Never mind specs for a moment, what I hope Dragon looks like

[Curt Pair]

Yes.. I'm a Red man... that uses the ARRI from time to time...(and would jump ship in a heart beat if anyone bought out a better camera).

I'm still unsure what this "look" or "Secret sauce" you talk of.. ? I can very very easily make my RED images look exactly like Alexa images... Im in RAW, I can grade people blue if i want, but one thing I cannot do is make my Alexa images look like RED images... RED is pin sharp Alexa is soft (compared to RED) .. RED is shooting massive 75 meg per fram images, not 6.6 meg per fram.... You really notice how soft the Alexa is when viewing both on a 4k monitor or even better on a 4k projector... it's seriously like chalk and cheese. But once again, it's so easy to match colour and tone... and if you want the soft look you hit the blur tool.

But the fact still is, I can match skin colour for skin color with the ALEXA any day of the week and do so with ease.... I'm not sure if you guys judge your images from an onset monitor in stead of a proper grading suite, if your grading from DPX? or some silly lut created by some numpty on set... but for the life of me, I have know idea what your talking about. I can grade both exactly the way I want.

It's a pity that this subject keeps raising its head, as I truly think its a myth... forged by the inexperience.

Regarding Dragon... I'm just hoping for a cleaner chip at hi ISO and a bit more range. I'm quite happy all ready.

Mark stole everything I wanted to say... Ditto!

 

[Dave Blackham]

In regards to Alexa here - many posts seem to suggest that it is gaining ground because people are misinformed or it's easier in post etc. As an owner of both systems I can give you my perspective.

The Epic wins in the following areas:

Weight (by a long shot)

Size (again by a long shot)

Resolution (Also by a long shot)

Menu access (as in the Alexa I personally find frustrating with it's presets system which take time and are limiting for colour balance etc). Also it's dumb side menu system is difficult for the operator to access so is slower whilst getting the first AC to change something.

Slow motion - (both in that the Alexa takes forty seconds to get to (and from) high speed - and in terms of speeds on offer)

Anamorphic desqueeze: On the Red this is simple - on the Alexa it is (a) a costly upgrade and (b) requires all manner of fiddling with framelines whilst switching to get it correct and also looks wrong in the eyepiece to me - too wide.

PL mount: All my PL lenses (5 sets) fit on the Epic - not all of them fit into the Alexa - the Red RPP's for example are too large to sit into and close the mount. As this is a micron based standard I don't know why this is.

Power consumption: The Alexa is a power hungry beast compared to the Epic - I've never seen 26v batteries get eaten so fast.

Where Alexa wins:

Image: The picture is more filmic.

Colour: Colour reproduction seems more full and pleasing to me. Closer to film. Skin tones especially.

DR - Seems to offer way more highlight roll off than the Epic. Flares etc seem to roll off more pleasingly.

In short - In my opinion, the Epic beats the Alexa in all areas except in one key area: Look. However for me as a DP who hands footage to a producer and has little control over it's post house journey after that, I prefer handing over rushes that don't need any post tinkering (often out of my direct control) to make work. In short, as a DP, in my opinion, the Alexa offers a chance for me to do what I do best - light and shoot then hand in finished work which I am happy with. For others, ie: shooters who 'do it all' and can afford to do their own post work this may not apply.

Having looked at Alexa recently One area that may affect an organic look is Alexa processes data in Log internally. I can see how this can affect the Prores SxS on board recording but would have thought that his can be processed similarly with in RC-X for a Red workflow. I don't know the math used for RC-X but would be interested to know. I also dont know how or if the log processing in Alexa affect a Arri RAW recording and post processing but Im sure if forms part of the visual signature from the camera if you using a RAW workflow. I did see Skyfall yesterday which was as youd expect Stellar. Some posts mention the Ariels were shot with Epic which cut perfectly fine. I thought Skintones were exceptional. My only comment from watching the movie on a Digital projector in a cinema is I found my self wanting more resolution as a viewer.

 

[Brian Boyer]

I see where Brian's going with this. And I agree. It would be great to get a more "rounded" image (mainly a nicer falloff on the highlights). I feel the Sony F65 has already made the case for more resolution, sharpness, clarity. The camera looks great but it still "looks" to me like a more-perfect technical advancement and not necessarily a more "filmic" advancement. I hope Dragon is not just the next technological leap. There is definitely something to be said about the "character" of an image. And as Bruce pointed out everything has a look. Many of us here have worked with these images for years. I, for one, can usually tell the difference between cameras like the Epic and Alexa. I don't buy the argument that you can make one look like the other. Most viewers will never notice a difference but we are not "most viewers". I think even Mr. Toia would agree with me here...

