RED KOMODO my thought :)

[rand thompson]

I kinda like grading the Komodo footage over the Monstro which surprised me!

Komodo



Monstro





Komodo



Monstro





Everytime I tried to correct the Komodo equivalent of this Monstro scene it crashed Resolve. It was like Resolve just said fuck it .


Monstro






Monstro






Monstro Recovered low light scenes the best


Monstro






Komodo

[M Harvey]

Thanks for sharing your testing, Luigi. It's helpful for evaluating the camera, and it also looks great.

[luigivaltulini]

Thanks for sharing your testing, Luigi. It's helpful for evaluating the camera, and it also looks great.

Welcome ;)

Thanks for posting this, Luigi.

Your continued contributions to this forum are awesome!!!

:)

Thanks to you Martin ;)

[Les Hillis]

Had another look at those outdoor frames.

Proper exposure on the subject, with unrecoverable highlights -





With recoverable highlights, but underexposed subject -





With recoverable highlights, but letting them go a bit to lift up the subject (not something I'd really want to do) -





Again, this is just done with a single node in Resolve, with no qualifiers or anything, and with all the tweaking done to get the Komodo image closer to the ALEXA.

I'm guessing they could be matched more accurately, but that's beyond me.

At the same time, I think there's something to be said for respecting the way each developer has decided to go about capturing and rendering the data off their sensors. Even at that level there are creative decisions and interpretations that have to be made, and I for one really like what the team at RED have come up with.

[Karim D. Ghantous]

Which is all just to say, I think the extra highlight retention you get with the Alexa isn't that huge in reality. It's there, but if you're playing in that area you're probably going to be relying on more than that to save, correct or protect your highlights anyway, so...does it really make that much of difference?

I'm amazed at how the Komodo is performing. I mean, think about it: the sensor is half the size, roughly, of the LF's, and has 50% more resolution. But in some ways it's not surprising. Red hasn't stood still, while ARRI has.

[Les Hillis]

I doubt ARRI have stood still behind the scenes.

It'll be interesting to see what their next sensor and RED's DSMC3 turn out to be.

[David Rasberry]

Image sample grade from RCX.

[Les Hillis]

Looks like the Komodo and Monstro have a different way of dealing with those Orange/Yellow Sodium(?) lights in the background -





Either way looks alright to me, crap lighting is crap lighting, but matching the cameras might not be so easy.

[Blair S. Paulsen]

IMO, the key challenge of turning a Bayer data set into balanced RGB is dealing with green. Each platform uses different techniques to extract as much image information as they can from the green channel, without letting it run wild. Certain lighting sources have such weird spectral distribution that even sophisticated formulae for linear RGB representation can go a bit sideways. Based solely on posted images viewed on an MBP, it appears that the Komodo deals with green skew a bit better than Monstro. I suspect it's a combination of sensor tech and Graeme's ongoing refinement of his core color science algorithms.

FWIW, seeing the comparos, I was struck by what appears to be a more modulated gamma characteristic just above the mid tones on Alexa. There are certainly a number of potential reasons for that perceptual difference, so I wouldn't draw any conclusions. That said, it would be fun to have a play in a proper color suite.

Cheers - #19

[Aaron Lochert]

Looks like the Komodo and Monstro have a different way of dealing with those Orange/Yellow Sodium(?) lights in the background

Komodo's OLPF is said to be skin-tone highlight based and these shots were done with the standard OLPF on Komodo. I'd be curious to see the test done again with the STH OLPF on Monstro before making any judgments on sodium vapor rendering.