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

[geoffboyle]

Cool test Ergin. It seems that the way Dragon was exposed in this test gave it more clarity in the darks at the expense of highlight protection. Any chance you'd post some images?

Anyone who wants to do their own comparisons can get all the RAW files from all the cameras at http://www.cinematography.net/UWE/index.html

Check the results for yourselves on your kit the way that you work.

I will add updated Dragon and F65 and a couple of other cameras in the future but first I need to do my "real" job!

 

[Dave Blackham]

Anyone who wants to do their own comparisons can get all the RAW files from all the cameras at http://www.cinematography.net/UWE/index.html

Check the results for yourselves on your kit the way that you work.

I will add updated Dragon and F65 and a couple of other cameras in the future but first I need to do my "real" job!

Great to see Geoff here. Thanks to all for this.

 

[Ketch Rossi]

White flag raised... and Geoff accepted. We can all be friends now.. and move along.


Hi Jared,

I agree completely.
We both overreacted, but that's who we are.
We both care passionately about what we do.
If I could have replied on Reduser to the people who were posting messages
attacking my methodology and approach before they had ever seen anything as
well as questioning my integrity we could have avoided all this.
I over-reacted and posted the only place I could, it was a surge of
frustration, I was trying to tell people how good Dragon is, in its current
form never mind later, and I was being reviled.
Please let me reply quietly and politely on Reduser in future and we can
prevent all this.
Oh and the reason that RED is listed as a sponsor is that a guy called Jim
Jannard sent me a cheque years ago that covered sponsorship up until mid
2015...
Geoff


-----Original Message-----
From: Jarred Land
Sent: 11 February 2014 03:34
To: Geoff Boyle
Subject: We need

to get past this stuff.. and I am just as guilty in this regard. It's not
good for you, for me, or any of our readers/customers.

I'm glad to be back here, thanks Jared.
Maybe we can now talk about this directly and calmly.

The tests were conducted with the cameras that were currently available, there were "experts" in each camera present where possible, nobody from KineRAW or SinaCAM.

The tests were carried out in the University of the West of England and senior staff from there were present to make sure that everything was fair.

My methodology on DR is very simple, it's not as pedantic as some, I want to know what usable DR there is, not what the last point a pixel is visible through the noise!

To check what I would describe as shooting DR we overexpose a grey scale at half stop intervals until the difference between the brightest 2 chips disappears, the half stop before that is the limit. I don't use highly specialised kit to measure this, I use my eyes on a regular monitor from a regular viewing distance. The object after all is to see what the shooting DR is in normal conditions.

We then repeat this with exposures reducing, again, the point where I can't see the difference between the 2 darkest chips is a half stop beyond the DR limit.

If the camera is very noisy at this point I will comment, as I will also comment if it is very quiet. You may recall that I did mention this in regard to the Dragon.

Now can we all please be civilised about this.

Oh and I will be contacting Jared in July 2015 asking him to continue sponsorship of CML!!

Very happy of this attitude on both sides...

Even so I don't agree with all of Geoff's tests results, and or opinion about some things, I surely do respect his knowledge and time taken to share it in his great Blog and in Forums, and again while I but headed with him in some things, I enjoy reading him, and very happy that He is back on Reduser...

Sorry again Geoff we didn't get to work on that 3D project together I am sure it would have been a fun experience...


Welcome back, and if you are any were near Barcelona... shout!!

 

[luigivaltulini]

Alexa is strong in the highlights ..
Now with the new calibrations and OLPF close the gap and maybe we'll have something more.
After a couple of days spent with Alexa, and DRagon I struggle to see the difference in a real world of lights controlled., The thing you notice most of the monitor are the colors, the softness of the image to 6k it is very pleasant, and low noise. Dragon in this is better than Alexa.
For me, these tests are unitli chart to understand the extreme limits of the sensors.
I can not wait to try the new OLPF with a low cow, maybe Phil before we post something , I'm curious to see otherwise I have to wait another three weeks ...

 

[Niko Pueringer]

Anyone who wants to do their own comparisons can get all the RAW files from all the cameras at http://www.cinematography.net/UWE/index.html

Check the results for yourselves on your kit the way that you work.

I will add updated Dragon and F65 and a couple of other cameras in the future but first I need to do my "real" job!

Cool, thanks Geoff.