Thanks Adrian. I don't think one can have a discussion about improving sensor designs and not take Alexa and Epic MX into account. However, Mark and a few others have become fixated on the Alexa portion of what I said.

My preference, or anyone else's, for Alexa's image quality is immaterial to the statements I made about Dragon. So, in order to keep things moving, let's say Mark is right and I'm imagining things about Alexa's image or I need new glasses. Who cares?

The question is, should Dragon look like a sizable leap beyond what either Alexa or Epic MX currently offer in terms of image quality? If so, in what way?

Or, should Dragon merely look like an enhanced version of what Epic already offers? This question is dependent upon whether you think Epic has a look or not. Mark and company say no. Bruce, Adrian and myself say yes.

BTW, I essentially said what Bruce said in the post right before it but it's probably not worth mentioning. In fact, forget I said anything about being first. I'm okay with letting such a profound statement go unnoticed.

Should Dragon's image have more of an organic signature look of it's own or should it be more of a blank canvas?

For the sake of argument...

Let's say RED deliberately designed or inadvertently stumbled upon a means to make Dragon capable of capturing 95% of the look that Kodak Vision3 5219 color negative film produces, as a baseline look, without the grain but with the added benefits of much improved low light sensitivity over the 5219 and, of course, RAW workflow. Let's also say that you could bump the look up to 98% if you added the appropriate grain in post. For those who are preparing to become fixated on my choice of film stocks, please use your noggin/imagination and substitute any high performing film stock of your choice.

Would this be the best thing to happen to digital sensors or a disastrous turn for the worse?

You'd still have the flexibility to change the look through grading like you can now. I know some will argue they can already make Epic look like any film stock they want through grading. But, is there a benefit to having that very popular and very pleasing organic look on hand from the get-go? The benefit I'm seeing is you'd get a more organic look no matter what you did in grading because of the architecture of the sensor.

Obviously, that represents a trade-off, but do you feel it's worth it?

I'm not suggesting RED should spend their time trying to emulate film stocks. I am asking if RED should spend their time producing a more organic looking sensor.

 

[D Fuller]

For the sake of argument...

Let's say RED deliberately designed or inadvertently stumbled upon a means to make Dragon capable of capturing 95% of the look that Kodak Vision3 5219 color negative film produces, as a baseline look, without the grain but with the added benefits of much improved low light sensitivity over the 5219 and, of course, RAW workflow. Let's also say that you could bump the look up to 98% if you added the appropriate grain in post. For those who are preparing to become fixated on my choice of film stocks, please use your noggin/imagination and substitute any high performing film stock of your choice.

Would this be the best thing to happen to digital sensors or a disastrous turn for the worse?

You'd still have the flexibility to change the look through grading like you can now. I know some will argue they can already make Epic look like any film stock they want through grading. But, is there a benefit to having that very popular and very pleasing organic look on hand from the get-go? The benefit I'm seeing is you'd get a more organic look no matter what you did in grading because of the architecture of the sensor.

Obviously, that represents a trade-off, but do you feel it's worth it?

I'm not suggesting RED should spend their time trying to emulate film stocks. I am asking if RED should spend their time producing a more organic looking sensor.

This whole post kinda mystifies me. You do realize that any look you associate with 5219 or any other film stock is the result of grading?

 

[Brian Boyer]

This whole post kinda mystifies me. You do realize that any look you associate with 5219 or any other film stock is the result of grading?

I could be wrong but I don't think that's true in the context that I'm asking the question.

Film doesn't need grading to look like film. Is there nothing more to film looking like it does than just color curves and grain? I know they're a big part of what we consider film to look like, but is that it? I'm not so sure.

If grading is so magical, I shouldn't be able to tell the difference between 5219 stock and Epic or Alexa footage graded to look like 5219 with some grain added in post. I bet you can tell the difference if you look closely enough. I know Christopher Nolan and plenty others think they can tell the difference.

Some shots may not give themselves away as readily as others but eventually some give up the ghost and their digital nature is revealed.

What I'm talking about is an ungraded image from Dragon being indistinguishable, in virtually every way, from an ungraded film image. Would that be a good thing or a bad thing?

Grading goes a long way but I've yet to see it do what I've been talking about as far as more organic looking images.

