DRAGON - ITS THE REAL DEAL !!!! - (PART 2)

[paul schefz]

Sorry but skin lacks saturation to my eye, and a bit too pink - should be a tad more toward orange. Again, subtle, but...

You can nail skin tones. The problem is nailing skin tones while keeping other colors as they should be without using a secondary qualifications and corrections. This frame is a bad example because there are no known colors in it other than skin - the lipstick's true color is unknown.

Just to be clear - the problem is very subtle, and occurs only under certain types of lighting, with certain types of complexion, but it's there. On an F65, no. On Dragon, I'm hoping, also no.

As to those who are calling me a detractor - as a thinking person I feel it is not my job to just say "Rah Rah Red!!!" no matter what, like a mindless sycophant. Thing is, clients, directors, studios etc., all have these concerns. If you just get angry at me, or them, instead of addressing these concerns they will simply request another camera, and that's no good for Red or for Red owners who promote their use.

sorry, but what you are saying makes no sense at all....we could put up 10 or 20 variations of that shot going from more yellow to more red with all kinds of saturation and everything in between, let people take a poll and most likely not a single version would get more then 25% of the vote....
there is no such thing as "perfect skin tone" does not exist....i am happy for you that you are calibrating your screens and calibrators to keep everything perfectly balanced on your end...but once you let ANYONE see ANYTHNG it all goes out the window...you do know that? right? 5 screens will show you 5 different "perfect skin tones"....and not only that...it is also completely subjective....add to that everybody sees color differently....

 

[Elsie N]

....add to that everybody sees color differently....

There may be a simple answer here to explain why Rob sees things differently. At first I thought he may be aesthete, but I think Peter cornered that with his image. But then I thought, what if Rob is a synesthete? Not all out but having just a mild case. That would explain how he sees color differently. Not that that is a bad thing necessarily, just unnecessary when it comes to making pretty images.

 

[Brant Hadfield]

I asked Roger Deakins, in the midst of a similar discussion, which shots in Skyfall were used with Epic on the octocopter by the second unit during the chase scenes in Istanbul, vs the other footage shot on Alexa. His reply:
"There were probably just two shots of Bond on the rooftop in Istanbul which were made with the Red and used in 'Skyfall'. They are quick shots and you would never notice the difference. I have done side by side tests on a portrait, however, and that is where you really see the subtle differences in colour representation, differences that cannot be counteracted in a DI suite."
I think there were actually more than just two shots from the octo and Epic, but point taken. More dense though smaller photo sites benefits resolution. Larger photo sites benefit variation in color sensitivity at the cost of resolution. I hope Dragon turns this model on its ear.
I do believe we've not yet really seen a good representation of Dragon's capability, given its current stage of development, and importantly, the color science.

 

[Vadim Bobkovsky]

In daylight through the window with slight diffusion, similar Caucasian (I assume) face would look a lot like this, without much of orange. I'm not sure if my eyes are calibrated or I'm partially color blind, but that's how I see it in real life most of the time when I'm sober. Mixed lighting and tungsten is another question, of course. Again, less noise from the sensor, cleaner RGB channels under different complex lighting conditions and increased sensitivity will make cinematographer's life that much easier, no doubts about it. MX is cool, but Dragon is my preciousss.

 

[Andy Roberts]

Rob have u seen that recent test on CML with f65, Alexa, Epic, BMCC, c500 and F55? Worth looking at:
http://www.omegabroadcast.com/fmp/omega-detail.php?-manuRecID=630&-recid=2472&-tab=0

Jake, your link was for a shoulder mount pad...

 

[luigivaltulini]

Jake, your link was for a shoulder mount pad...

http://www.cinematography.net/hannover-2013.html

I have never seen a test like this ...
or rather, I've never seen anyone so bad use the RED.
but how do you make a comparison with the middle of the video signal!!

too bad....

 

[Michael J Brennan]

Larger photo sites benefit variation in color sensitivity at the cost of resolution

Variation in colour sensitivity?
How so?



Mike Brennan

 

[Peter Majtan]

Sorry but skin lacks saturation to my eye, and a bit too pink - should be a tad more toward orange. Again, subtle, but...

You can nail skin tones. The problem is nailing skin tones while keeping other colors as they should be without using a secondary qualifications and corrections. This frame is a bad example because there are no known colors in it other than skin - the lipstick's true color is unknown.

Just to be clear - the problem is very subtle, and occurs only under certain types of lighting, with certain types of complexion, but it's there. On an F65, no. On Dragon, I'm hoping, also no.

