All the Epic/Dragon Vs. Alexa/Sony threads are belong to us...

[Sanjin Jukic]

Apparent contrast could also be due to the gamma.

I guess that there is better ways to check "actual" contrast than RG3 vs Alexa REC709.

Without bringing up the hornets nest of "how do they make the Alexa the way it is", which in many ways is irrelevant because it is what it is, I think that there are many other factors to a good image than pure DR, and besides having good DR, the Dragon image checks on many points I like. As mentioned a few times, I would have taken the upgrade for the image-characteristics alone, even with equal DR and resolution to the Epic.

Dragon is a new camera, and should be treated as that, all the way from set to post to analytics.

For any Alexa vs. Dragon comparison test

a film look image evaluation should match ISO alignment

by ratio 1:2 or even better at maximum ratio 1:2.5


What does it mean?

It means for indoor shots especially when Alexa has set @ ISO 800,

Dragon should be set in a range between ISO 1600 and ISO 2000...or so...

LINK>>>>

Or for easy understanding if you would shoot tungsten with a film stock @ 500 ASA

and if you would exchange that film camera with Alexa you should use it @ ISO 1000

and if you would like to use Dragon for that particular shot you should set ISO in a range between ISO 1600 and ISO 2000.

That means if you have Alexa and Dragon for a test you can't have them both @ the same ISO.

The ratio is 1:2 or even better at maximum the ratio is 1:2.5.

Keep this in mind!!!!

 

[luigivaltulini]

For any Alexa vs. Dragon comparison test

a film look image evaluation should match ISO alignment

by ratio 1:2 or even better at maximum ratio 1:2.5


What does it mean?

It means for indoor shots especially when Alexa has set @ ISO 800,

Dragon should be set in a range between ISO 1600 and ISO 2000...or so...

LINK>>>>

Or for easy understanding if you would shoot tungsten with a film stock @ 500 ASA

and if you would exchange that film camera with Alexa you should use it @ ISO 800

and if you would like to use Dragon for that particular shot you should set ISO in a range between ISO 1280 and ISO 1600.

That means if you have Alexa and Dragon for a test you can't get set them bot @ the same ISO.

The ratio is 1:2 or even better at maximum the ratio is 1:2.5.

Keep this in mind!!!!

Sanjin,
why do you say that should be set to 1600 Dragon?
Last week I made ​​of small testing DRagon with Alexa, medium gray 18% and both 800 iso, were perfectly aligned with the same aperture of the lens.
Did I miss something?
Damn, I deleted the files for you to show: (

 

[Sanjin Jukic]

Sanjin,
why do you say that should be set to 1600 Dragon?
Last week I made ​​of small testing DRagon with Alexa, medium gray 18% and both 800 iso, were perfectly aligned with the same aperture of the lens.
Did I miss something?
Damn, I deleted the files for you to show: (

Luigi is a quote from my writing at the tread that was already closed here>>>>>

http://www.reduser.net/forum/showth...3Ds-amp-MORE&p=1305398&viewfull=1#post1305398

Also I read somewhere on the internet from Dan Hudgins that:

"Different cameras types having different saturation points and that to compare cameras you can get

comparable results on the tonal level by finding their saturation ISO, then aligning the ISO settings/exposure so that they align.

Approximate Table to align the saturation points of the cameras in this test for ISO 250 color negative film,

Film=Arri=Red

4=25=100
8=50=200
16=100=400
32=200=800
64=400=1600
125=800=3200
[250=1600=6400]
500=3200=12800"

So what does it mean here for this test?

It means that when Alexa is exposed @ ISO 800 than # Dragon should be approximately exposed @ ISO 1600

that those two images can be compared and matched equally.

(# That means Dragon @ ISO 1600 should show about the same sensitivity as Epic X @ ISO 3200)


To conclude is to say that when Alexa is exposed @ ISO 800

than Dragon should be exposed in the range between ISO 1280 and ISO 1600

that these two cameras could be compared on approximately equal basis.

 

[Jean Maurizio]

So how did the Epic- the best camera in the world few months ago , just become SD camera like DVX? all of the sudden?
you guys are funny- bashing your child that was your favorite just a year ago...

