Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

My Dragon Thoughts

You guys are having all the fun!


Christopher Probst said:
BTW, I've handed over some files to Phil to poke around with. Maybe we can get some better uploads going. I'm still out of country and have no time to do the leg work.
Click to expand...

I'll put something together tomorrow with these.

Some initial observations from these grabs show just how much more is held in the highlights. It's pretty obvious here.

Also, this is a good test to showcase just how differently Dragon sees color than Mysterium-X. Noticeable in all channels actually.


John Marchant said:
I've been not posted any comparisons like these, as some people will likely misunderstand them... but another significant factor worthy of exposition is what happens at and beyond the point of clip on Dragon vs MX.

Viewing in linear light is a good way to check it out.


With apologies to my long suffering editor Holly...
Click to expand...

That's a fantastic test John. Something that a few people haven't noticed is that there currently isn't a DRX feature in Redcine-X Pro for Dragon footage and this pretty shows why you don't need to worry about that.

You can see the more "flavor packed" highlights with the burning out effect. I dig that image character a lot. Mysterium-X burns to clipping pretty earlier in comparison, but Dragon's got a whole lot more detail in there. That's the dynamic range doing it's thing. Also, some of that over-saturated color near the clipping point is held through and smoothly rolled off. This is actually effecting the skin tones in a positive way here as there isn't that sort of jagged gradient, but rather a smooth one from the clipping point into the saturated core and I can see more purity into the subsurface scattering across the skin on her face. Very cool.

Another observation. It looks like the new OLPF in Dragon and likely whatever else is going on in the new sensor box area is protecting a bit more contrast under these conditions. Incredibly subtle, but an interesting thing to note.
 
Nice... but we are almost in November and just a handful of Red test cameras is out there. Meanwhile, F5´s and F55´s are becoming increasingly (and justifiably) numerous and popular, ARRI et al are for sure working silently in new and better sensors and gear... I would greatly appreciate some news on the final mass upgrade of Epiics.
Guido
 
The mechanical detail of the new OLPF , sensor box etc. is far more intricate on Dragon. New shape, new mounting method, a beautiful little miniature mask clipped in around the edge - very smart stuff. A better seal at the back of the OLPF too - the OLPF frame meets the sensor surround with a thin adhesive strip around the edge. Zero possibility for foreign objects to make their way in.

Interestingly the OLPF is not the only place that improved IR rejection occurs - unfiltered by the OLPF the sensor makes true colour images, unlike MX which is heavily IR biased when 'bare'.

So much has changed with Dragon :)
 
THANKS Christopher!

THANKS Phil and John et al.
 
A very clever test method, with the light on a motor ! Congrats John !

Phil, can you process some r3d 's to show dark noise levels in the BG, letting the midtones blow out ?
That would be interesting to see the actual noise levels in the darks with the new sensor.
I hope a black SUV doesn't show up take you to the desert for doing this ;)
 
John Marchant said:
...

Interestingly the OLPF is not the only place that improved IR rejection occurs - unfiltered by the OLPF the sensor makes true colour images, unlike MX which is heavily IR biased when 'bare'...
Click to expand...
John, is this something a semi-competent person could do?... removing the OLPF filter, that is. And what would be benefit?... shooting at night in low light, perhaps?
 
Comparison with and without OLPF with 6K scaled to 4K might be interesting...
 
I would love to see a similar test to what Christopher Probst did on the last page, except that instead of changing ISO in post, you would compensate for the ISO changes with the T stop, and both on MX and Dragon, with R3D stills posted. So, for example:

ISO 50 - T4
ISO 100 - T5.6
ISO 200 - T8
ISO 400 - T11
ISO 800 - T16
ISO 1600 - T22

A beer goes to whoever can do this first!
 
John Marchant said:
Its not that difficult to do. You'd need a damn good reason... sharpness related rather than anything else. Don't go doing it for fun.
Click to expand...

i'm about to shoot some aerials of mountains/hills/foliage this evening... was actually considering pulling it off, what do y'all think?
 
paulherrin said:
i'm about to shoot some aerials of mountains/hills/foliage this evening... was actually considering pulling it off, what do y'all think?
Click to expand...

