RED Dragon vs. 65mm Film cont'd

[Dr. Sassi]

...
From the perspective of somebody who has shot and dealt with using *gasp* Solitaires for 65mm film recording, yes I would say the Red workflow is infinitely smoother and faster than 65/70mm. It's been a few years since I've done that work, but I think a full 15-perf IMAX image took a bit over 1.5 minutes to record to film per frame when exported digitally! There's only a handful of people in the world who still do that work and I do admire their determination and expertise. 5-perf 65mm laser recorded output (and scanning) is much, much faster. It's still a chunky workflow in terms of getting footage scanned and exported nicely. There are also camera related issues often to shooting 65mm. I remember seeing some of the light contamination on some of the more recent Batman films that needed to be fixed in post.

It's a gorgeous format in reality. It's wonderful that it's being celebrated by high end/high profile productions. Those people are wild and sadistic

I left the world of film capture when it comes to my camera selection and work. Much of that history though is what led me to Red in the first place. The Red One made me look and shoot with it, but by the time Epic existed and was available about 3 years later I couldn't say no to that.

Thanks for the additional information, Phil. Certainly: you have an interesting background! I really like to read your post here.
It is certainly an exciting work atmosphere to hear the 15 perfs running though the camera, to feel how much value each single frame finally has.

My last "celluloid" was a direct print from a practical+CG short (35mm/Technicolor), certainly a pristine quality, but so much effort... compared to anything digital ;o) -- I can't even imagine to would have done all of that for 65/70mm...

 

[Les Hillis]

Looks great Phil. I remember the sharpening tests you posted here when grading 'It's Just Paint'. I was and am impressed by the level of clarity your processing gives the images. Noticed it in some of the stills you've posted too. Hope we get to see more examples of it in your future projects.

 

[Stacey Spears]

Phil, what did you use to sharpen your image? I still need to look into Sapphire.

 

[Phil Holland]

Looks great Phil. I remember the sharpening tests you posted here when grading 'It's Just Paint'. I was and am impressed by the level of clarity your processing gives the images. Noticed it in some of the stills you've posted too. Hope we get to see more examples of it in your future projects.

Oh there will be samples soon

Phil, what did you use to sharpen your image? I still need to look into Sapphire.

This is a process I use through Adobe After Effects CC, not Sapphire though (however that's a great set of plugins). I have many post methods through different programs. I'd start with the sharpening options available in Redcine-X Pro first to experiment as it's quick and easy.

Occasionally these days when I finish material, if I'm "really going full ham" I'll do a final QC pass in AE and clean up things like chromatic aberrations if there are any. Might do a few other things. The piece I'm working on right now has a weird blemish on the wall I need to remove. Certain things can be done in things like Resolve as well.

Most of the time though I really don't need to do this at all for motion work. It's more about if you are trying to carve out more detail and perhaps a more pristine image.

 

[Dr. Sassi]

Not to noise up the thread too much with saying thank-yous, but: Thanks again for the details. Even I'm very interested in this stuff since long, I never saw any discussion about a 12K scanner, or better why they won't exist. Yes, when I scan single SLR-"film-frames", I reach very quickly the grain barrier. Sometimes nice and sometimes not at all.

You deliver so many details, with a great deal of background experience each time, that I feel, discussions about Film vs 6K acquisition might fade in the near future more and more.

 

[Les Dittert]

Phil,, thanks for posting more evidence of today's digital blowing away film.
If you like, share an r3d and I'll do something detailed with it.

 

[Jeremy Saville]

Now this is more like it. You look great in a turban Phil. In fact it looks like a "Phil of Arabia" sequel could be in the offing (Shot on Dragon naturally), but more likely it'll be "Phil of Korea" considering...

 

[Maik Müller]

Thanks Phil. Awesome insights and comparisons. Puts the Dragon into the right perspective, it's really something new, better than film and simply one of the biggest achievements in cinema history. Underlines what Jim said, he wasn't exaggerating things. Oh and yes, i miss his enthusiastic midnight-posts too.