Hello Jacek,

I think my statement was misunderstood.
I'm not bashing anything here.
It's obvious that, not matter what camera you have, also a 50 megapixel raw from an Hasselblad, if you wash out all the details with neat video, you lose all the definition.
I prefere some noise, and i can add some grain on it, instead of using neat video heavily, to transform a 5-6k image into something with not details at all. DVX100 was just an example/paradox.

Cheers

Jean

 

[David Mullen ASC]

I didn't understand what Dan Hudgins was talking about. I still don't re-reading it. If he's talking about clip point, he's assuming at it's the same at 1600 ISO for Dragon as it is for 800 ISO on the Alexa, meaning he's assuming that it has a stop less overexposure latitude at 800 ISO compared to the Alexa, which may be true for the MX sensor but how does he know it is true for Dragon??? And is he assuming that the noise at 1600 ISO on the Dragon is the same as it is at 800 ISO on the Alexa? How does he know that either?

 

[Gunleik Groven]

I didn't understand what Dan Hudgins was talking about. I still don't re-reading it. If he's talking about clip point, he's assuming at it's the same at 1600 ISO for Dragon as it is for 800 ISO on the Alexa, meaning he's assuming that it has a stop less overexposure latitude at 800 ISO compared to the Alexa, which may be true for the MX sensor but how does he know it is true for Dragon??? And is he assuming that the noise at 1600 ISO on the Dragon is the same as it is at 800 ISO on the Alexa? How does he know that either?

Hm...

I am not sure exactly how the Dragon and Alexa compares in the top yet.

MX and Dragon has a bit if a difference, though.

12 frames from both...:
http://www.gunleik.com/dragonepic/r3d_01.zip

 

[Evin Grant]

It means that when Alexa is exposed @ ISO 800 than # Dragon should be approximately exposed @ ISO 1600

I disagree with this since when I shoot on Alexas I always have to underexpose by 2/3 of a stop to get the skin tones to have the right tonal balance, otherwise they look quite bright and the facial highlights are quite distracting. I have never needed to do this with the Epic and so far the Dragon seems to behave similarly, I can bring a skin tone to 50% and it still looks good, while I have to keep the highlight at 40ish on an Alexa (using a leader waveform). Granted some of this is because we are almost always using the 709 curve on the Alexa but still, it seems to me that this is just a difference is how the curve is applied.
There are many philosophies on how to properly expose, regardless of camera. The important thing is to test and evaluate which settings work best for you particular project or aesthetic goal.

 

[Sanjin Jukic]

What Dan Hudgins has said I just implemented on the test shots using RAW files from both cameras, Dragon and Epic.

If you don't believe me what I'm talking about than try Offhollywood test shots for yourself.

Below is the quote and the link from the findings what I've had:

For any Alexa = Dragon comparison test

a film look image evaluation should match ISO alignment

by ratio 1:2 or even better at maximum ratio 1:2.5



What does it mean?

It means for indoor shots especially when Alexa has set @ ISO 800,

Dragon should be set in a range between ISO 1600 and ISO 2000.



Below is an absolute matching @ 2K that is now ready for more precise colour grade later.

Alexa original 2K size and Dragon @ ISO 2000, de-noised, softened, de-saturated (-20%), cropped and resized in 2K.

A film look applied: Dragon R3D, ISO 2000, K 3200, Aperture T/8, PS6 Color Effex, Kodak Elite Chrome 200, Dragon, de-noised, softened, de-saturated (-20%), cropped and resized in 2K

A film look applied: Alexa RAW, ISO 800, K 3200, Aperture T/8, PS6 Color Effex, Kodak Elite Chrome 200

Right-clicking opens an original size image in a new tab!!!

Comparison by dynamic range between Alexa and Dragon should show that Dragon has one more stops of DR in shadows

and also one more stop of DR in highlights.

You can see these additional DR stops clearly at nicer highlights roll off in the face of a guy

and with more detail in the background at the shadow with the lamp.

That means if Alexa claims 14 stops of DR than Dragon should claim about 16 stops of DR.

I think that Jim was totally right when he was writing about Dragon's 16 stops of DR here at the forum a couple months ago.

So with this small and simple test between Alexa and Dragon,

Dragon showed that is a leader in digital cinema camera market today.