 

[Brian Boyer]

Even the algorithms in RED and Alexa that write to a RAW format impart a 'look'. So far only Sony have published their formula for getting the data back to the linear light of the scene. Most camera manufactures depart from the objectivity of their sensors in a special recipe designed to maximise the DR, resolution, shutter speed and writing speed all while trying to refridgerate it down to acceptable levels. Oils and Oils and RAWs and RAWs.

But the truly salient point made by Toia IMO is the art of RAW manipulation. Past a certain threshold of picture information the _potential_ to alter the appearance of the image transcends any a priori traits of the camera. Just watching Spider Man now and marveling at how it looks like itself and at the same time it looks like RED and at the same time looks like film.

John Lennon said 'give me a tuba, I'll get you something out of it' about being an artist. Well, lets say the 5D is a tuba and the RED is a Gibson Les Paul.

I don't disagree with anyone who extols the virtues of the flexibility of RAW manipulation.

Due to the fact that virtually no one answered the question, that is in itself an answer.

Am I to take from this most people's position is basically "who cares what the image initially looks like when I can make it look like whatever I want in post?"?

Now I'm starting to understand why people couldn't help themselves when it came to throwing around frame rates and resolutions and prices and every other number when speculating about Dragon.

Except, when every camera maker and his brother does RAW, has 18+ stops of DR, 4K+ resolution and high frame rates in a small, lightweight package and numbers no longer are enough to differentiate between them all, the ones with the "best looking image" are going to come out on top.

When that shift happens, I guarantee you every mofo on this forum won't shut up about it.

And I'll tell you another thing, if I ever read Mark Toia talking about how organic a RED camera's image looks, I'm going to let him have it.

 

[Adrian Jebef]

Truth is that the Epic and the Alexa are just different cameras. The Log encoding that the Alexa uses to record to ProRes (and to display ARRIRAW) doesn't change any of the MATERIAL TEXTURE of the image. That texture is simply there, created when you turned the thing on. Every digital camera has a "look". Just as every film stock has a "look". Grading will change things but it WILL NOT change the fundamental baseline. Unless the grade is so harsh that any subtleties are crushed (see "Rampart"). It seems like there are plenty of RED users who can only "see" specs and not necessarily the "look". I've always found the R1 and EPIC to have a very subtle "video look". There's not a lot of character in the immediate image besides clean, sharp lines and the usual clipped highlights. When the camera is handled correctly the images can be stunning but out of the box it's quite Vanilla to me personally. This is not necessarily a bad thing. As others have commented, most love this attribute as it allows them to create their vision in post as they see fit. I prefer my cameras to have a more immediate, recognizable character. It's just my preference. And I'd love to see more "character" in the Dragon sensor. But a bit more highlight detail / natural roll off I'd settle for.

BTW: digital cameras that I think have their own unique personality (ie: "special sauce"): DVX100, Viper Filmstream, Alexa, Ikonoskop ACam DII.

And just to be clear, I feel we are at a threshold technically. I don't buy the argument for increased resolution. 4K is more than enough (displayed), anything above that is just showing off. I look forward to more DR at the micro level. 18 stops doesn't do me any good when they are all at the top and bottom of the curve. Subtleties in micro contrast above and below middle-grey is what is really important. The gradual transition from stop to stop or level to level wins me every time (it's really the one area that the Alexa continues to excel in, even vs. the F65). And you can NEVER have enough "natural looking" skin tones. Again the Alexa is pretty on target here, the F65 does well, and the EPIC has certainly gotten better in these past few months. So that's my Dragon wish-list: better DR subtlety (especially highlights) and better color-reproduction.

 

[Brian Boyer]

Truth is that the Epic and the Alexa are just different cameras. The Log encoding that the Alexa uses to record to ProRes (and to display ARRIRAW) doesn't change any of the MATERIAL TEXTURE of the image. That texture is simply there, created when you turned the thing on. Every digital camera has a "look". Just as every film stock has a "look". Grading will change things but it WILL NOT change the fundamental baseline. Unless the grade is so harsh that any subtleties are crushed (see "Rampart"). It seems like there are plenty of RED users who can only "see" specs and not necessarily the "look". I've always found the R1 and EPIC to have a very subtle "video look". There's not a lot of character in the immediate image besides clean, sharp lines and the usual clipped highlights. When the camera is handled correctly the images can be stunning but out of the box it's quite Vanilla to me personally. This is not necessarily a bad thing. As others have commented, most love this attribute as it allows them to create their vision in post as they see fit. I prefer my cameras to have a more immediate, recognizable character. It's just my preference. And I'd love to see more "character" in the Dragon sensor. But a bit more highlight detail / natural roll off I'd settle for.