As to those who are calling me a detractor - as a thinking person I feel it is not my job to just say "Rah Rah Red!!!" no matter what, like a mindless sycophant. Thing is, clients, directors, studios etc., all have these concerns. If you just get angry at me, or them, instead of addressing these concerns they will simply request another camera, and that's no good for Red or for Red owners who promote their use.

So first "the problem" is too much yellow in the skin-tones and now it is too little? Rob, with all due respect - you have no idea what that model's skin looked like - just like you have no idea what Mark daughter's skin looks like. I just did a BTS of a major fashion shoot here in Tokyo for GILT. Three European and two Japanese models - all with different skin-tone and hair color. Not a single shot exhibits "your problem"...

Here is the fashion photographer himself against a reasonably neutral grey wall (and the color of his Hassie is also well known...) - again straight out of the camera with no grading...:

And here is a Japanese model (the lipstick shot was an European model) - behind her to the left is a neutral grey wall...:

Here are all 5 in the same shot (plenty of known colors in there)...:

I am sorry to say so - but you might have spent way too much time in your grading suite. For the record - I am colorist myself and ironically - some of my early color work was for Mark back in 2000/2001. There clearly is no problem. Not with the cameras that is...

:sifone: Peter

 

[Dave Blackham]

So first "the problem" is too much yellow in the skin-tones and now it is too little? Rob, with all due respect - you have no idea what that model's skin looked like - just like you have no idea what Mark daughter's skin looks like. I just did a BTS of a major fashion shoot here in Tokyo for GILT. Three European and two Japanese models - all with different skin-tone and hair color. Not a single shot exhibits "your problem"...

Here is the fashion photographer himself against a reasonably neutral grey wall (and the color of his Hassie is also well known...) - again straight out of the camera with no grading...:

And here is a Japanese model (the lipstick shot was an European model) - behind her to the left is a neutral grey wall...:

Here are all 5 in the same shot (plenty of known colors in there)...:

I am sorry to say so - but you might have spent way too much time in your grading suite. For the record - I am colorist myself and ironically - some of my early color work was for Mark back in 2000/2001. There clearly is no problem. Not with the cameras that is...

:sifone: Peter

Peter, Great post. Every time I had thought theres some skin tone issue and Ive looked hard at has been shot or lit I can track the issue down to something else other than the camera. More recently we have been plagued with poor lighting CRI, which can be very subtle but is an issue. We have a lot of tools available to us and I for one am still learning.

 

[Blair S. Paulsen]

Talk about the deleterious effects of poor CRI lighting sources, try shooting the BMCC with them.

My point would be, that the more accurate the capturing device, then the more one notices the color shifts causes by the lighting source. Color temperature and tint are excellent ways to compensate for the impact of lighting type, but since they are global adjustments they cannot precisely track vagaries throughout the spectrum.

Very excited by the move to remote phosphor lighting technology that appears to offer both high efficiency and high CRI (color rendering index).

Cheers - #19

 

[Robert Ruffo New]

There may be a simple answer here to explain why Rob sees things differently. At first I thought he may be aesthete, but I think Peter cornered that with his image. But then I thought, what if Rob is a synesthete? Not all out but having just a mild case. That would explain how he sees color differently. Not that that is a bad thing necessarily, just unnecessary when it comes to making pretty images.

See Roger Deakins comments above. He's no idiot and neither am I. I do not have color blindness. My color acuity is in the top fraction of a percentile - i.e. 99.5% of people have less. I am not imagining things. There is such a thing as a sense-memory skin tone that looks just right to us for all races - it is not really all that subjective. Out of the box, (meaning white-balanced on a grey card next to the actor, but no other grading done) Red MX with RG3 is very close but just a bit off, to a varying degree depending on subject skin color (tan, race) and light temperature . Other cameras nail this better. I love Red, we shoot most of our projects with Red, but color precession has never been my favorite thing about the system. It's not off enough for me to prefer another system, it's very good as it is, but it does have some room for improvement. The issues are not very obvious all the time, but they do happen, more than with some other cameras. If you did a survey of top DPs most would probably all say that Red MX color (RG3) is very good but not perfect yet.

Now wether you have trained your eye to be highly discerning about things like this or not, I can't tell you. But I do know that those of us who have trained our eyes and minds to see color very carefully notice these problems - maybe Mark disagrees, and he inarguably does great work, but I am not alone in this position and it is based on many hours spent grading Red footage, and working with the many Red cameras we own, not on hearsay of any kind.

Again, it's a very subtle problem - not a terrible problem.

