Defending RED, Camera (Brand) Wars, Shot On What?!

[Aharon Rothschild]

@luigi, I need to dig through CML, but does anybody know what ND filters they used in that side by side?

I have alarmingly different results, but I think the variance is the ND itself.

From the site:
"The cameras, where appropriate, were set to whatever resolution came closest to a "standard" S35 or Full Frame image.

Lighting was Blonde's and ARRI 2Kw Fresnel's for the Tungsten tests and ARRI Sky Panels for the comparison with LED's

The reference charts were courtesy of DSClabs and also X-Rite & Kodak ones purchased by CML.

The Cameras were exposed in half stop intervals from T1.5 to T16 with a 180 degree shutter. We added ND 1.2 and then exposed from T4 to T16 to give us a complete range of exposures. ND's used were mainly the built in ones but where these were not available we used Tiffen, no colour correction was made for built in ND's any change is inherent in the camera.

All cameras were set to the manufacturers recommended EI.

Light level was 400 foot candles."

I'd be very interested in how your xyla charts compare.... Any chance you could post a. R3D snapshot?

 

[Phil Holland]

Here's everything I found with Monstro, which was a year ago this month, but wow it feels like longer than that.

http://www.reduser.net/forum/showthread.php?165811-RED-Monstro-Dynamic-Range-Examined

Xyla is incredibly useful and is the resultant patches here, but it's important to test DR in even more controlled setups, which is why I confirm my results two other ways, sadly those are also the most time consuming. Particularly the patch test. But glass aperture and ND are not controls for a DR test. Too much variation particularly in the first 3-4 stops on a typical lens and wild variation of channel info on NDs, which from the sounds of this was "just some OG Tiffen ND". Much better modern options than that.

My last big test/research project actually had to deal with the limits of Dynamic Range in human vision, which I'll be publishing shortly. Some fascinating data and surrounding theories came out of all that.

 

[Tom Gleeson]

But glass aperture and ND are not controls for a DR test. Too much variation particularly in the first 3-4 stops on a typical lens and wild variation of channel info on NDs, .

Good points in regards to both camera tests and comparisons. Ideally the same set of NDs and the same lens for each camera. In my experience the rating of an ND filter is more an approximation than exact science. For future tests would the use of shutter speed be a better choice than ND filters?

 

[Phil Holland]

better choice

I find myself in an odd place as I test cameras professionally and for a day rate for various entities a few times a year and should keep a whole bunch of this to myself, but I'll sweep away my common sense for exactly one moment in order to be helpful.

If you are testing or comparing camera sensors, glass of any sort should not be the variable at all. And shutter speed should be a constant unless testing variance related to it, i.e. don't change it.

There are separate tests that really should be done for color/DR stability and accuracy of shutter speeds, angles, and integration times. For instance, one camera (not the two in question here) had an issue on this front early on which was clearly detected under regular shooting conditions.

If you are testing lenses or ND Filters, go right ahead. On this front as you'll discover what ND works best for each sensor t check and optical stack behind the lens.

But in this case and even on a lab setting under a reference optic the light is what should be manipulated for a comparison such as this to eliminate any question of what is being looked at. Which takes work as well and often not as simple as dimming or gelling a somewhat wild light source. This is common in the world of testing cameras for cinema application, even more so today.

Keeping the format size consistant-ish here was correct to a degree in aperture adjustment however. But too much going on uniquely to be a good way to measure in this case.

 

[PatrickFaith]

Rather then a technical problem I see this more as a brand thing. For the Arri brand the story is real simple "sensor hasn't changed in 10 years, and if you want a extreme high end look rent the super large sensor size". For the RED brand the story is complex, with ipp2 and the current dsmc2 line up I think it makes sense, but the story needs to be told over a few beers.

I btw think the brand split on hydrogen brand was a brilliant move, to me their was a problem on the branding with the raven which has been cleaned up. The hydrogen brand split makes the story much better: hydrogen is for all the new cell formats ... all the way from high end consumer to some some professional use cases.

 

[Hrvoje Simic]

Good points in regards to both camera tests and comparisons. Ideally the same set of NDs and the same lens for each camera.

For a general DR test with lit scene, and only applied to few bottom stops limited by iris usability. For Xyla, no NDs.
For colour response test, no NDs.

 

[Hans von Sonntag]

I think it's surprising that all of the Oscar-nominated films, all of the Sundance dramatic award winners, and most of the Sundance documentaries avoided using Red cameras. There's nothing inherently wrong or bad about Red cameras that would preclude using them for documentaries.

