Low Light OLPF VS Skin Tone OLPF. Photos Inside.

[Corey Benoit]

So I just wanted to see if these looked like they are calibrated correctly and if the results I got are good. The room was average ambient tungsten light. ISO range 1600-12,800. I black shaded for both tests and switched the config for each OLPF in the menu.



Skin Tone OLPF @ 1600

Low Light OLPF @ 1600

Skin Tone OLPF @ 3200

Low Light OLPF @ 3200

Skin Ton OLPF @ 6400

Low Light OLPF @ 6400

Skin Tone OLPF @ 12,800

Low Light OLPF @ 12,800

 

[Nick Morrison]

Looks good! What firmare? The latest?

 

[David Battistella]

Great test Corey,

did is you rate and expose for each ISO or did you just push process?

i think where you would see a difference is if you under exposed by a few stops and then push processed, the sth would not hold up as well in that situation.

the real problems show up when the sensor is light starved. What was the exposure here?

 

[Tom Gleeson]

Corey,

Great to see somebody doing tests with their camera. Having said that its hard to view your results and see any noise with heavily compressed 1024pix JPEG's. Really need to see the R3Ds or minimum a codec like 1920 Pro Res LT ? It would also be better to see motion as its hard to look at noise structure on a still frame. Apologies for being a Party pooper.

 

[Patrick Tresch]

A 100% crop Side by Side video would be great!

Pat

 

[Julio Quintana]

Since you posted this in the Scarlet forum, I assume this is Scarlet Dragon at 5K?

 

[Jamon Lewis]

Good test, pleasantly surprised at the results from the STH. In general with Dragon I find it hard to understand why there would be noise in areas that aren't under-exposed especially with the LLO.

 

[Nick Morrison]

Good test, pleasantly surprised at the results from the STH. In general with Dragon I find it hard to understand why there would be noise in areas that aren't under-exposed especially with the LLO.

It can happen if you dont blackshade properly...

 

[Jamon Lewis]

Nah, I'm sure I black shaded properly & according to the experts it's what's to be expected.

Looks about right for an ISO 800 REDCODE 8:1 image to me with that exposure.

 

[Phil Holland]

Good test Corey. Something I'd recommend everybody doing honestly. And if possible checking out various apertures and light levels as well. This is one of the best ways to understanding ISO ratings and texture on all cameras really.

I do remember, Dragon's "recommended" ISO range is 250-2000. The Low Light Optimized OLPF provides a bit more freedom on that front. I've managed to shoot very usable ISO 4000 and 5000 material with it. Honestly though, I mostly work in the 250-1280 range with both OLPFs.

 

[Craig Lees]

Hi Corey,

Good stuff.

I would say posting video is good as you can see the noise MOOOVE.

From these tests, it looks like the images are a lot cleaner than what they are actually are I think.

Good to see people doing tests though as which OLPF is a massive thing in your production decision making and lighting/lensing budgets obviously.

Cheers,

: )

Craig Lees
ED UK
www.3dstories.co.uk

 

[Corey Benoit]

Thanks everyone. I tend to like the skin and outdoor colors better on the STH however the low light looks to be a stop better on the LLO.

i black shaded and was at 23.98 & 1/48

 

[Craig Lees]

Yes LLO OLPF is 1.3x stops cleaner in the blacks. Vice versa for the highlights vis-a-vis Phil's tests. The blacks on lowlight OLPF are better than the Alexa (when both rated at ISO800) though the highlight retention on the Alexa is amazing still after all these years!

Horses for courses.

I think the dragon Swappable OLPF system is great for different shooting conditions as it creates a versatile approach.

Colours are still pretty darn good on the lowlight OLPF. Check out Christopher Probst's latest Taylor Swift music vid. Really great work and shot LLO OLPF.

Love the dragon sensor.

Craig Lees
ED UK
www.3dstories.co.uk

 

[Marc Wielage]

I bet it would be possible to color-time these to precisely match. Once all the images match in terms of saturation and contrast, I'd be curious as to how different they are in any other characteristics.

I have been part of camera tests before where we matched everything up and got to the point where everybody in the room got confused as to which camera we were seeing. Only when you really knew what you were looking for and used test charts were the differences really obvious. I recall one major production where they tested between the F55, Red Dragon, Alexa, and Canon C500, and they were all so close the clients opted to shoot on film.

 

[Corey Benoit]

I filmed a wedding the other day. It was dark in the reception room, i did light it decently and the ST OLPF did well. I know the LL OLPF would be much better however the dots/flares on the sensor are bad with the LL OLPF while aimed in sunlight or any light for that matter. Switching can take 10 minutes to do it properly and at a wedding thats hard to do sometimes.

Still learning the system. Here is a shot tho of the hotel where the reception was. In 6k. I also notice shooting @ 800 ISO and using ND or stopping the lens down increasing the dynamic range. I notice its much better shooting at 25-320 ISO in darker situations and using light and lenses to set exposure. REDs whole metadata and FLUT implementation is really, really good!