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 :)

CML Camera Evaluation 2015

Check out what Dragoncolor 2 adds to the images. The test was conducted with Dragoncolor, but if you switch it over to Dragoncolor 2, things look even more impressive. I personally really like Dragoncolor2.

As for Dragon Skin tone, the test was conducted at 400ISO (and exposed for at 400ISO) so you're losing latitude on the high end in exchange for more on the low end. I'm assuming Geoff went that route because of the higher noise levels with the skintone olpf. That's why you see the +4 on the skintone olpf clip while the Low Light at +4 still holds on the high end. Low Light was exposed for 800ISO.
 
Christopher*Seguine said:
I dont see it listed anywhere but it looks like they used a cooke 50mm 5i from the metadata, kinda a crumby lens to use for a test.
Click to expand...

Yes, how crumby he may have used one of the best professional cinema lenses of our modern age.

care to explain?
 
I thought that as well, but when I reset the metadata to how it was shot it seemed to be 800, perhaps the r3d was adjusted just before the snapshot was taken.

I guess geoff could confirm, but I doubt he would re-light or expose just the dragon STH differently than every other cam in the test.

Maybe he's checking the thread and could chime in or perhaps that is in the notes of the testing procedure.

Battistella


Matt Ryan said:
Check out what Dragoncolor 2 adds to the images. The test was conducted with Dragoncolor, but if you switch it over to Dragoncolor 2, things look even more impressive. I personally really like Dragoncolor2.

As for Dragon Skin tone, the test was conducted at 400ISO (and exposed for at 400ISO) so you're losing latitude on the high end in exchange for more on the low end. I'm assuming Geoff went that route because of the higher noise levels with the skintone olpf. That's why you see the +4 on the skintone olpf clip while the Low Light at +4 still holds on the high end. Low Light was exposed for 800ISO.
Click to expand...
 
I'd say grade those images. There is much more info to be had in the STH. Here is the logic. If you drop the STH to say 400 ISO anduse that as a base ISO for grading, you'll end up with a cleaner overall image. With the latitude mapped that way you can easily pull it back for excellent results.

It might seem like counter intuitve logic, but if you needed to stop down the STH in this situation to match the LLO, that means it's actually more sensitive than the LLO which people would never think right?

Interesting.

Battistella
 
David Battistella said:
I'd say grade those images. There is much more info to be had in the STH. Here is the logic. If you drop the STH to say 400 ISO anduse that as a base ISO for grading, you'll end up with a cleaner overall image. With the latitude mapped that way you can easily pull it back for excellent results.

Its counter intuitve logic, but if you needed to stop down the STH in this situation, that means it's actually more sensitive than the LLO which people would never think right?

Interesting.

Battistella
Click to expand...

The problem is the skintone olpf is clipped and the LLO olpf still appears cleaner and remains unclipped. So in this test it appears the LLO olpf has more DR.
 
Right. I agree. On first look, out of the box that might be true and how it looks, but which one do you think is more elastic. In terms of pushing saturation I feel the STH is a bit better. We both know the deb cleans up the STH quite well, but it's all a preference thing. I think for products, landscape or anything where color accuracy and prescision is critical (like, at times for me reproducing the colors of a renaissance painting) it becomes important.

So if it seems like I harp on color, then I just wanted to let folks know why. :)

battistella
 
David Battistella said:
Naw, I don't think it is a disservice, choice is never a disservice. When the whole discussion was going on RED was saying, "we believe in the STH and this is why it will ship and these are the reasons."

now, we see those reasons in this great test by Geoff. It's actually like looking at two different cameras. I have both and I am going to let the STH live in the dragon for a while after doing a bunch of tests. Really, with DEB in RCX and in resolve through the SDK, I'm very happy.

i also think choice is a good thing and there are people who will prefer the LLO for its "clean out of the box" look. I just got to used to the color reproduction on the STH.

battistella
Click to expand...

David, now with 3rd Parties creating an OLPF market and the watergates fully open, we will see many different flavours of Dragon, being the STH OLPF one of many. So I can imagine future discussions about colour fidelity, noise, etc... in many, many stages of confusion including not only the filter but also colour science Dragoncolour 1 and 2, and different colour setups for different OPLF technologies, in the moment 2.