BTW: digital cameras that I think have their own unique personality (ie: "special sauce"): DVX100, Viper Filmstream, Alexa, Ikonoskop ACam DII.

And just to be clear, I feel we are at a threshold technically. I don't buy the argument for increased resolution. 4K is more than enough (displayed), anything above that is just showing off. I look forward to more DR at the micro level. 18 stops doesn't do me any good when they are all at the top and bottom of the curve. Subtleties in micro contrast above and below middle-grey is what is really important. The gradual transition from stop to stop or level to level wins me every time (it's really the one area that the Alexa continues to excel in, even vs. the F65). And you can NEVER have enough "natural looking" skin tones. Again the Alexa is pretty on target here, the F65 does well, and the EPIC has certainly gotten better in these past few months. So that's my Dragon wish-list: better DR subtlety (especially highlights) and better color-reproduction.

Precisely.

There are two types of highlights:

1) The ones that are the brightest part of an object and approach or exceed the upper clipping range of the sensor. These generally blow out to white (much quicker than film does, no matter what the original color of the source - another area that needs work) and are the ones most people associate with a video look and refer to when talking about highlight protection.

2) The ones that are the brightest part of an object but come nowhere near clipping the sensor. These are the ones I think are subject to what Adrian described as the "subtleties in micro contrast above and below middle-grey".

Part of the video look I've been talking about is apparent around the second type of highlight.

I know virtually nothing about sensor design. Maybe an increasingly wider dynamic range is all that's needed to achieve a "gradual transition from stop to stop or level to level" due to the fact you're adding more luminance values with each stop. It seems it should also help with reproducing better looking skin tones, along with a better color filter and wider color gamut. But that gradual transition Adrian refers to, however it's accomplished, is key to help round out and refine the image in its entirety and not just at the extremes.

If wider DR is the answer, Dragon should be a noticeable improvement in this regard over MX with approximately 2 stops more.

P.S. I used to own a DVX-100. There is something about its look that's appealing, despite its MiniDV limitations. It's the camera that made me think Panasonic was going to be the company to beat...until they suddenly weren't. It's like they didn't know where to go from there once HD came around and got lost in making me-too video tools. The AF100 was a relatively admirable attempt at a low cost cinema-like camera for a brief time, no doubt spurred on by the original 3K for $3K Scarlet announcement.

 

[Mark Kern]

I would love a test where someone tries to dial a Red into the Alexa look. Anybody with access to an Alexa up for trying this?

This isn't the first time I've heard the claim that low-con OLPF is the only difference, and that you can easily throw one on your glass and get indistinguishable results with a RED, but I haven't seen any tests. Not only that, but there are others that claim the lowcons just add noise to the shadows (by "brightening them) which essentially nullifies their "increased dynamic range".

A little off topic, but does anyone have any links to RED footage with and without 1/8 or 1/4 LowCon filters?

 

[Björn Benckert]

Thanks Adrian. I don't think one can have a discussion about improving sensor designs and not take Alexa and Epic MX into account. However, Mark and a few others have become fixated on the Alexa portion of what I said.

My preference, or anyone else's, for Alexa's image quality is immaterial to the statements I made about Dragon. So, in order to keep things moving, let's say Mark is right and I'm imagining things about Alexa's image or I need new glasses. Who cares?

The question is, should Dragon look like a sizable leap beyond what either Alexa or Epic MX currently offer in terms of image quality? If so, in what way?

Or, should Dragon merely look like an enhanced version of what Epic already offers? This question is dependent upon whether you think Epic has a look or not. Mark and company say no. Bruce, Adrian and myself say yes.

BTW, I essentially said what Bruce said in the post right before it but it's probably not worth mentioning. In fact, forget I said anything about being first. I'm okay with letting such a profound statement go unnoticed.

Should Dragon's image have more of an organic signature look of it's own or should it be more of a blank canvas?

For the sake of argument...

Let's say RED deliberately designed or inadvertently stumbled upon a means to make Dragon capable of capturing 95% of the look that Kodak Vision3 5219 color negative film produces, as a baseline look, without the grain but with the added benefits of much improved low light sensitivity over the 5219 and, of course, RAW workflow. Let's also say that you could bump the look up to 98% if you added the appropriate grain in post. For those who are preparing to become fixated on my choice of film stocks, please use your noggin/imagination and substitute any high performing film stock of your choice.