Now what would be nice, to get the last bit of a "bigger look", is one of those focal reducers, specifically for Dragon and 70mm-cine-lenses and/or Medium Format still lenses. But i guess RED has already something else in the oven (not hard to guess if you remember those early roadmap-pictures).

 

[Phil Holland]

Now this is more like it. You look great in a turban Phil. In fact it looks like a "Phil of Arabia" sequel could be in the offing (Shot on Dragon naturally), but more likely it'll be "Phil of Korea" considering...

Heh. I may sport that look at NAB this year!

Phil, when ADD for motion is enabled, do you have a sense it will upscale to an even higher resolution than even 6k acquisition?

This was my great hope for ODEMAX... that is, encoding to Redcode in different formats for different venues ranging from Vista Vision sized outdoor theaters (some sort of variant of a drive-in movie) to a small art-house screen. With Dragon, we have the cameras... all that's missing is new exhibition venues. At least, that's the way I see it.

ADD cleans up the image more and especially benefits higher ISO work. Technically speaking it would allow for better scaling. However, in practice if uprez'in you may find better results or more time efficient results with different methods. Combining those techniques with ADD might be worth experimenting with though. We'll see.

Interestingly enough when it comes to going from 6K to 8K you can create some pretty damn good results using various methods.

As for ODEMAX. My heart hurts thinking about it's potential and lack of existence thus far. The REDRAY technology and .RED are very much a good thing though and likely we'll see or hear more about their future in a couple days.

I will say this. Times are a changin'. Comcast/Timewarner Cable haven't exactly showed their full skin yet in the UHD game. Netflix has shown what they are up to. DirectTV is dropping hints here and there. Google has 4K through Youtube and they are pushing hard and fast towards VP9. The X Factor is various SmartTV features and apps and how people can get content into them. There's also various other UHD devices and services coming that we have and haven't seen. One of which I've contributed work to.

In this way as we race towards our UHD future much is up in the air for content distribution. Right now it's a mountain of hay and nobody is doing it "right and proper". However, whoever has the content and delivers it to the viewer in the best way will win out. So that to me means a focus on quality content and a focus on user experience/interface design. I have very strong thoughts on these matters and more or less how to implement them into consumer devices.

Can't wait to see what new developments NAB brings. I have a REDRAY hooked up to my 70" UHD 4K Sharp television here that also gets a 4K Youtube and Netflix right on the display. I'll be adding the Nuvola NP-1 and a Playstation 4 into the mix soon. I also can hook my PC straight into the device and achieve 4K playback. What's next and what's the best are really yet to be seen.

I will say this from a secure theatrical and media delivery perspective I like .RED, for web and general broadcast needs VP9 and H.265 have been promising.

If you missed it, here's a UHD 4K Google VP9 sample I threw together:
http://www.reduser.net/forum/showthread.php?111230-Google-VP9-4K-HD-Sample&highlight=

I think it was the first 4K VP9 sample available on the net actually.

Thanks Phil. Awesome insights and comparisons. Puts the Dragon into the right perspective, it's really something new, better than film and simply one of the biggest achievements in cinema history. Underlines what Jim said, he wasn't exaggerating things. Oh and yes, i miss his enthusiastic midnight-posts too.

Now what would be nice, to get the last bit of a "bigger look", is one of those focal reducers, specifically for Dragon and 70mm-cine-lenses and/or Medium Format still lenses. But i guess RED has already something else in the oven (not hard to guess if you remember those early roadmap-pictures).

Some have criticized Jim and Red in the past about overselling their products. In Dragon's case I feel this is pretty much more or less dead on the money in terms of what Red was hoping to achieved and what they are actually delivering.

I said it once and I'll say it again.

Red's initial goal was to produce a film alternative in a digital cinema camera system. Dragon however is in my mind a film replacement and exceeds Red's initial goals. I've enjoyed the Red cameras over the years. Especially what Epic w/ Mysterium-X has to offer. But Dragon really made me feel comfortable with my investment. The image quality is solid and good. Beyond all the tech talk and testing it feels right. That is what's most important to me.

Phil,, thanks for posting more evidence of today's digital blowing away film.
If you like, share an r3d and I'll do something detailed with it.