 

[Robert Ruffo New]

So first "the problem" is too much yellow in the skin-tones and now it is too little? Rob, with all due respect - you have no idea what that model's skin looked like - just like you have no idea what Mark daughter's skin looks like. I just did a BTS of a major fashion shoot here in Tokyo for GILT. Three European and two Japanese models - all with different skin-tone and hair color. Not a single shot exhibits "your problem"...

Here is the fashion photographer himself against a reasonably neutral grey wall (and the color of his Hassie is also well known...) - again straight out of the camera with no grading...:

And here is a Japanese model (the lipstick shot was an European model) - behind her to the left is a neutral grey wall...:

Here are all 5 in the same shot (plenty of known colors in there)...:

I am sorry to say so - but you might have spent way too much time in your grading suite. For the record - I am colorist myself and ironically - some of my early color work was for Mark back in 2000/2001. There clearly is no problem. Not with the cameras that is...

:sifone: Peter

The color of your skin tones is very close but its slightly off - the saturation still just a touch under, the shading looks ever so lightly off, which is a spectral response thing (as far as I know, I;m not 100% sure what causes that, but I've seen this very subtle color distortion before) Sorry, but ask any colorist about Epic MX skin tones. They are slightly off "out of the box". Side by side with an Alexa, out of the box, the Alexa does always look slightly more correct - under some lighting situations much more correct. It's so subtle most of the time (and not terribly hard to fix, just a bit of a pain) that it's not worth the weight and complication of shooting ARRI raw, not to mention the resolution hit, and I always nonetheless recommend Epic instead, but if Dragon gets us that last 5% (maybe 2%) of the way, then there will be no more trade-offs. This is what I am hoping for.

 

[Blair S. Paulsen]

There are doubtless a number of factors that impact skin tone rendering with any camera system/processing pipeline. Moreover, the way in which mixed or low CRI lighting affects color rendering varies rather widely. With the MX chip, I find daylight Kino's noticeably outperform any other lighting source in terms of natural looking skin values and overall color precision.

Once we have Dragon color science that's ready for prime time, I'll be curious to see how well the color values track with challenging sources. I'll wager the Kinos will still look great ;-)

Cheers - #19

 

[paul schefz]

My color acuity is in the top fraction of a percentile - i.e. 99.5% of people have less.

i wonder how THAT is measured....we get together, i show you pantone swatches and you tell me the correct number?

i thought color was light reflecting off a surface, seen by an eye, interpreted by a brain....even that seems to have a lot of variables in it...and that is just 2 people standing next to each other, seeing the same scene....i wonder who told you that you see that scene 99.5% more correct then others....

 

[paul schefz]

Remember this Red color example.

View attachment 41481

PS

How do I post and make the image appear larger?

is that an official red color example? or from a thread?

 

[Eric Haase]

See Roger Deakins comments above. He's no idiot and neither am I. I do not have color blindness. My color acuity is in the top fraction of a percentile - i.e. 99.5% of people have less. I am not imagining things. There is such a thing as a sense-memory skin tone that looks just right to us for all races - it is not really all that subjective. Out of the box, (meaning white-balanced on a grey card next to the actor, but no other grading done) Red MX with RG3 is very close but just a bit off, to a varying degree depending on subject skin color (tan, race) and light temperature . Other cameras nail this better. I love Red, we shoot most of our projects with Red, but color precession has never been my favorite thing about the system. It's not off enough for me to prefer another system, it's very good as it is, but it does have some room for improvement. The issues are not very obvious all the time, but they do happen, more than with some other cameras. If you did a survey of top DPs most would probably all say that Red MX color (RG3) is very good but not perfect yet.

Now wether you have trained your eye to be highly discerning about things like this or not, I can't tell you. But I do know that those of us who have trained our eyes and minds to see color very carefully notice these problems - maybe Mark disagrees, and he inarguably does great work, but I am not alone in this position and it is based on many hours spent grading Red footage, and working with the many Red cameras we own, not on hearsay of any kind.

Again, it's a very subtle problem - not a terrible problem.

I agree with you on this. I'm not sure why everyone is so defensive about this subject. You are clearly trying to point out a small shortcoming of the MX sensor with the hopes that RED will note it and improve it. I've been hoping for this kind of improvement since they announced a new sensor. Yes you can get there with the skin tones with some secondaries in DI but the RED sensor and color science COULD be improved for more accurate color rendition OVERALL and specifically for how the sensor sees and color science develops the skin tone. What Rob Ruffo is describing is what most DP's mean when they say they don't like the "skin tones" on the Epic. A lot of redusers including the guys at RED have stated there is no issue at all. But I think a lot of DPs see, feel, and experience a subtle but important difference in the way RED MX sees color compared to other sensors. As he stated, it's not a dealbreaker, just something that could be improved.