Dragon can easily simulate and evaluate any of Alexa shot with 100% mach but vice versa

Alexa can't simulate and evaluate Dragon just because of less dynamic range and lack of more spatial resolution.

http://www.reduser.net/forum/showthread.php?109981-OFFHOLLYWOOD-DRAGON-R3Ds-amp-MORE&p=1306148&viewfull=1#post1306148

 

[Gunleik Groven]

Signal chains are tricky things...

 

[luigivaltulini]

so if you shoot with Alexa, Dragon must be set to 1 stop higher, right?
But why this huge difference between Arri and RED?
So to a DP that makes the lights for 800 iso in fact you should put 1600 on Dragon and Alexa 800?
But wait, maybe stupid question, but ISO 800 should not be the same for everyone?

 

[Gunleik Groven]

so if you shoot with Alexa, Dragon must be set to 1 stop higher, right?
But why this huge difference between Arri and RED?
So to a DP that makes the lights for 800 iso in fact you should put 1600 on Dragon and Alexa 800?
But wait, maybe stupid question, but ISO 800 should not be the same for everyone?

I'd be pretty surprised if that was the situation...

ISO 800 is probably very similar for middle gray at any given light. As it should be.

 

[Evin Grant]

So far my incident meter and an 18% grey card have matched perfectly so I think you need to explain yourself better Sanjin.

If 18% grey=50% brightness on the Dragon, Alexa and an incident meter then how exactly is shooting at ISO 1600 on the Dragon the same as 800 on an Alexa?
Are you saying that to match the same clip level? Cause, the Dragon will clip at the same point weather you shoot EI 250 or 2500, it's just how you distribute the stops in between.
If what you are saying is that the Dragon has an equal number of stops above and below middle grey at EI 1600 that may be the case, I haven't tested that. However that still doesn't explain how your matching scenario with 800 Alexa=1600 Dragon.

 

[Sanjin Jukic]

So far my incident meter and an 18% grey card have matched perfectly so I think you need to explain yourself better Sanjin.

If 18% grey=50% brightness on the Dragon, Alexa and an incident meter then how exactly is shooting at ISO 1600 on the Dragon the same as 800 on an Alexa?
Are you saying that to match the same clip level? Cause, the Dragon will clip at the same point weather you shoot EI 250 or 2500, it's just how you distribute the stops in between.
If what you are saying is that the Dragon has an equal number of stops above and below middle grey at EI 1600 that may be the case, I haven't tested that. However that still doesn't explain how your matching scenario with 800 Alexa=1600 Dragon.

Evin, I did not make Offhollywood Dragon vs Alexa test.

I've downloaded files only.

Then I wanted to match the shots with the same ISO.

It didn't work as I expected.

Then I raised ISO on Dragon shots.

Almost simultaneously Dan Hudgins has wrote something similar on CML.

So I tried ones more and it seems to me that worked.

Btw. ISO setting doesn't anything directly to do with DR.

The latest film stocks have 14 sops of DR guaranteed.

Why film stock with 500 ASA (or ISO) beats any digital in the highlights rolloff?

There is no film stock 800 ASA or even 1600 ASA still available.

Also ISO setting doesn't mean anything in terms of DR,

for example Canon C500 can be dialled to (fantastic!!!) ISO 80.000 at the maximum but still has about 12 stops of DR.

I'll make a diagram about that when I'm back home later. Stay tuned!!!!

Anyway, but what I discovered is that Dragon slightly beats Alexa in both, shadows and highlights.

How much I can't tell you but it is a great technological achievement for 6K digital CINE (CMOS) sensor.

Dragon needs a new colour science that should be set and sorted in a right CINE-matic way.

And that's it.

 

[Evin Grant]

I know it doesn't change the DR, and I'm pretty sure Mark and Off Hollywood know how to operate both cameras. I also do not have the time currently to download the files but I'm curious what your diagnosis is as to what might be causing this peculiar difference between the exposures of the cameras? I can certainly tell you that the Dragon is not underrated by 1 stop, at least if middle gray is any indication.

(Edit) So after really diving into the thread I'm pretty sure the perceived exposure discrepancy comes from the difference in the curves between Log-c and Redlogfilm.