 

[James_Mills]

I've been recently considering canceling my Dragon upgrade and selling my RED package to purchase the Amira. This is primarily because I'm not confident I'll ever see my Dragon upgrade at this point considering I lost my place in line when they upgraded my Scarlet to an Epic instead (unbeknownst to myself, a 100% legit bummer).

I'd really like to see some Dragon VS Alexa tests for dynamic range, because the tests I've seen thus far clearly show that the Dragon has more range than the MX, but it doesn't look more than Alexa's 14+ stops to my eye (and certainly not more than 65mm film like the RED team claimed, at least not imo)

Our own RED fanboy bias' aside, what do you guys think of this idea? Has anyone else considered this?

 

[luigivaltulini]

I've been recently considering canceling my Dragon upgrade and selling my RED package to purchase the Amira. This is primarily because I'm not confident I'll ever see my Dragon upgrade at this point considering I lost my place in line when they upgraded my Scarlet to an Epic instead (unbeknownst to myself, a 100% legit bummer).

I'd really like to see some Dragon VS Alexa tests for dynamic range, because the tests I've seen thus far clearly show that the Dragon has more range than the MX, but it doesn't look more than Alexa's 14+ stops to my eye (and certainly not more than 65mm film like the RED team claimed, at least not imo)

Our own RED fanboy bias' aside, what do you guys think of this idea? Has anyone else considered this?

James, I think you should do your personal comparison.
Needless to watch these test ...
Get an Alexa and a Dragon and shoot something together.
I'm sure you'll be happy to see the differences.

 

[James_Mills]

James, I think you should do your personal comparison.
Needless to watch these test ...
Get an Alexa and a Dragon and shoot something together.
I'm sure you'll be happy to see the differences.

I definitely agree, just so hard to get ahold of either at this point so my mind is wandering and dreaming until then naturally.

These tests really concern me too:
http://www.cinematography.net/UWE/alexa.html
http://www.cinematography.net/UWE/dragon.html

If Dragon has a full 2+ stops more than Alexa, why is it very obvious in these tests that the Alexa is getting more information and detail in the black charts and color charts on both the first two images (over exposed) and the last two images (under exposed)? You can't even see any color detail on the chart in the first two images of the Dragon or all the black scales but you can clearly see them in the Alexa. Shouldn't this suggest the Dragon has less ​dynamic range yet it's claiming to have 2 stops more?

 

[Mike P.]

I kind of agree with James in terms of how impressive Alexa continues to be; Dragon's 16+ stops should be easily/clearly distinguishable from Alexa's 14+ stops, in the same way that Alexa's 14+ stops is so clearly distinguishable from MX's "13.5" stops. I'd even say the Amira form-factor is more applicable to more types of jobs/shooting than the DSMC package (size and weight considerations included).

That said, going to Dragon is $9500, which is way cheaper than getting an Amira... Plus, as much as I think 2.88k upscaled to 4k looks perfectly acceptable, spending top dollar to get that lower resolution when 4k is where things are so clearly headed, is still a hard pill to swallow. In all honesty, I think being able to sit on MX until the dust settles (Alexa4k at NAB? DragonGamma+OLPF offering more useable stops of DR on Dragon?) is actually a pretty good short-term game plan... In other words, I think an Amira is a bad purchase between now and April (unless you've got a gig lined up that will cover the cost).

 

[James_Mills]

That said, going to Dragon is $9500, which is way cheaper than getting an Amira... Plus, as much as I think 2.88k upscaled to 4k looks perfectly acceptable, spending top dollar to get that lower resolution when 4k is where things are so clearly headed, is still a hard pill to swallow. In all honesty, I think being able to sit on MX until the dust settles (Alexa4k at NAB? DragonGamma+OLPF offering more useable stops of DR on Dragon?) is actually a pretty good short-term game plan... In other words, I think an Amira is a bad purchase between now and April (unless you've got a gig lined up that will cover the cost).

The more research I do, the more I agree. You're spending a lot more money on an Amira for what I believe may be slightly better DR and skin tone / color but at the expense of only 2k capabilities in an ever evolving market (which it's pretty safe to say will see Dragon still looking great ten years from now). Plus it's not raw. Also, I have the same suspicion of an Alexa 4K trumping their other products and thus making them sooner to be obsolete.

I think the Dragon is the better option all in all, but I don't think it's right to claim 16+ stops of dynamic range when professionally performed tests suggest it to be less than Alexa's 14. The Dragon has so much going for it that I don't think it needs to reach in those regards.