Technically I would say, yes, it's surprising to see Red so underrepresented because they can render outstanding pictures. But as usual, it's not all about specs and form factors or Xylia charts. It's more than often about reputation. Now, imagine you are a conscious filmmaker, imagine your are technology agnostic, no nerd and mainstream do the trick for you. Perhaps heavy metal isn't your music. Would you choose the camera with the SS-skulls all over the place if there were alternatives with an immaculate reputation, say Arri, that doesn't distract you visually?

Red does a lot right when it comes to technology. But their marketing skills...? Very segregating. Red lives in a male bubble, love Hummers and black SUVs, tattoos, a military language like assault plate or tactical whatnot, call their flagship weapon and sticks on everything crosswalk stripes and skulls. That is ridiculous and immature and surely is a significant deterrence to anyone sensitive.

Nonetheless, I'm upgrading my Epic Dragon to a Gemini. But that's me because I know what a Red camera can do in the right hands. But I also understand what inevitably is happening when a Red camera is in inexperienced hands. There is a reason why there are so many threads about noise on these boards. You won't find such posts in Sony or Arri land.

There are many reasons not to use a Red camera.

Hans

 

[Mike Nagel]

Tests are only there to indicate behavior. The more "lab grade" ur test is, the more unrealistic it can be when it comes to practical application. The Xylia charts indicate DR levels, which at the extremes, often mean very little when it comes to real life usage. Hence why very few care (unless there is a massive discrepancy in DR between two cams)...

The CML tests - while "quick and dirty" - also indicate certain behavioral patterns... And many will argue that these are much more important than any "lab grade" tests...

Arri holds its color throughout over/under the best... Which on any given set, over/under will happen and in the edit and grade this will line up the best... Besides that they got the best color... as in: most pleasing OOTB

The Arri and RED already capture more DR and color than "needed", meaning what makes the bulk of the image the viewer actually consciously perceives... Yet the top echelon prefers Arri... go figure.

All of this discussed ad nauseum since 2007. RED is a camera maker, they are not color science perfectionists like Arri is. Good thing is with digital and raw u can tweak, with great results. Something that is above the pay grade of 99% of RED users - and clearly RED themselves ;-)

The "simple" CML tests should be a clear message to novices, as they will run into this...

 

[Phil Holland]

Tests are only there to indicate behavior. The more lab grade ur test is, the more unrealistic it can be when it comes to practical application. The Xylia charts indicate DR levels, which at the extremes, often mean very little when it comes to real life usage.

That to me seems really odd. Most DPs I know who in particular use meters like to understand how many stops above and below 18% gray you have before you lose information at various ISO ratings. I mean, one of several reasons I do the Xyla and patch tests.

I guess you can tell a gaffer to wing it

 

[Mike Nagel]

That to me seems really odd. Most DPs I know who in particular use meters like to understand how many stops above and below 18% gray you have before you lose information at various ISO ratings. I mean, one of several reasons I do the Xyla and patch tests.

I guess you can tell a gaffer to wing it

yeah Phil they do that... and then on real sets for 100+ years they shoot 1-2 stops brighter than the scene will actually end up to be (after post)... there's a million ways to skin a cat, and u gotta know how to shoot right for the capture medium at hand...

also, those DPs then usually do NOT push to the very extremes, which is exactly what my sentence alluded to... because most of the high end cams only differ in the top/bottom extremes on these charts... u can use all of them, if u know how...

but maybe u can show ur Xylia chart tests to Deakins (ur old buddy) and convince him to shoot RED... ;-))))

just to be clear: I've shot Xylia chart tests as well, but I know how to place the returned data into proper context when it comes to real life application...

DR is not the "problem" any longer for a while now... face the music

 

[Andy Jarosz]

yeah Phil they do that... and then on real sets for 100+ years they shoot 1-2 stops brighter than the scene will actually end up to be (after post)... there's a million ways to skin a cat, and u gotta know how to shoot right for the capture medium at hand...

Can you let me know how to get to this magical land where DPs shoot 2 stops over? Because for the past year I've been on project after project that are horribly underexposed, and they don't realize until post that they can't grade properly. Cameras have more DR than ever...and DPs are only using 1/2 of it.

 

[Phil Holland]

but maybe u can show ur Xylia chart tests to Deakins (ur old buddy) and convince him to shoot RED... ;-))))

I'll be seeing him tomorrow, but we rarely talk about that sort of thing. Slightly more interesting discussions occur.