 

[Kirk.Attard]

I agree Jacek. I dont see much point in pushing the Epic to 3200 iso. I know its just a test but its like trying to take an f-18 in the stratosphere. Its a great jet but not meant to fly at that altitude. Its pointless trying to test it that way. I am so happy I did the upgrade from Scarlet to Epic. I know my cameras limitations and I try to work around them the best I can. I dont like the denoising either so I do my best to expose as best as I could. Then again I don't do much run n gun. Hardly any.

Still I wish I had a Dragon for lowlight but I will never feel limited to shoot in any situation because I 'just' got an Epic.

What I think is that at this point we should try to push the Dragon to see what its capable of but its of little use trying to break the images from Epic. HAving said this I still appreciate and read all the post concerning test because people invest time and effort in doing so and almost always they offer good insight.

 

[Sanjin Jukic]

I know it doesn't change the DR, and I'm pretty sure Mark and Off Hollywood know how to operate both cameras. I also do not have the time currently to download the files but I'm curious what your diagnosis is as to what might be causing this peculiar difference between the exposures of the cameras? I can certainly tell you that the Dragon is not underrated by 1 stop, at least if middle gray is any indication.

(Edit) So after really diving into the thread I'm pretty sure the perceived exposure discrepancy comes from the difference in the curves between Log-c and Redlogfilm.

At the moment I can't say anything more than to show how it looks in REDCINE X PRO 22 and ARRIRAW Converter 3 also with the full size outputs from both apps in jpg>>>>

Dragon, ISO 1600, T8, RED Log Film, REDCINE X PRO 22
Use the right mouse button and choose Open Link in New Tab to see the full image!

Alexa, ISO 800, T8, Arri Log C Wide Gamut, ARRIRAW Converter 3
Use the right mouse button and choose Open Link in New Tab to see the full image!

Dragon, ISO 800, T8, RED Log Film, REDCINE X PRO 22
Use the right mouse button and choose Open Link in New Tab to see the full image!

Dragon, ISO 1600, T8, RED Log Film output
Use the right mouse button and choose Open Link in New Tab to see the full image!

Alexa, ISO 800, T8, Arri Log C Wide Gamut output
Use the right mouse button and choose Open Link in New Tab to see the full image!

Dragon, ISO 800, T8, RED Log Film output
Use the right mouse button and choose Open Link in New Tab to see the full image!



Now at ISO 1600 Alexa is getting pretty noisy!!!!

Alexa, ISO 1600, T8, Arri Log C Film, ARRIRAW Converter 3
Use the right mouse button and choose Open Link in New Tab to see the full image!

Alexa, ISO 1600, T8, Arri Log C Film output
Use the right mouse button and choose Open Link in New Tab to see the full image!



Dynamic Range is maybe very similar to both cameras by counting the stops (14) but Dragon show more flexibility (+++) and spatial resolution (+++) !!!

Right now RED should finally fix Dragon Colour Science and that's it!!!

The future belongs to Dragon!!!

Alexa should be the past!!!

But we know from the real life situation how difficult is to fight conservative opinion and behaviour...!!!

Just saying it!!!

 

[luigivaltulini]

thanks for testing Sanjin

 

[luigivaltulini]

ok guys ..
yesterday evening after a day of shooting, I stopped in the return to the city for a moment.
It was late .. Epic and Dragon I had with me, I did some shots together ..
Sorry for the focus is not perfect, but I was really tired, and I'm sorry but at some point it was so cold that the Epic has gone outside temperature. Maybe the black were affected.
However, I like the Carbon, in 5 minutes is ready and perfect temperature wowo, does not suffer like my Epic, even after 15 minutes the temperature was .. well maybe because last night was -3 degrees celsus. However, the carcass Carbon insulate well, I hope he did not mind this summer.
I made these last shots because my Epic 599 left this morning to upgrade to Dragon.
I used a Canon 24mm.
So you can play with a bit of nightlife files from the road.


Matt to you, the first grade hahaha

https://next-d.copy.com/qoHiO3YVWHrT

 

[Niko Pueringer]

Hi Luigi,

Your archive seems to be corrupted. I'm unable to extract half the files. Tried downloading twice!

 

[Sanjin Jukic]

Hi Luigi,

Your archive seems to be corrupted. I'm unable to extract half the files. Tried downloading twice!

Everything is fine the files, I've got them all OK!!!