 

[Niko Pueringer]

I'm also curious why this test shows less range than the Alexa when John Merchant profiled it at 16.5. If the Alexa underwent the same DR profiling that Marchant did, would it show a higher number?

And yes, I'm aware the new OLPF and color science add a stop. In the meantime, I'm just trying to learn the technical approach to these tests and why we see the results we do.

 

[Ergin Ozturk]

The more research I do, the more I agree. You're spending a lot more money on an Amira for what I believe may be slightly better DR and skin tone / color but at the expense of only 2k capabilities in an ever evolving market (which it's pretty safe to say will see Dragon still looking great ten years from now). Plus it's not raw. Also, I have the same suspicion of an Alexa 4K trumping their other products and thus making them sooner to be obsolete.

I think the Dragon is the better option all in all, but I don't think it's right to claim 16+ stops of dynamic range when professionally performed tests suggest it to be less than Alexa's 14. The Dragon has so much going for it that I don't think it needs to reach in those regards.

i have the same concerns. and changing my idea everyday

DR, highlights and grain are the one part of the game and relatively its easy to measue. and as far as i see from CML test Alexa still has +1 stop more in highlights. But I'm sure Dragon will be equal and even more with new OLFP filter. So i am almost confident on that.

BUT! my biggest question is the COLOR. Alexa has beautiful colours. It is not measurable i know but just look at movie "drive" how vivid, saturated, beautiful colour. I sent some of the Dragon footages to 2 professional colourists and ask their opinion i hope they will say "yes it is as good as alexa even better" than there is no question marks for me to upgrade dragon.

 

[Gunleik Groven]

It was the colors that had me go Dragon.
all the rest were just bonuses. Just sayin'

And with the upgrade it looks like it will get even better.

 

[Ergin Ozturk]

It was the colors that had me go Dragon.
all the rest were just bonuses. Just sayin'

And with the upgrade it looks like it will get even better.

Gunleik i trust you

but i am still waiting something like;

 

[D Anderson]

Ergin, That truly is amazing....














250g of HOT CHOCOLATE for only $1.69

 

[James_Mills]

It was the colors that had me go Dragon.
all the rest were just bonuses. Just sayin'

And with the upgrade it looks like it will get even better.

You don't mind the ever-so-slight green tint that has been brought up earlier in this thread? I've noticed it with my Epic MX at times and am discouraged to see it prevalent in these Dragon tests. I didn't notice them in Peter Collister's tests from earlier, however.

 

[luigivaltulini]

James I can understand your concerns, I also had these moments.
When I saw in my tests that Alexa had 1 stop more ', I was pissed and confused, I thought I had done something wrong in testing, I sent a few PM Gunleik to understand, but he already knew the answer to this difference in the high lights.
Then I tried to shoot something real with lights made by DP, looking at the monitor. As I mentioned the differences in favor of Dragon was little noise, a soft beautiful picture in 6k with nice colors and finally a nice rolloff, I wanted to do some testing with a low cow in front to see what was going on but I forgot ..
This made me forget everything and say how Dragon is a nice camera.

Phil showed us a preview of what will happen with the new OLPF and I look forward to.

.... a few weeks after the new OLPF and calibration:

DR: Alexa and DRagon same, maybe something more to DRagon
Noise: Dragon wins, much less noisy
Colors: Definitely better on DRagon, especially with tungsten light.
Resolution: here there should be no doubts
Overexposure: rolloff is very nice on both DRagon and Alexa.

For the rest should be a personal choice, or at least to those who think that it can serve you more at this time.
A camera, complete and as large as ARRI, or the modularity of RED.

Personally why I have chosen the Dragon as well as above:
- I can shoot alone with a light body, the modularity I love it. And the Carbon is fantastic.
- I can install on gimbal and octocopter.
- Be able to extract great pictures at high resolution
- Timelaps with double exposure (HDRx)
- I have one that comes to 300fps slow motion
- I have a prerecorder for years and this has allowed me to save a lot of space without losing the moment. Especially in nature. I think about Arri came only from SUP9 this year.
- Price
- RAW, a lot of flexibility in post.

just a thought ...

 

[James_Mills]

Thanks for the response, Luigi. While I'd like to believe that the new OLPF has equal to more dynamic range as the Alexa, I'm naturally suspicious seeing that RED had claimed the original sensor had 16+ stops before there was ever any need for a new one. We had also heard about the colors and skintone of Dragon beating everything else, and when tests somewhat seem to question or disprove this, we are now told that the new sensor is going to dramatically benefit all of these different areas. Such an abrupt change seems very sudden to me.