 

[Mike Nagel]

I'll be seeing him tomorrow, but we rarely talk about that sort of thing. Slightly more interesting discussions occur.

out of interest, since u as a RED owner and operator (like many of us) have started this thread and listed a few things that in your opinion makes RED a valid choice (many of us agree), what does your friend Roger state as his reasons why he prefers Arri ?

 

[Phil Holland]

out of interest, since u as a RED owner and operator (like many of us) have started this thread and listed a few things that in your opinion makes RED a valid choice (many of us agree), what does your friend Roger state as his reasons why he prefers Arri ?

Not something I'd discuss in a public forum. But Roger can do whatever he wants, he's one of my favorite DPs.

 

[Misha Engel]

what does your friend Roger state as his reasons why he prefers Arri ?

To my knowledge he doesn't, he'll shoot an entire movie with a smartphone when he has to and it will still look good.

 

[Karim D. Ghantous]

I think it's ultra lame to tell other DPs what equipment they should use. Unless, of course, you are asked for your opinion. I don't care if a DP is an ASC member or someone who just discovered cinematography. If anything, users of RED cameras can let other DPs know that if they want help or advice with RED cameras, they can get in touch. Be helpful, lead by example. Don't preach, don't nag.

This is a RED camera website. You can certainly give unsolicited opinions to fellow members - just like I'm doing right now. As far as I'm concerned, be as blunt as you need to be. But outside these forums, don't try and 'convert' people, whether it's to 35mm film, or Leica lenses, or RED cameras.

Edit: I love this quote by Deakins, as he is 100% right:

I genuinely feel that cinematography, like photography in general, is not something that can be learned but, pretentious as it may sound, can only be discovered.

 

[Mike Nagel]

To my knowledge he doesn't, he'll shoot an entire movie with a smartphone when he has to and it will still look good.

people do a lot of things when "they have to"... same as Lubezky, he doesn't "have to", he picks...

 

[David J. Buchanan]

I agree that it's pretty lame nowadays to tell any DP what camera they "should" be using.

This OP made this thread because people have been bashing RED for ridiculous reasons, and fair enough... but there's no real necessity to even talk about RED vs ARRI anymore unless you're genuinely curious.

It's like film vs digital, no one brings that up anymore.

 

[Hrvoje Simic]

Technically I would say, yes, it's surprising to see Red so underrepresented because they can render outstanding pictures. But as usual, it's not all about specs and form factors or Xylia charts. It's more than often about reputation. Now, imagine you are a conscious filmmaker, imagine your are technology agnostic, no nerd and mainstream do the trick for you. Perhaps heavy metal isn't your music. Would you choose the camera with the SS-skulls all over the place if there were alternatives with an immaculate reputation, say Arri, that doesn't distract you visually?

It's not a frikking SS skull. It's a visual analogy.

Red had "camera brains" before. Which was innacurate analogy.
It's a camera head. Head = skull outside, brain inside.

If it associates to that, it's a matter of particular design approach and context (especially with militaristic naming), not the symbol itself.

 

[Hrvoje Simic]

Can you let me know how to get to this magical land where DPs shoot 2 stops over?

That was the approach with film due to highlight lattitude and with introduction of digital in many cases changed to paranoia over highlights, due to poor DR and highlight lattitude of first era of digital cinema and quality of transformations from log. Overexposure with digital now is left as a practice more for low light scenarios, to avoid the thing under b) in the next part of the text.

I've had a bunch of talks with directors and DoPs before shoots and on the sets, insisting low light scene blacks are exactly exposed as they see it on the set gamma corrected, relying on that gamma transformation as a reference and underestimating post production process requirements, but many directors and DoPs don't require education in that department and bring perfect negative for post. Some DoPs mostly look at the scene and use the monitor to observe the "health" of log signal, with only occasional glance at gamma corrected approximation. When the scene requires large DR variables, DoP's will naturally incline to a camera which does the upper part best. Not just in range but in lattitude and tonal stability.

Because for the past year I've been on project after project that are horribly underexposed, and they don't realize until post that they can't grade properly.

Besides ignorance and paranoia, underexposure tendencies are further motivated by:

a) generalized and misleading ETTR trend
b) the trend of relying on digging data out of shadows, inherited from raw digital photography, when pixels shrunk, mid grey moved to the left and highlight lattitude decreased, the trend which is for the most part flawed in principle for cinematography, depending on specific light design, contrast ratios and tonal/texture continuity, complex post production and display/presentation factors
c) 65472 types of log gammas between manufacturers and 6453423748 monitoring pipeline variables
d) log transformations overboosting the image, for the camera users unaware they are basing exposure decisions on already pushed image