 

[Elsie N]

See Roger Deakins comments above. He's no idiot and neither am I. I do not have color blindness. My color acuity is in the top fraction of a percentile - i.e. 99.5% of people have less. I am not imagining things. There is such a thing as a sense-memory skin tone that looks just right to us for all races - it is not really all that subjective. Out of the box, (meaning white-balanced on a grey card next to the actor, but no other grading done) Red MX with RG3 is very close but just a bit off, to a varying degree depending on subject skin color (tan, race) and light temperature . Other cameras nail this better. I love Red, we shoot most of our projects with Red, but color precession has never been my favorite thing about the system. It's not off enough for me to prefer another system, it's very good as it is, but it does have some room for improvement. The issues are not very obvious all the time, but they do happen, more than with some other cameras. If you did a survey of top DPs most would probably all say that Red MX color (RG3) is very good but not perfect yet.

Now wether you have trained your eye to be highly discerning about things like this or not, I can't tell you. But I do know that those of us who have trained our eyes and minds to see color very carefully notice these problems - maybe Mark disagrees, and he inarguably does great work, but I am not alone in this position and it is based on many hours spent grading Red footage, and working with the many Red cameras we own, not on hearsay of any kind.

Again, it's a very subtle problem - not a terrible problem.

See? By your own admission you do have greater color acuity than the majority... although I didn't know you could measure that.

And I'm not challenging your abilities either as a colorist or natural color recognition. All I'm sayin' is, that other 99.5% will never see what you have identified... only the .05 percenters like yourself.

I just think you are overthinking this.

 

[Elsie N]

i wonder how THAT is measured....we get together, i show you pantone swatches and you tell me the correct number?

i thought color was light reflecting off a surface, seen by an eye, interpreted by a brain....even that seems to have a lot of variables in it...and that is just 2 people standing next to each other, seeing the same scene....i wonder who told you that you see that scene 99.5% more correct then others....

I'm only guessing here, but I think a lot of old hippies from the 60s and 70s who took too many LSD trips are used as controls for measuring psychedelic color recognition. '-)

(Sorry, just too good an opportunity to suggest those old burnt out bags of bones came to some good. Back on topic now. '-)

 

[Scott Crawley]

See? By your own admission you do have greater color acuity than the majority... although I didn't know you could measure that.

Well... there is THIS ... FWIW.

 

[paul schefz]

I agree with you on this. I'm not sure why everyone is so defensive about this subject. You are clearly trying to point out a small shortcoming of the MX sensor with the hopes that RED will note it and improve it. I've been hoping for this kind of improvement since they announced a new sensor. Yes you can get there with the skin tones with some secondaries in DI but the RED sensor and color science COULD be improved for more accurate color rendition OVERALL and specifically for how the sensor sees and color science develops the skin tone. What Rob Ruffo is describing is what most DP's mean when they say they don't like the "skin tones" on the Epic. A lot of redusers including the guys at RED have stated there is no issue at all. But I think a lot of DPs see, feel, and experience a subtle but important difference in the way RED MX sees color compared to other sensors. As he stated, it's not a dealbreaker, just something that could be improved.

i am not sure why this thread has to be turned into a discussion about how "accurate" dragon color might be...especially when it is judged from ungraded compressed footage from a beta camera with beta color science....
i completely agree that every sensor /camera captures color differently...of course it does....and there is nothing wrong with prefering one interpretation over another....i just can't agree with one being better/more accurate/whatever then another....and especially not in this (test/beta) case....and it really does not make a lot of sense to me when talking about skin tone....
i shoot fashion and getting the color right with fabrics often is extremely critical....and i am not talking about pleasing or nice or looking good....i am tlaking about match the pantone swatch right....and the epic has some of the best color representation out there....even compared to high end phase backs....
i can understand that some people don't like what they are getting with epic....some people don't like how canon DSLRs have a yellowish skin tone (which of course also can be dialed down)....some people prefer the more "neutral" nikon look....most people prefer the canons because of the more "pleasant" look straight out of camera...which really means they don't know anything because they are looking at a jpeg....but often that helps with a client....

regardless.....if anyone does not like the way the camera captures and interprets certain tones or one just can't get the right tones out of it, there is nothing wrong with working with a different camera....

i just don't think this thread is the right place for this discussion...the first thread got shut down, i don't see the point in working on doing the same to this one...

i would prefer mark and other people who can actually answer questions about dragon did not have to wade through all this and loose interest....