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 :)

Dragon/Epic comparison

James,

First, a few definitions.

Dynamic Range = the full range of light-intensities a sensor can record, from value 1000 (clip) to value 0 (black)
That is the total "potential. (the numbers could have been any range of numbers, they are just illustrations. I could have spoken in f-stops)

Let us for this example say that the Dragon can record the full range of intensities, while the Epic clips at 900 and goes mushy at 200


Let us say we have an outdoor scene with snow, a dark doorway with a person in and that you want to expose for that persons face.

Snow has a reflective level of 1000, the doorway is at 200 and the persons face is at 500 where you would want it.

Let us say that you would like to expose for that persons face, but not clip the snow, and preferably not make the hallway black.

Get it?

Also, if you can avoid NDs in your signal path, it does not hurt to leave them off...

I will try to take a quick shot A/B of this scene.

Problem with slapping in an ND, is that the face gets overexposed and the doorway gets too dark.
Problem with slapping in ND and increase ISO, is that black is still black, and you increase noise.
Solution 3 is to "bring in the lighting crew and the genny"
Solution 4, could easily be the Dragon.

That is 1 scenario where the highlits are of importance.
Second is noise.

Let us say that the Dragon has a stop more highlight tolerance. And that colors look good all the way "up" (which is not common in digital cameras)

THEN I could possibly shoot at 400 ISO or possibly 200 ISO, and get a LOT cleaner image and smoother black.

Sure.
I would have used less of the dynamic total capacity, but I would have made a "smooth image" priority with what I have.

On the Epic, I would rarely do that.

On the Dragon, I see that as a very relevant option.


I will try to illustrate soonish...
 
James_Mills said:
I wasn't talking about shooting at 4K on an Epic (I would never do this either and have urged directors not to when they asked me if we could on my Epic). I meant that, as far as I know, the actual 5K image from the Epic is the same FOV as the Scarlet's 4K. You said that it has a wider field of view and that we are thus zoomed out from the noise, but I don't believe this is true?
Click to expand...

Yes James, it's true. The MX sensor is LARGER THAN Super 35. When you shoot 5K on Epic, you are using the ENTIRE SENSOR. The Scarlet has the same sensor, but not the horsepower to use it all, so instead you shoot cropped in at 4K (for 23.9 and 29.97).
 
Actually, I would love to see some grades and comments from the three comparisons I have posted...

I know they can all easilly be matched.

Point it is if you stretch them a bit in dynamics and colors...

Bring them on.

(the other two batches of R3Ds are on the frontpage...)

Have fun

G
 
Gunleik Groven said:
Actually, I would love to see some grades and comments from the three comparisons I have posted...

I know they can all easilly be matched.

Point it is if you stretch them a bit in dynamics and colors...

Bring them on.

(the other two batches of R3Ds are on the frontpage...)

Have fun

G
Click to expand...

EDIT...Removed for investigation.
 
Thanks Brice!

I will take that last example offline as I get very conflicting results with that.

Not conflicting as to "what I expect" but just very conflicting.

Must investigate a bit!

Thx!
 
Robert Ruffo New said:
This same producer will demand Dragon, yes, and again I will agree with him/her (within reason of cost difference, right now at 2K a day it's too much for many prods).
Click to expand...

Is that what most are billing for Dragon? With how complete a kit? Sounds like I'm too cheap ;)
 
Gunleik Groven said:
Thanks Brice!

I will take that last example offline as I get very conflicting results with that.

Not conflicting as to "what I expect" but just very conflicting.

Must investigate a bit!

Thx!
Click to expand...

Gunleik thanks for providing so much explaining about dragon. I already got it and I am saving. I was very picky about noise improvement but after this dynamic range examples and being a very happy Scarlet owner, I am already saving for the upgrade to dragon. dR make such a difference that my doubts about low light noise performance is no longer as important as it was before. Also because I know that it does perform better with noise and we have 6k to down sample and get better image overall.

Thanks again to you! Phil and the rest of you putting lots of time to explain and share
Alain
 
Gunleik Groven said:
James,

First, a few definitions.

Dynamic Range = the full range of light-intensities a sensor can record, from value 1000 (clip) to value 0 (black)
That is the total "potential. (the numbers could have been any range of numbers, they are just illustrations. I could have spoken in f-stops)

Let us for this example say that the Dragon can record the full range of intensities, while the Epic clips at 900 and goes mushy at 200


Let us say we have an outdoor scene with snow, a dark doorway with a person in and that you want to expose for that persons face.

