Dragon Latitude

[Stewart Read]

Hi all

I am trying to wrap my head around using my Sekonic 758 Cine Light Meter with my Dragon.

For this test I rated camera at ISO 320 with STH OLPF:

Shooting out a window the spot meter was giving me f32 peaks.

I adjusted exposure to 5 stops below this at f8.0. In RAW view and watching traffic lights was expecting this exposure to give me enough latitude to hold those f32 peaks but this as not the case. I think I was finding it closer to f11 which would be 4 stops headroom.

Am I missing something ? Quite possibly I admit!

Should I be using ISO 640 to reach further into the highlights and then ND ?

Grateful for any advice re this as I am keen to understand highlight protection better.

Thanks !

 

[Michael Tiemann]

Hi all

I am trying to wrap my head around using my Sekonic 758 Cine Light Meter with my Dragon.

For this test I rated camera at ISO 320 with STH OLPF:

Shooting out a window the spot meter was giving me f32 peaks.

I adjusted exposure to 5 stops below this at f8.0. In RAW view and watching traffic lights was expecting this exposure to give me enough latitude to hold those f32 peaks but this as not the case. I think I was finding it closer to f11 which would be 4 stops headroom.

Am I missing something ? Quite possibly I admit!

Should I be using ISO 640 to reach further into the highlights and then ND ?

Grateful for any advice re this as I am keen to understand highlight protection better.

Thanks !

A spot meter is typically a 1 degree field of view. A 35mm ARRI Master Prime has a 39.4 degree field of view across an ANSI Super 35mm frame, which is 24.9mm wide. The DRAGON pixel pitch is 5 microns, which means there are 4980 pixels across that 39.4 degree horizontal field of view, or 126 pixels per degree. The area of a 1 degree circular spot is pi * radius squared, or about 12500 pixels. If you are measuring a subject whose illumination is perfectly uniform across the entire spot circle, your light meter and your dragon should agree. But if the illumination is not perfectly uniform, there's a chance that some of those 12500 pixels illuminated less or more than the rest, with some illuminated enough to clip. If enough of those pixels are clipping, they will light the traffic lights, even if the other ~12000 pixels are in range.

If you use a super wide angle lens like an 8R, there will be about 10x fewer pixels fighting against your averages, but if you use a longer lens then many, many more pixels have the opportunity to catch light that's enough to clip and trip the traffic lights.

 

[Stewart Read]

Thank you Michael - very interesting information.

 

[Nick Morrison]

640 would shift your headroom up a stop, yes...at the expense of 1 stop in blacks.

 

[Phil Holland]

640 would shift your headroom up a stop, yes...at the expense of 1 stop in blacks.

Not necessarily at the expense of "losing" a stop in the blacks however. This would really depend on the scene and if crushing was occurring in the blacks.

The general concept is when you are shooting you are always capturing the total Dynamic Range that Dragon has. Depending on who you talk to that's different things, if you're talking to me under practical conditions that's around 16/17 stops. The ISO Rating is just moving and mapping tones and colors within that total captured Dynamic Range. If the scene supports it you'll have a workable latitude that exists within the exposure if nothing is clipping or crushing. Though if you are digging into the shadows a great deal you will be seeing more noise/texture/grain.

To answer your question Stewart, yes rating a higher ISO value allows for more highlight values.

A few graphics that can help.

Examples of gray patches across the exposure range and the effect of ISO rating and adjustments:

And here's a scene and 100% crop from the Low Light Optimized OLPF showing the extra highlight detail and increase of noise/texture/grain in the shadows as you increase your ISO.

 

[Michael Tiemann]

Phil's post brings up another important point which many either forget or completely misunderstand. It is true that RAW allows you to choose your ISO after the fact. But you will always, always, always get better images if you expose properly for your intended ISO. In other words, if you are shooting at ISO 800 and you notice that things are looking about two stops under, don't say to yourself "no problem, I've shot ISO 3200 before and it's worked...I'll just dial up the ISO in post." Better to change the ISO then and there and deal with the realities it's telling you. Conversely, if you know that you don't like the shadow noise that ISO 3200 gives you, then by all means make other adjustments so that ISO 800 is on the mark.

 

[Stewart Read]

Thank you Phil

 

[Stuart Graham]

Thanks Phil,
I'm still getting 'the director is concerned shooting on RED as it has highlight issues' Zzz

Stu.

 

[Michael Blieden]

Hey Stewart. I am still learning about my 758C as well, and it in turn is helping me learn a lot about the dragon sensor.

This is slightly off the topic of your question but I'm interested in how these two amazing tools work in concert, so I wanted to add to the thread, hopefully without taking it in too different a direction.

I've spent a lot of time trying to create a custom sekonic profile for the dragon using the sekonic desktop software, but I was unable to capture the full dynamic range of the sensor.

I wrote a blog post about it here.

After several different attempts this is the best I could do:

As you can see this indicates a DR of 12.1 EV. Which I'm hearing from several others DP's is still an admirable effective DR, so I'm not too disappointed. But in a labratory kind of way I'd like to understand my results. I've seen at least one other Sekonic Profile created by Ryan Walters for Dragon that captures more otal usable stops.

You can download my custom profile here. It was created with output images that were uncorrected Filmlog. I created other profiles with export images from RedGamma4 but as you can expect the effective dynamic range of that profile was even smaller.

I had the meter calibrated recently so I think it's accurate, btw. But this profile adds an exposure offset as well, and I'm not certain that I like this offset, after comparing its results to the calibrated, stock exposure.

If you or anyone who uses this meter has any insights would love to hear about it.

 

[Nick Morrison]

Thanks Phil,
I'm still getting 'the director is concerned shooting on RED as it has highlight issues' Zzz

Stu.

Really? Just shoot LLO 1280 and get all the highlights you need!!

 

[Phil Holland]

Thanks Phil,
I'm still getting 'the director is concerned shooting on RED as it has highlight issues' Zzz

Stu.

If this was Mysterium-X we are talking about I can understand certain concerns, specifically when it comes to clipping, Dynamic Range, and DRX. However, you are shooting on Dragon and there is "no DRX" anymore as the Highlight Roll-Off fades to clip rather nicely with the boost in DR.

Show the director a few tests and if possible against other cameras, that's what I do. My last 3 jobs all required me to do so and all of them have been shot on Dragon between 5K through 6K on both OLPFs.