Would this be the best thing to happen to digital sensors or a disastrous turn for the worse?

You'd still have the flexibility to change the look through grading like you can now. I know some will argue they can already make Epic look like any film stock they want through grading. But, is there a benefit to having that very popular and very pleasing organic look on hand from the get-go? The benefit I'm seeing is you'd get a more organic look no matter what you did in grading because of the architecture of the sensor.

Obviously, that represents a trade-off, but do you feel it's worth it?

I'm not suggesting RED should spend their time trying to emulate film stocks. I am asking if RED should spend their time producing a more organic looking sensor.

I think if you got a Red log film file and make sure your histogram is about centered and toss a Vision 3 5219 3D lut on it it will look like Vision 3 5219 period. IF you then also custom match the detail, noise and grain. I can assure you there would not be a problem having every second shot coming from film and the next from the epic. So this fuss about looks I find is just something that comes from the fact that it's close to impossible to match a certain films stock when just using the tools available in red cine or such grading application. If you would use Cineon or any real film digiital colortiming tool on the other hand that would end the whole discussion.

Try the log2RBG's in this zip file and tell me if they do not perfectly mimic both kodak and fuji.

http://www.syndicate.se/Default.aspx?Id=116

user and pass is: lut

 

[Dave Blackham]

Thanks for this Bjorn. So in your view is it possible to build what one could call an Alexa LUT and apply it to an Epic raw file. I realise all else would ned to be equal same glass etc. Sensor resolution wont match of course or OLPF but would be interesting to see how close you could get.

 

[Brian Boyer]

I think if you got a Red log film file and make sure your histogram is about centered and toss a Vision 3 5219 3D lut on it it will look like Vision 3 5219 period. IF you then also custom match the detail, noise and grain. I can assure you there would not be a problem having every second shot coming from film and the next from the epic. So this fuss about looks I find is just something that comes from the fact that it's close to impossible to match a certain films stock when just using the tools available in red cine or such grading application. If you would use Cineon or any real film digiital colortiming tool on the other hand that would end the whole discussion.

Except...I'm not having the discussion you think I am.

I agree. You could cut between Epic and film without anyone noticing in many instances. J.J. Abrams cut between film and R1MX on Super 8 and no one was the wiser.

But this thread has almost nothing to do with film. It has almost nothing to do with Alexa. I mentioned them as a frame of reference in order to try to illustrate my thoughts about Dragon. I failed miserably.

My main point was this: Under some, not all, conditions, Epic (and Alexa) look more videoish than others. I can see it. I'm certain many others can as well, including some of the good folks at RED. The MX is one of the best digital sensors to date, but there's room for improvement. I hope Dragon goes considerably further in diminishing the video look.

 

[Akos Nyoszoli]

Comment no.1:
I saw von Trier - Melancholia (shot on Alexa) projected from 35mm, and I couldn't tell it was not shot on film, until some night exteriors lit by some messy super-yellow lights.
I saw a feature (Tomboy) shot on 5DmkII from 35mm print, and it looked decent on the canvas.
I saw movies shot on 35mm, projected by a digital projector in another cinema, and they all looked digital.
(I know 90% of the people watch films on monitors nowadays)

Comment no.2:
Tree of Life couldnt have been shot on any digital camera. Maybe in a few years, maybe on Dragon.

 

[Carlos Reyes]

Comment no.1:
I saw von Trier - Melancholia (shot on Alexa) projected from 35mm, and I couldn't tell it was not shot on film, until some night exteriors lit by some messy super-yellow lights.
I saw a feature (Tomboy) shot on 5DmkII from 35mm print, and it looked decent on the canvas.
I saw movies shot on 35mm, projected by a digital projector in another cinema, and they all looked digital.
(I know 90% of the people watch films on monitors nowadays)

Comment no.2:
Tree of Life couldnt have been shot on any digital camera. Maybe in a few years, maybe on Dragon.

I don't know if I agree with your second point, and that's one of my favorite films of all time. I think chivo could pull it off with either alexa or epic.

 

[Brett Harrison]

Except...I'm not having the discussion you think I am.

I agree. You could cut between Epic and film without anyone noticing in many instances. J.J. Abrams cut between film and R1MX on Super 8 and no one was the wiser.

But this thread has almost nothing to do with film. It has almost nothing to do with Alexa. I mentioned them as a frame of reference in order to try to illustrate my thoughts about Dragon. I failed miserably.