Have we been talking about ISO and noise, heavily depending on the shooting environment an the individual approach of exposure, and often not really satisfying for the original poster, will be now so many new variables added to the mix that I foresee a cam-tech talk babel that can't be disentangled anymore, a gordian knot of cinematographical questions. Haha.

Ok. Seriously, Red is THE nerd cam in the market. Fine for me - in the moment I still see through the Red-universe but I know many skilled DPs that do not do that, don't want that and hence prefer cameras from a different vendor, mostly the chaps form Munich. Choices are good. But if choices make things complicated and not much better, at least in the eyes of a majority of users things are taking the wrong road. Or to put it differently, does Red's new OLPF politics increase sales or pull people from the Alexa bandwagon back? In the moment I doubt that. But I could be wrong.

Imagine Geoff's next UWE test in 2017: There will be 4 Dragon OLPFs to be tested, if Arri joins Red's OLPF trip, he will have to test at least 2 Alexa versions, and Sony won't be shy either bringing 3 different F-65 to the shootout. Now that Sony is in the OLPF races BM don't want be left behind, the same counts for the AJA cam. And yes, the AJA MK2 Version will have 4 strengths of locon OLPFs available to help the not so brilliant latitude. Haha. Digital cameras in the pro-level are much, much more complicated that film ever was. Does it help the quality?



Hans
 
Probably the best thing one can do for the quality of their images is run the through a pipeline and be completely comfortable with the exposure path for the scenarios in the film. There are plenty of owner operators or DPs that want a one size fits all solution and that is most likely the LLO or the STH.

As for specialized OLPFs the underwater guys will probably be estatic.

Then one day, a big name DP will drop a quote in an article about how they used a home cooked OLPF for the look of the film and everyone with need it because they need something to replace LUTS or magic bullet looks because that is just the way this wheels seems to turn. :)

we need geeks, we need artists. We need geeks to be artists and we need artists to be geeks. That's what my look at history has taught me. :) most architects do not execute their vision without an engineer. Some are both and can fully explore both sides of the brain. But let's never limit the brain. :)

battistella
 
I think the OLPF thing is good for one thing, putting back focus on optical signal-path even on digital cameras.
Digital cameras have to a certain degree been discussed as the if the optical path was not relevant.

Whether this is good or not for RED, well that is a separate discussion.

Like Hans, I know a lot of excellent DoP's who have no interest in being camera-geeks.
They learn a tool (from their own non-geeky premises) and will happily and confidently use that tool the way they have learned it and invest their names in those results.

So while I see Davids argument, there is a lot of merit to Hans.
But what approach ends up as the good one, we'll know in 10 years time. Or... By then there will be at least a couple more iterations/generations of cinema cameras, and maybe some of the DR/color challenges approached by the OLPF solutions have been settled.

Still exciting to follow development. That's for sure.
(from a geeky perspective)

So many great cameras now that can make good pictures.
(from a photography point of view)
 
Okay...after further inspection, what I thought was the case...isn't. It all comes down to the labeling for the ISO 400 tungsten shots versus the ISO 800 ones.

Geoff states on this page that the ISO 400 cameras labeled "Stop 0" are actually -1 stop (1 stop underexposed) because they kept the light levels the same and lit for ISO 800 on the tungsten day of shooting.

Initially, I thought they opened up the lens +1 stop and called that "Stop 0" on the ISO 400 cams but that's not correct. They left the lens at T5.6 for both ISO 800 and ISO 400 cams and called that "Stop 0" for both, even though the ISO 400 cams were 1 stop underexposed at that light level.

So, for the tungsten shots, you can't simply go by the labels/slates when comparing Dragon Skin (set at ISO 400) to Dragon Low Light (set at ISO 800) or comparing any ISO 400 camera to any other ISO 800 camera.

In terms of light levels, "Stop +4" on Dragon Skin is comparable to "Stop +3" on Dragon Low Light. "Stop -4" on Dragon Low Light is comparable to "Stop -3" on Dragon Skin. Dragon Skin "Stop -4" is actually 5 stops underexposed.

Do I have it right this time? If so, people are going to have to be careful when trying to do 1-to-1 comparisons based on how each shot is marked.