Snow has a reflective level of 1000, the doorway is at 200 and the persons face is at 500 where you would want it.

Let us say that you would like to expose for that persons face, but not clip the snow, and preferably not make the hallway black.

Get it?

Also, if you can avoid NDs in your signal path, it does not hurt to leave them off...

I will try to take a quick shot A/B of this scene.

Problem with slapping in an ND, is that the face gets overexposed and the doorway gets too dark.
Problem with slapping in ND and increase ISO, is that black is still black, and you increase noise.
Solution 3 is to "bring in the lighting crew and the genny"
Solution 4, could easily be the Dragon.

That is 1 scenario where the highlits are of importance.
Second is noise.

Let us say that the Dragon has a stop more highlight tolerance. And that colors look good all the way "up" (which is not common in digital cameras)

THEN I could possibly shoot at 400 ISO or possibly 200 ISO, and get a LOT cleaner image and smoother black.

Sure.
I would have used less of the dynamic total capacity, but I would have made a "smooth image" priority with what I have.

On the Epic, I would rarely do that.

On the Dragon, I see that as a very relevant option.


I will try to illustrate soonish...
Click to expand...

I'm well aware of all this, I was simply stating that this test isn't the best for comparing the usable DR differences between sensors (opposed to the great example you gave in this post). These tests are great in other regards, I'd just like to see one where a wider range is evident for both without having to over and underexpose.
 
OK.

I kind of disagree, and gere is why.

Actually I think my first example here is key.

It shows that you can expose higher in a HDR situation, and thus give skintones etc priority and still be able to protect highlights., and thus get significantly less noise than you would get in a similar situation with the Epic.

Unfortunately I do not have an Epic around right now, and also not the Dragon for a few days, but I think I have figured the perfect setup to show the difference and its consequences.

Add the higher bit/color precision, and I think I have figured a logical (and not emotional) reason I and many others are so excited about working with it.

I KNOW I feel it, I have been a bit unsure as to "why exactly", but I think I finally got it, and THAT will make it easier for me to make better images with the Dragon...

STill the crap in - crap out rule applies AND as I said again and again back in August.
The Dragon needs a bit of contrast to shine.

But THEN it shines.

Main point is:
You should not expose the Dragon like you would expose the Epic... :)

So I guess my next batcch of tests will show off the same scene, but exposed for the Dragon and exposed for the Epic.

First batch I was religious in trying to give them the same amount of light to see the differences.

next batch will be all about exploring the differences.
I think that is where all the fun I have practically seen comes from.
That you treat them differently.
It is after all a whole new cam.

They can and should be exposed in different manners.
That is what opens up the pandoras box of new fun.


But right now I have neither cam available...


Best

G
 
Gunleik Groven said:
OK.
So I guess my next batcch of tests will show off the same scene, but exposed for the Dragon and exposed for the Epic.

First batch I was religious in trying to give them the same amount of light to see the differences.

next batch will be all about exploring the differences.
I think that is where all the fun I have practically seen comes from.
That you treat them differently.
Best
G
Click to expand...

Wich make me think you probably under expose the dragon shot in the example you removed.
;-)
Anyway my guts are telling me the dragon sensor will not need that much lowcon filtering anymore.
 
Brice Ansel said:
Wich make me think you probably under expose the Epîc shot in the example you removed.
;-)
Anyway my guts are telling me the dragon sensor will not need that much lowcon filtering anymore.
Click to expand...


It does not

I removed that example because I was not sure I did not mess it up.
First I thought it was a RCX bug, then other things.
I do not like to be unsure about the stuff I post. (I can be wrong, but I do not want to doubt what I say myself, if I make a statement. THAT image was a very strong statement. If it was misleading, it should not be used... I will check for that scenario again later)

G
 
Gunleik Groven said:
They are not the same, and should thus not be exposed the same...
(unless you need them to match...)
Click to expand...



Make sense.
This is why they were easy to match then.
 
Dunno yet. More than one. I hope more than 2, but even one would make a huge difference in noise and tonality. and shadow renderition.

Actually I overexposed the Epic and underexposed the Dragon, which made them match fairly well...

Next test would be to expose both within range on the same subject and lights.

And I kinda know what I will see then...
 
You must log in or register to reply here.
Back
Top