My main point was this: Under some, not all, conditions, Epic (and Alexa) look more videoish than others. I can see it. I'm certain many others can as well, including some of the good folks at RED. The MX is one of the best digital sensors to date, but there's room for improvement. I hope Dragon goes considerably further in diminishing the video look.

Under some conditions film (especially the more recent stocks) looks like video. Isn't there a lesson there?

Also, isn't it kind of arbitrary to hold emulsion up as the gold standard of looks (something that IMO is steeped in the history of emulsion and bent glass in fact). I know kids growing up with movies like The Hobbit and Spider Man won't question the way it looks. It's time to develop a new aesthetic and video is the basis of this. Where we go from here is up to the creativity of engineers, shooters and colorists to innovate in the digital realm.

 

[Brian Boyer]

Under some conditions film (especially the more recent stocks) looks like video. Isn't there a lesson there?

Also, isn't it kind of arbitrary to hold emulsion up as the gold standard of looks (something that IMO is steeped in the history of emulsion and bent glass in fact). I know kids growing up with movies like The Hobbit and Spider Man won't question the way it looks. It's time to develop a new aesthetic and video is the basis of this. Where we go from here is up to the creativity of engineers, shooters and colorists to innovate in the digital realm.

I've heard that complaint about newer film stocks. And I agree, sometimes film can take on characteristics people more often associate with video, although I won't go so far as to say I'd mistake it for video.

As for holding emulsion up as the gold standard, once again, I feel like I'm defending against a point I never made.

It's ironic. I try to avoid being misunderstood by being detailed. Being detailed means I tend to type more. Typing more means people have to read more. People end up skimming over or ignoring whole sections of what I type. I end up misunderstood.

Earlier on, I stated I purposely refrained from mentioning film. This was specifically to avoid the comment you made. I never said digital cinema should hold film as the gold standard, only that it shouldn't look like video. Do you like that look for movies? Are the Alexa and Epic so popular because they look more like video or less?

As I also stated before, the video look isn't just a digital thing. Analog video has the same look. It's not a new aesthetic. I'd be perfectly happy if the new aesthetic meant digital sensors got to the point where they didn't always look like film but they never looked like video.

I just want digital cameras to have the quality that makes film so appealing. Not necessarily the exact look of film.

Audio went through a similar thing. At first, just getting it into a digital form was the trick and it was bloody expensive to do it with high fidelity.

Now that it's commonplace and computing power is such that you can record CD quality audio on your phone, engineers and programmers are going back to try to emulate (analog modeling, for example) or recapture the character of analog recording that got stripped away when zeros and ones took over in the early days.

They do it because, in addition to being nostalgic for some and a cost saving measure for others, there's something about the sound that's very appealing and just plain worth saving.

As for digital cinema, a new generation of kids probably wouldn't question the look if they were never given a chance to prefer something different.

 

[Brett Harrison]

I've heard that complaint about newer film stocks. And I agree, sometimes film can take on characteristics people more often associate with video, although I won't go so far as to say I'd mistake it for video.

As for holding emulsion up as the gold standard, once again, I feel like I'm defending against a point I never made.

It's ironic. I try to avoid being misunderstood by being detailed. Being detailed means I tend to type more. Typing more means people have to read more. People end up skimming over or ignoring whole sections of what I type. I end up misunderstood.

Earlier on, I stated I purposely refrained from mentioning film. This was specifically to avoid the comment you made. I never said digital cinema should hold film as the gold standard, only that it shouldn't look like video. Do you like that look for movies? Are the Alexa and Epic so popular because they look more like video or less?

As I also stated before, the video look isn't just a digital thing. Analog video has the same look. It's not a new aesthetic. I'd be perfectly happy if the new aesthetic meant digital sensors got to the point where they didn't always look like film but they never looked like video.

I just want digital cameras to have the quality that makes film so appealing. Not necessarily the exact look of film.

Audio went through a similar thing. At first, just getting it into a digital form was the trick and it was bloody expensive to do it with high fidelity.

Now that it's commonplace and computing power is such that you can record CD quality audio on your phone, engineers and programmers are going back to try to emulate (analog modeling, for example) or recapture the character of analog recording that got stripped away when zeros and ones took over in the early days.

They do it because, in addition to being nostalgic for some and a cost saving measure for others, there's something about the sound that's very appealing and just plain worth saving.

As for digital cinema, a new generation of kids probably wouldn't question the look if they were never given a chance to prefer something different.

