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.

 

[Jake Wilganowski]

Komodo overall is way more magenta than any red cam

 

[Les Hillis]

Was just an observation. Didn't think to look at what OLPF the Monstro used.

Curious too what difference it makes.

 

[Aaron Lochert]

Komodo overall is way more magenta than any red cam

I noticed that too. With other sensors, I'm used to dialing in a negative tint value where Komodo I sometimes need to dial in a positive tint value. But an overall tint difference is easily correctable with the WB controls. The thing with the sodium vapor stuff is it's less of a simple white balance thing and more of a frequency response thing. When you match the two against one light source, another goes awry.

Komodo

Monstro (standard OLPF)

These were both roughly matched for the foreground grass via WB and Tint. You can see how different the street lighting is coming across. If you try and match the street colors, you get this:

Where you will notice that matching the color of the road has unmatched the colors in the foreground grass and the far background lights to a purple tint.

This is why I am curious to see Monstro with the STH OLPF in the same conditions to see if it's a frequency response difference at the OLPF level or the sensor level.

 

[Les Hillis]

This Monstro image also raises other questions -



1. Why does that road, complete with lane markings, lead straight into a concrete wall?
2. How exactly does a bus get in and out of that bus stop?
3. Why are there at least six man-hole covers all in a row on top of that embankment?
4. Would it be inappropriate to call the 'Sale' number written on that garage door in the center of frame and see if they have the answers to these questions?

 

[yann LHENORET]

;-)

This Monstro image also raises other questions -



1. Why does that road, complete with lane markings, lead straight into a concrete wall?
2. How exactly does a bus get in and out of that bus stop?
3. Why are there at least six man-hole covers all in a row on top of that embankment?
4. Would it be inappropriate to call the 'Sale' number written on that garage door in the center of frame and see if they have the answers to these questions?

 

[rand thompson]

It's an overhang from another building or a higher overpass road. It hovers high above and far away from the road the photo gives the illusion that it's on the road.

 

[Christopher A. Bell]

This Monstro image also raises other questions -



1. Why does that road, complete with lane markings, lead straight into a concrete wall?
2. How exactly does a bus get in and out of that bus stop?
3. Why are there at least six man-hole covers all in a row on top of that embankment?
4. Would it be inappropriate to call the 'Sale' number written on that garage door in the center of frame and see if they have the answers to these questions?

And why does a still image have an appearing/disappearing chevron pattern on the fence(3 second intervals)(RMBP, Safari 12.1.2)(not scrolling the image/page)?

 

[Jon Dishler]

It is a gif - something odd here.

 

[Les Hillis]

Was just joking around, don't want to de-rail the thread.

I just thought it was a funny/weird bit of urban planning where they just went and built something on top of an existing road and left it at that (could be part of that train station seen in the other shots).

It can happen sometimes, like when you really needed a purpose-built training ground for man-hole digging and covering and have nowhere else to put it.

 

[Sabovic Adis]

Lemme see:

This Monstro image also raises other questions -



1. Why does that road, complete with lane markings, lead straight into a concrete wall?

They ran out of bricks so they put concrete wall instead. Gotta build walls!

2. How exactly does a bus get in and out of that bus stop?

It's mini-bus stop.

3. Why are there at least six man-hole covers all in a row on top of that embankment?

Six? Odd number in this case. Aliens?

4. Would it be inappropriate to call the 'Sale' number written on that garage door in the center of frame and see if they have the answers to these questions?

Luigi?
:biggrin: