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 v Epic MX v Alexa Test

Status
Not open for further replies.
Marc Ruiz

Marc Ruiz

Well-known member
Joined
Aug 29, 2009
Messages
145
Reaction score
0
Points
16
Location
Detroit
So let me preface this by saying 2 things...

1) I am grateful for Stratton Camera doing the test with me on this. They supplied the Epic and the Alexa XT AND the Ultra Prime glass. We are all busy and everyone's time involved is valuable and I really appreciate their interest in this and willingness to be part of it.

2) This is a flawed test. We did not have a ton of time but we did want to know some of the improvements and comparisons. So I do not show this test to say this is the be all end all but rather here is our experience and reference. I imagine had we had more time we could have made this a pretty flawless test. Alas, I answered some questions I wanted to know and this is my effort to share this with the group.

So that said...

Dynamic Range - Dragon v Epic MX:Using the False Color as a bench mark. They hypothesis being that if Red = X% Clipping than the additional dynamic range that Red suggests the camera has would be seen by the Red in the False Color happening at a different place on the lens (referring to the stop).

When we did this using the snow we noticed the Dragon had an extra stop before it clipped to Red than the Epic MX.

Using this same hypothesis that if Purple =Y% in the shadow than the additional dynamic range that Red suggests the camera has would be seen by the the Purple in False Color happening at a different place on the lens (again referring to the stop).

When we did this using a low light situation we noticed the Dragon had an extra 1.5 stops before it clipped to Purple than the Epic MX.

Problems with this test include: Human error of when the Red and Purple was considered a stopping point. I think we were careful and our measurements represent the capabilities.

Lensing on Alexa 4x3 v Dragon:When we put on a 40mm Ultra Prime on the Dragon it was most similar to the 32mm Ultra Prime on the Alexa (Horizontally). The Alexa was slightly wider so a conservative estimate is the difference is approximately 5mm.

Dragon V. Epic MX:

In a dark room a lighter was used as the ONLY light source. Cameras were set to their maximum resolution and lowest compression available. 2000 ISO. Lenses were opened up.

Epic 5k

https://www.dropbox.com/s/sv29t9rz6w09ce3/5k.tif

Dragon 6k

https://www.dropbox.com/s/ijzyjnll2aj0n1i/6k.tif

At first glance they feel pretty close so I decided to do equal adjustment in RCX (documented on the image). 20% increase in Exposure, decrease in Saturation and light move on the curve.

Epic 5k with light RCX adjustment

https://www.dropbox.com/s/7rfxt6in19na9cu/5k - sat - level -exposure - .2 .tif

Dragon 6k with same light RCX adjustments

https://www.dropbox.com/s/bh9rrx61ttv6dw3/6k - sat - level -exposure - .2.tif

Upon seeing this it became clear to me the Dragon footage was holding up better. There is obviously some room for user error argument.

So I decided to push them all a little harder in RCX adjusting Exposure up 50% or half a stop and saturation down .5.

5k with heavy RCX adjustment

https://www.dropbox.com/s/6u2jsy3pmj0hkh9/5k - sat - level -exposure - .5 .tif.tif

6k with same heavy RCX adjustments

https://www.dropbox.com/s/bjx3695kt3139up/6k - sat - level -exposure - .5 .tif.tif

The Dragon seems to hold up in stride while the Epic clearly falls apart in my opinion.

So I grabbed this screenshot from the Alexa footage which was shot with the same ISO, Shutter, FPS.

https://www.dropbox.com/s/20sffq3kqbesg1m/Alexa - Screenshot.png

This is not the most fair test to the Alexa and I will be the first on to say that. 1) This is a screen grab instead of a Tif and 2) It is a screen grab of the APR not the Raw.

That said, I have to say that while the Alexa at 2000 is pretty unusable and noisy it seems to pick up some extra detail in the shadow.

Again, I will be the first to say it is an imperfect test. But...I think it gives a lot of data and I thought it would be fun to share here with the group. With more time it would have been nice to do tests with more dynamic lighting schemes, multiple ISO settings like 800 Alexa v 800 Dragon and 800 Alexa v 2000 Dragon but time was limited. Perhaps next time should time allow.
 

Attachments

  • 1604821_10151874462385773_1281358900_n.jpg
    1604821_10151874462385773_1281358900_n.jpg
    74.4 KB · Views: 1
Last edited:
Honestly, I have to check what view mode the camera was in to confirm that.
 
It was not running the newest firmware that just came out. However, I will confirm the exact firmware shortly.
 
I just tried your links Marc, and only the very first …5K.tif and the very last Alexa Screenshot links are working. Dropbox is saying that the files were either moved or deleted.

Could you double-check the rest on your end? I'd love to see your results.
 
So they are...

The first 6k one should be catching up shortly - everything else should work now...
 
I gotta tend to some budgets, prep and those pesky things that pay for the toys :)

but I'll see what I can do tmrw or later this week to upload more and the r3d's - I can say I unfortunately I do not have the arri raw though.
 
Why did you post screen grab of Alexa and not the tiff?
 
The Alexa footage was a MOV file and I would have had to take through a non-linear system to isolate the TIFF etc. I figured for the time being it was a decent representation. I was also quick to note this as an error of the test and presentation. To be honest the footage looks a lot more noisy with more detail.
 
Marc Ruiz said:
The Alexa footage was a MOV file and I would have had to take through a non-linear system to isolate the TIFF etc. I figured for the time being it was a decent representation. I was also quick to note this as an error of the test and presentation. To be honest the footage looks a lot more noisy with more detail.
Click to expand...



So you just shot ProRes on the ALEXA not raw?
 
Be aware that with the Dragon "Exposure" tool false color overlay in 'RAW' viewing mode, red false color may be seen prior to clipping. I noticed this the other day when the false color was showing red on some specular highlights but none of the stop lights were lit for any color channel and there was no clip indicated in the left-hand goal post.

Either this means that red false color may now indicate, for example, "90%+ white up to clipped", or the firmware isn't yet tuned precisely for exposure tools for Dragon.
 
Dragon at Clairmont camera event

Dragon at Clairmont camera event

I had a chance to play with another Dragonized Epic at a Clairmont camera event with a Cooke 5i 100mm wide open at T1.4. It was nice to see images on the 9" touch LCD. The false color shifted drastically upon changing the ISO from 2000 to 320. The same was true for the histogram. I was previewing in RAW and toggling back between RAW and RG3. ISO 320 barely produced any noticeable differences in luminance between RAW and RG3 but vastly different histogram and false color. What am I missing? What are we basing the exposure off? Clairmont's proprietary module(prototype) for Epic is one of the best I have seen till date. I wish I could purchase one for my camera.
 
Sulekh Suman said:
I had a chance to play with another Dragonized Epic at a Clairmont camera event with a Cooke 5i 100mm wide open at T1.4. It was nice to see images on the 9" touch LCD. The false color shifted drastically upon changing the ISO from 2000 to 320. The same was true for the histogram. I was previewing in RAW and toggling back between RAW and RG3. ISO 320 barely produced any noticeable differences in luminance between RAW and RG3 but vastly different histogram and false color. What am I missing? What are we basing the exposure off? Clairmont's proprietary module(prototype) for Epic is one of the best I have seen till date. I wish I could purchase one for my camera.
Click to expand...

any photos of that module?
 
Status
Not open for further replies.
Back
Top