I think it's all associational. An aesthetic is the consequence of arbitrary methods (mediums give rise to methods) that are refined later. No one elected for film emulsion particularly; it presented itself as a sound reproduction of reality, and at the beginning of movies you have it looking quite different with very little dynamic range. All the same, people cowered in their seats when a train seemed to emerge from the screen in the Nickelodeon theaters. I.e. that arbitrary aesthetic was a blank slate (no association) and therefore defaulted to 'reality'.

Then people shaped what film was. The film curve we love was a development. I believe that through the late 70s to early 80s was a big change (if anyone has a good text on the history of film emulsion I'd appreciate a post reply or PM). Then video became strongly associated with 'newsreel reality' (the style of 'on the March' news reels employed in Citizen Kane for example) and that is another example of changing aesthetic as association changes.

Perhaps try and state your contention in once sentence, but if you want to separate quality from look, well the quality of film is the way it looks and nothing else. Unless you mean 'quality' as a measure of value. I think a lot about this and I would say that what you're getting at is complexity, i.e. plasticity in the hands of an artist to mould it into something expressive. Film is quite complex in its effects. I think we are honestly above that threshold already. But here's my point: the history of video aesthetic will eventually develop into a new cinema aesthetic, influenced by other video mediums like video games and even ENG video, as seen in the found footage movies.

It's true that a culture needs to be eased into a new aesthetic like a person learning a foreign language, but at some point it becomes conversant - if you learn it before you're too old you won't retain an accent. Digital audio is a prime example as you rightly point out - but for my point as well: look at all the in the 'in the box' music producers who don't even use instruments or even solid state synths - it's all digital. They tend to spurn so called analogue modeling systems. 90 percent of the synths on Inception are a software synthesiser.

 

[Bruce Allen]

I think if you got a Red log film file and make sure your histogram is about centered and toss a Vision 3 5219 3D lut on it it will look like Vision 3 5219 period. IF you then also custom match the detail, noise and grain. I can assure you there would not be a problem having every second shot coming from film and the next from the epic. So this fuss about looks I find is just something that comes from the fact that it's close to impossible to match a certain films stock when just using the tools available in red cine or such grading application. If you would use Cineon or any real film digiital colortiming tool on the other hand that would end the whole discussion.

Try the log2RBG's in this zip file and tell me if they do not perfectly mimic both kodak and fuji.

http://www.syndicate.se/Default.aspx?Id=116

user and pass is: lut

Hey Bjorn

Thanks for throwing down the gauntlet and going to the trouble of posting LUTs

I'll try them on some footage tomorrow.

Personally... as I have said before, I think that when it comes to skin tones (and not just overall color of the skin but subtle color variations across skin), you can't make RED look 100% like film without practically painting every frame in your grading app.

But I would love to be proved wrong!

Anyway, I've never been able to make them exactly match - and I haven't seen a film shot on RED yet that actually has color that I like.

Alexa... I kinda like. But I can tell it a mile away. My day job is working on big movie trailers - we got in a new film and within 15 seconds of watching footage I could tell it was shot on Alexa. It just has that "trying to look like film and getting 90% there but the skintones are too uniform straw-tinted" look

Anyway, to separate fact from fiction, I'm going to do my own attempt at a test next year.

I just don't have the budget to do a test for the hell of it but when I prep to shoot my next short (unfortunately these don't happen that often because I have to raise the money myself by just saving it up!)... I will prep with a test shoot with multiple ethnicities, under multiple lighting conditions, with and without makeup (this seems to strongly affect things - I think due to RED either blocking or receiving the wrong stuff in the near-IR spectrum, which makeup affects). With some wide gamut color targets as well. Plus a bunch of color chips which I will measure with a spectraphotometer.

I'll use Fuji and Kodak stocks (need to buy the Fuji stuff before they close up shop!). And Dragon, 4K Alexa, F55, C500, Panasonic's 4K stuff if it's ready, and anything else (Blackmagic? Apertus? My friend MacGregor's F35?). And Dragon with a blue filter. And Dragon with a slight diffusion to get an Alexa-style look. Plus stills with Canon and Nikon DSLRs. Also will test in front of mirror glass for 3D rig and see if that affects things (there are some theories that polarization is actually part of what screws up skintones - because the subsurface scattering component of skin has different polarization characteristics to the specular reflections of skin - and I want to see this for myself).

If anyone wants to come in with me on this, they're very welcome!

The only catch is that in addition to shooting boring charts, I also want to do a simple test: attempt to recreate the look of a favorite shot from a movie.