Given people's tendency to read the directions last, I think the assumption that all "Stop Whatevers" are created equal is going to cause a bit of confusion. Perhaps the discrepancy should be mentioned more prominently.
 
If the lenses on both camera were at t5.6 then setting both to 800 or 400 or 1000 ISO makes no difference.

By by your logic, set the STH to 800 and all things are equal, because they were both set to T5.6

the rest is just meta data errors.
 
Brian Boyer said:
Initially, I thought they opened up the lens +1 stop and called that "Stop 0" on the ISO 400 cams but that's not correct. They left the lens at T5.6 for both ISO 800 and ISO 400 cams and called that "Stop 0" for both, even though the ISO 400 cams were 1 stop underexposed at that light level.
Click to expand...

heres what does not make sense. Both were lit exactly the same way? Both were opened to T5.6 one was set to 400 and the other to 800. If this is the case it's a metadata error and the 400 is NOT one stop under exposed.

Battistella
 
Ohh, it completely escaped me that STH was lit and exposed for 400 iso and LLO was lit/exposed to 800. So there's a stop more light hitting the sensor in the STH Day, taking it to clip town.
 
This still is not clear to me.

Can we all agree that if the lighting was consistent and the T stop was 5.6, it does not matter that the camera was set to 400, 800 or 1000? we understand this, correct?

Now, if he lit the STH for 400 and set it to t 5.6, that is another story but that is not how I am reading what Brian wrote.

battistella
 
As it happened the lighting did change, Geoff says on the site of the test: "On the daylight day we changed levels, doubling the light level for the 400 ISO cameras, so that exposures were constant."

So I think STH Day got lumped in with the 400 iso cameras somehow and received the extra light whereas LLO Day was done at 800 and got a stop less!
 
David Battistella said:
This still is not clear to me.

Can we all agree that if the lighting was consistent and the T stop was 5.6, it does not matter that the camera was set to 400, 800 or 1000? we understand this, correct?

Now, if he lit the STH for 400 and set it to t 5.6, that is another story but that is not how I am reading what Brian wrote.

battistella
Click to expand...

This stuff gets me thinking in circles sometimes but...

For any given light level, if you meter it for ISO 800, the meter will tell you what t-stop to set in order to properly expose middle gray. If you then change the ISO on the meter, it will give you a different t-stop setting. Okay so far?

If you take a spot reading of a middle gray chip at ISO 800 and adjust the lighting instruments so the meter reads T5.6, changing the ISO to 400 on the meter will also yield a different t-stop number. Still in agreement?

So, what I think is happening with this test is they set up the tungsten shoot day by metering to the gray chip on the chart and adjusting the light levels until the meter read T5.6 at ISO 800.

That's the amount of light you would need to get that chip the correct brightness (in this case, T5.6) at ISO 800, which is less light than you would need to get that chip to read T5.6 at ISO 400. Therefore, the ISO 400 cams are a stop underexposed because they only have ISO 800's worth of light hitting the sensor.

I'm basing this on what Geoff wrote. Maybe I'm misunderstanding him and trying to bend reality to fit what I think he meant.


EDIT: I double checked this by comparing the RGB parades of the tungsten Dragon Low Light (ISO 800) to Dragon Skin (ISO 400) shots at 1 stop differences and they line up. Dragon Low Light "Stop +1" is roughly equivalent to Dragon Skin "Stop +2," and so on.
 
Last edited:
Justin Carrig said:
As it happened the lighting did change, Geoff says on the site of the test: "On the daylight day we changed levels, doubling the light level for the 400 ISO cameras, so that exposures were constant."

So I think STH Day got lumped in with the 400 iso cameras somehow and received the extra light whereas LLO Day was done at 800 and got a stop less!
Click to expand...

Which is not at all an ilogical postulate of many of the things that has been written about the OLPFs
 
But does not make it apples to apples.
That can obviously create a boatload of confusion when comparing dragon STH and Dragon LLO as, evidently the lighting was inconsistent.

This means plus or minus post gain on what the sensor captured and 1 stop is a difference that funnels down.

EDIT: looking at the +3 STH and the +4 LLO makes sense, but it still confused things a bit as it is not an apples to apples comparison.

Battistella
 
Last edited:
You must log in or register to reply here.
Back
Top