To the objective camera enthusiast looking in on this matter, I'd venture that this seems more like a cop-out than anything else. These tests done here seem fairly conclusive to me given the parameters in which they were held and the professionals that held them. Will we get the same tests done on the new OLPF? I'm not so sure, but I do believe that people will continue to claim that the new OLPF surpasses these previous test's results with no proof to back that up and I don't think that's fair to the Alexa.

 

[Phil Holland]

You don't mind the ever-so-slight green tint that has been brought up earlier in this thread? I've noticed it with my Epic MX at times and am discouraged to see it prevalent in these Dragon tests. I didn't notice them in Peter Collister's tests from earlier, however.

I am not worried and these tests weren't white balanced from what I've been told. However, the white balance is indeed slightly different with the new tech.

New OLPF/New Calibration compared to the old:

Daylight, No ND, No WB or Tint Adjustment.




This illustrates the New Calibration's change of base white balance at 5600K. Every digital cinema camera is a bit different in regards to white balance under various lighting conditions. No two are alike. On some systems, no two cameras are 100% dead on either.

Thanks for the response, Luigi. While I'd like to believe that the new OLPF has equal to more dynamic range as the Alexa, I'm naturally suspicious seeing that RED had claimed the original sensor had 16+ stops before there was ever any need for a new one. We had also heard about the colors and skintone of Dragon beating everything else, and when tests somewhat seem to question or disprove this, we are now told that the new sensor is going to dramatically benefit all of these different areas. Such an abrupt change seems very sudden to me.

To the objective camera enthusiast looking in on this matter, I'd venture that this seems more like a cop-out than anything else. These tests done here seem fairly conclusive to me given the parameters in which they were held and the professionals that held them. Will we get the same tests done on the new OLPF? I'm not so sure, but I do believe that people will continue to claim that the new OLPF surpasses these previous test's results with no proof to back that up and I don't think that's fair to the Alexa.

The Old OLPF and Calibration do measure up to a practical 16.3 stops of usable dynamic range. The key to Dragon's Dynamic Range and Color is how you choose to unlock it. The bigger reasons behind the New Calibration and OLPF are a few things:

- Better color "purity".
- Dynamic Range has been "remapped" which now places approximately 1.3 more stops in the highlights from a base exposure of 18% Gray in comparison to the Old Tech.
- New OLPF and Calibration every so slightly increases tonal separation and detail in shadow areas and also lessens the potential for certain types of shadow noise.
- White Balance is slightly different, favoring a more "warm gray" rather than "greenish gray" under 5600K light.

Eventually when I have free time and the new tech available I'll make proper measurements to illustrate the improved DR and Color further (though I do need to shoot things other than tests to make a living!). The best way is to read chip separation up and down off 18% Gray in 1/3 or 1/4 stop increments (that at least sets a relevant baseline/control), which I assure you is a decent amount of work to do properly. The moments where you can no longer measure tonal separation between chips is where the "wall" is. More complicated still is how you process the material to get the most out of it. I did a test against Alexa and F65 back in late December 2013 and it indeed totals more, however, that doesn't mean that it totals more in the exact same places as the other two cameras. This is actually what's a bit illustrated in the UWE tests. I'm extremely curious to see the New Calibration and OLPF against both of those cameras again, though since I've tested the new tech I feel it's less needed, but it does need to be done.


Just a couple quick shots from the Older OLPF/Calibration:

Putty in broad daylight.



Daylight Rainbow

 

[James_Mills]

I am not worried and these tests weren't white balanced from what I've been told. However, the white balance is indeed slightly different with the new tech.

Can someone confirm this? Given the credentials behind these test takers, I'd be shocked if something as rudimentary as white balancing wasn't controlled. If that was indeed the case then I'd begin to question just how much stock I'm giving these results as well.

The Old OLPF and Calibration do measure up to a practical 16.3 stops of usable dynamic range. The key to Dragon's Dynamic Range and Color is how you choose to unlock it.

So we're arguing semantics then? If we have allowed the distinction of latitude to become so ambiguous and immeasurable, shouldn't Alexa claim to have 18 stops of dynamic range?

Beautiful colors on the bottom two images btw, thanks for sharing.