Because hey... if I can make Dragon (or any other digital camera) look like the films I love, then that's good enough for me!

And seriously, if the end result is a LUT that actually makes RED look like film (or, say 10 LUTS for different lighting conditions), I would be happiest of all. I really don't want the expense of film.

Bruce Allen
www.boacinema.com

 

[Brian Boyer]

I think it's all associational. An aesthetic is the consequence of arbitrary methods (mediums give rise to methods) that are refined later. No one elected for film emulsion particularly; it presented itself as a sound reproduction of reality, and at the beginning of movies you have it looking quite different with very little dynamic range. All the same, people cowered in their seats when a train seemed to emerge from the screen in the Nickelodeon theaters. I.e. that arbitrary aesthetic was a blank slate (no association) and therefore defaulted to 'reality'.

Then people shaped what film was. The film curve we love was a development. I believe that through the late 70s to early 80s was a big change (if anyone has a good text on the history of film emulsion I'd appreciate a post reply or PM). Then video became strongly associated with 'newsreel reality' (the style of 'on the March' news reels employed in Citizen Kane for example) and that is another example of changing aesthetic as association changes.

I agree, kind of. I always say, in response to people bemoaning the demise of both film and vinyl records, they were the state-of-the-art forms of capturing the visual and auditory world around us for their time. They had a good run but technology advances and new methods emerge. Part of my point about video is that it's been around for decades and therefore had plenty of time to surface as the preferred medium for narrative storytelling. That didn't happen. Instead, it's look (and ease of use) made it better suited for other types of visual communication like news, sports, game shows, etc. People prefer to shoot and watch movies on film, not 60i video. That's why film's examination is a better starting point moving forward. Video had its chance and it didn't measure up.

As for the train pulling into the station, it wasn't really the aesthetic of film that was the blank slate. People had already seen printed reproductions of the natural world in still photographs. The projection of motion pictures was the blank slate that had more impact. The aesthetic of film at that time was far less important than the fact that audiences had never seen anything like motion pictures before, except in real life. The moving images created the association to reality.

The cowering was akin to an old trick where there's a rattlesnake in a glass case. Someone bets you that you'll move your hand when the snake strikes. You figure it can't bite through the glass, so you take that bet. You end up losing because, even though you know better, your instincts tell you there's danger and you move your hand involuntarily.

Perhaps try and state your contention in once sentence, but if you want to separate quality from look, well the quality of film is the way it looks and nothing else. Unless you mean 'quality' as a measure of value. I think a lot about this and I would say that what you're getting at is complexity, i.e. plasticity in the hands of an artist to mould it into something expressive. Film is quite complex in its effects. I think we are honestly above that threshold already. But here's my point: the history of video aesthetic will eventually develop into a new cinema aesthetic, influenced by other video mediums like video games and even ENG video, as seen in the found footage movies.

The phrase I used was "exact look of film", meaning you may not need grain or color characteristics of specific stocks in order to arrive at what we find most appealing about the filmed image. There's a scene in the Hitchcock trailer where a guy gets hit in the head with a shovel. That shot looks as good to me as any outdoor scene I've seen, from film or digital. I think it's possible to deconstruct the image film produces and separate out the most appealing aspect(s).

Maybe what makes film so appealing is how it handles light falling onto it. Let's say digital sensors eventually behave in the same manner. What we once considered the film look could simply be thought of as the characteristic humans find most appealing about an image that both film and digital now share.

It's true that a culture needs to be eased into a new aesthetic like a person learning a foreign language, but at some point it becomes conversant - if you learn it before you're too old you won't retain an accent. Digital audio is a prime example as you rightly point out - but for my point as well: look at all the in the 'in the box' music producers who don't even use instruments or even solid state synths - it's all digital. They tend to spurn so called analogue modeling systems. 90 percent of the synths on Inception are a software synthesiser.

90% of the synths on Inception may be from a software synth but what percentage of those are emulating either physical instruments or analog electronics versus creating sounds that one might conclude sprang forth purely from the digital domain and have no real world counterpart?

We humans love our organic imperfections and many of us keep coming back to them every chance we get, especially when advances in digital technology get us that much closer to it.

The most advanced and celebrated digital tools will probably be deemed so due to their ability to accurately replicate something that occurs in nature.

 

[Brett Harrison]

The sounds are all synth based (I have the patches from TDK). I think 'organic' is just a byword for complexity. You could argue that organic things are just combinations of the four genetic letters and therefore digital, albeit super complex.

Motion picture is an aesthetic.