mid grey in your dynamic range.

I don´t think you can define a NATIVE ISO in a raw camera. In a RGB camera that incorporate curves (Varicam or Sony F35 for example), it´s easier to define a working ISO, but you will notice that it will change depending on the curve that you are using. And this makes sense, since using Film Mode in a Varicam increases the dynamic range by changing the shape of the gamma curve. This decreases the mapping position of the mid gray into a lower zone. You get more noise yes, but you get a couple extra stops in highlights, which are great for creating a nice shoulder that mimics film. And also the result is an increase in the exposure setting, therefore if you use a photometer you´ll get a higher reading (let´s say 200 vs 400ISO).
Same thing happens with the F35 and the hypergammas.

The RED ONE initializes at 320ISO. If anyone is really concerned about selling a camera to their costumers that is trully 320ISO, that´s fine.
The problem will come when the DP measures the light in the set and the director looks at the monitor. He will complain about everything looking too dark.
Exposing with the lightmeter at 320ISO will probably capture a really nice image. It will have a bit of grain, but the highlights will be well preserved.

The inconsistency right now resides in that the outputs are not showing a matched image to what you should expect from your lightmeter. And if the outputs fails, False color fails too.

Possible solution:
- add curves to the outputs of the RED ONE instead of just a boosting the white input level (all cameras i know use curves. The RED is too linear in the way it shows the image. There is no rolloff in 709mode basically because the red processor it´s just applying a 2.2 gamma over the linear image (Or at least, this is what seems to be happening). This works fine if you´re setting your mid grey at a very high zone (what RED calls ISO 250-320), but the camera would create a better looking image if the mid grey was mapped into a lower zone together WITH some kind of highlight compression (knee) in the video output. I belive this was the attempt when RedSpace was created. And this is what people is manually doing with Look profiles. Why not making it standard?).

This would preserve highlights in the preview and increase the mid grey up to a point where false color gets the same reading as your reflected light measured by any lightmeter.

 

[Ben Holmes]

As much as I enjoy having to wade through pages of infantile and needlessly aggressive posts when trying to find the best ways to process and view RED footage (that was sarcasm btw), could I suggest, Sanjin, that you just try MacGregor's method, and see how it works for you? If it doesn't, then carry on doing it your own way. It's really that simple.

I would suggest that anyone trying to negotiate a better price for your camera based on these threads/suggestions was insane, and should not have expensive equipment left in their care at all - or any sharp objects at all.

I have an actual living to make from this, whilst still trying to spend time with my family and friends - and REDUser is bad enough to wade through without all this extra nonsense.

Many thanks.

 

[Luca Immesi]

Two days ago I met in Rome one of the best italian d.o.p. He has to shoot a movie on red in Italy with a famous international director. He was quite disappointed about a camera (the red) that can only shoot at 320 Iso 5000/5600° temperature. He said cinema is at 3200k not daylight and he felt very limited to shoot this way, with film he said I could choose different kinds of film stocks for different situations. Then we did some tests and he told me he thoughts the camera iso ratings was closer to 250 iso than 320 and after he was more relaxed watching the quality of the images projected in HD. You know, these big guys are used to film and they have always something to complain about digital (especially in Italy) but beside this experience I thought I'd like to have more profiles (red looks) like different kinds of film stocks. A 250 daylight profile, a 500 asa tugsten and so on... I know they are metadata but they influences how you take the exposure.
I know also we have pretty much unlimited possibilities in post but I think having different profiles for different situations could be a good option. And "sometimes" having a profile that helps not to clipping highlights it's very useful!

 

[Sanjin Jukic]

Now I'm back again but must say that this tread has got a pretty boring.

After all thinking about the topic I could openly agree with Geoff Boyle FBKS that cinematographer's job is

"to make images not to know how to use a particular technology".

Also I went yesterday to a short trip and shot a bit of handheld tests at windy weather

that was just before the rain storm in a sort of "all the time change" lighting conditions with almost dark sky at Neusiedler See, Podersdorf, Austria.

And of course that I took care about clipping highlights as mush as possible but not so extreme.

During lens one exchange at open and strong wind situation it happened to drop a little dirt/speck on lens, filter or sensor...but after everything was pretty OK.

Shot on RED1 handheld with Leica Summilux 50mm f/1.4, ISO 320, K5600, 25 fps, 1/50, @ f/4, filters: Hoya ND4 + Hoya ND2 + Leica IR-UV 55mm.

Here is Quick & Dirty test footage to download 1920x960 HD 216.6 MB>>

 

[HBrennan]

REDUser is bad enough to wade through without all this extra nonsense.

heh... couldnt agree more...

 

[Sanjin Jukic]

As much as I enjoy having to wade through pages of infantile and needlessly aggressive posts when trying to find the best ways to process and view RED footage (that was sarcasm btw), could I suggest, Sanjin, that you just try MacGregor's method, and see how it works for you? If it doesn't, then carry on doing it your own way. It's really that simple.

I would suggest that anyone trying to negotiate a better price for your camera based on these threads/suggestions was insane, and should not have expensive equipment left in their care at all - or any sharp objects at all.

I have an actual living to make from this, whilst still trying to spend time with my family and friends - and REDUser is bad enough to wade through without all this extra nonsense.

Many thanks.

heh... couldnt agree more...

Terribly sorry guys if I still didn't "satisfy" you but

I mostly never waste a time and always try to learn from professionals in the field:

LINK>>>