I don't know Paul... John was referring to the Dragon sensor and how it handles IR pollution sans OLPF. Not sure if this is advisable with the MX sensor.
 
good point. only reason i've considered it before was to shoot IR, just thinkin about whether i'd really get any artifacts in that situation. my thinking is that these shots are really hard for the compression and it might actually make it easier (or at least, end up with more detail) if it's sharper to begin with, but i dunno.
 
paulherrin said:
good point. only reason i've considered it before was to shoot IR, just thinkin about whether i'd really get any artifacts in that situation. my thinking is that these shots are really hard for the compression and it might actually make it easier (or at least, end up with more detail) if it's sharper to begin with, but i dunno.
Click to expand...

I've been curious too how the image would look without the OLPF on the MX sensor. Actually had a short written that begged for that when a character viewed everything in the IR spectrum. But only had a day's warning and a day of actors availability to shoot so that was the last thing on my mind. If I ever attempt a reshoot I just may try taking off the OLPF if that will give me true IR footage.
 
If anything, the compression is harder without the OLPF, as there is way more high frequency detail, at least in the MX version .
I shoot without it all the time, but added my own IR cutoff in it's place. You need some glass in there because your back focus is way off with just air in there.
I am guessing the Dragon OLPF has much less image spread and nay not be worthwhile to tinker with anyhow. We will see.
 
I DID do iris changes. Phil will post some of those soon.

David Kruta said:
I would love to see a similar test to what Christopher Probst did on the last page, except that instead of changing ISO in post, you would compensate for the ISO changes with the T stop, and both on MX and Dragon, with R3D stills posted. So, for example:

ISO 50 - T4
ISO 100 - T5.6
ISO 200 - T8
ISO 400 - T11
ISO 800 - T16
ISO 1600 - T22

A beer goes to whoever can do this first!
Click to expand...
 
Christopher Probst said:
I DID do iris changes. Phil will post some of those soon.
Click to expand...

Very soon. Like now soon.

cp_RedQDRTest_01_T1p4_ISO0320.jpg


Okay. Here's a look at Christopher's Quick Dynamic Range Test between Mysterium-X and Dragon. He has provided samples at T1.4, 5.6, and 22. I've made various "notable" ISO adjustments for everybody to inspect in REDcolor3 and REDgamma3.

Each link below will open a larger image in a new window.

Red Quick Dynamic Range Test - T1.4

Red Quick Dynamic Range Test - T5.6

Red Quick Dynamic Range Test - T22


Specific things to look out for are where tonal and color information are held and not lost or distorted. This is most visible in the ISO 50-320 samples. Additionally it's very clear to see Dragon's new way of seeing color overall in the ISO 800-2000 samples at T22.
 
This is what I was most curious about - the claims of a "Penalty Free" ISO adjustment. Seems like there is still a "sweet spot" (read: mandatory) ISO range to constrain yourself to as you start introducing all sorts of weird shifts at the lower end of the ISO spectrum. Looks like I wont be shooting 50 ISO in the daytime.

Thanks for these, Christopher and Phil - very informative!

Phil Holland said:
Very soon. Like now soon.

cp_RedQDRTest_01_T1p4_ISO0320.jpg


Okay. Here's a look at Christopher's Quick Dynamic Range Test between Mysterium-X and Dragon. He has provided samples at T1.4, 5.6, and 22. I've made various "notable" ISO adjustments for everybody to inspect in REDcolor3 and REDgamma3.

Each link below will open a larger image in a new window.

Red Quick Dynamic Range Test - T1.4

Red Quick Dynamic Range Test - T5.6

Red Quick Dynamic Range Test - T22


Specific things to look out for are where tonal and color information are held and not lost or distorted. This is most visible in the ISO 50-320 samples. Additionally it's very clear to see Dragon's new way of seeing color overall in the ISO 800-2000 samples at T22.
Click to expand...
 
You must log in or register to reply here.
Back
Top