Cheers Les. I'll get you some material after NAB. I have like maybe 12 hours on this edit to go before I leave for Vegas!

Not to noise up the thread too much with saying thank-yous, but: Thanks again for the details. Even I'm very interested in this stuff since long, I never saw any discussion about a 12K scanner, or better why they won't exist. Yes, when I scan single SLR-"film-frames", I reach very quickly the grain barrier. Sometimes nice and sometimes not at all.

You deliver so many details, with a great deal of background experience each time, that I feel, discussions about Film vs 6K acquisition might fade in the near future more and more.

Thanks again. For me the discussions between film and 6K Dragon are sort of ludicrous at this point. Dragon beats out film with it's technical merits and performance. However, if we are talking about the general character of film as a format that of course it is still a preference to some. However, I am confident in saying you can make this Dragon material and making it more "film like" through processing if that's desired. Just really matters what you want your material to look and feel like at the end of the day.

 

[Sergio Perez]

Love this post Phil. Love the camera, the system. A true revolution. Been here for 3 years and hope for many more.

 

[Eric Z]

Nice write-up, Phil.

Looking at your 6K grabs, I can't help but wonder if we could get rid of the OLPF altogether, both for stills and motion.
Despite its obvious benefits, it introduces "blur" to the images, and then you need to sharpen up in post...but it's not as good as the real thing.
I mean, people here are investing a lot of money on good sets of cine lenses, which provide excellent sharpness, only to lose it due to the OLPF.

Just saying...

 

[Alain M]

Awesome!!!

 

[Phil Holland]

Nice write-up, Phil.

Looking at your 6K grabs, I can't help but wonder if we could get rid of the OLPF altogether, both for stills and motion.
Despite its obvious benefits, it introduces "blur" to the images, and then you need to sharpen up in post...but it's not as good as the real thing.
I mean, people here are investing a lot of money on good sets of cine lenses, which provide excellent sharpness, only to lose it due to the OLPF.

Just saying...

Well that's just the thing. You don't "need" to sharpen in post. In fact many prefer not to as often sharpening interferes with the natural sensation that motion blur produces. Also, if finishing for 4K the benefits of scaling down to 4K from 6K obviously increases micro details.

The key to all of this is deciding how you want to finish your motion work. It's not very common to sharpen up footage in post in the industry as usually everybody is chasing a more organic feel. I do sharpen and "do things" to some of my material. Those birds flying around are a good example as I wanted that to be almost an ink drawing in terms of detail.

The OLPF isn't really hurting the resolution that much. It is protecting you from other bad things like edge aliasing and color aliasing, which you really don't want honestly. Especially over the course of a production.

You certainly have the option of "shooting naked" without the OLPF, but I don't really see a reason why honestly. Leave it the way it is, scale down, or sharpen up. The choice is there for you to decide.

 

[Hans von Sonntag]

You don't "need" to sharpen in post.

Phil, we all know that Red delivers high resolution pictures. But they are, in comparison to others like Arri for instance, visibly softer. This may be seen as "the red-style" but obviously many are seeking for a higher perceived sharpness. Interestingly, some are trying to get rid if the OLPF searching for the "right", "natural" "non-softed" sharpness. Generally one does not need to sharpen Red originated pictures in post because their resolution is unquestionable compelling and delicate. However adding a tad of sharpness in post (with the right tools, such as the sharpen nodes in ADSK Flame/Smoke) adds a lot of vibrancy and conciseness to the pictures without videoish fringing.

An OLPF that delvers maximal sharpness but holds down aliasing sufficiently is what red and other manufacturers strive for when designing the OLPF. My wild and unconfirmed guess is that in Red's case the OLPF is not only tuned for the right amount of anti-aliasing vs. sharpness but also to prepare the RAW data in an optimised way to allow for an artefact-free compression. We all remember the tree of death that broke Red's codec for quite a while (I think this were the build-16 times). A softer pictures is less artefact-prone when compressing it than pictures with high details. Shooting 3K with Epic MX and RC-8 will introduce lots of compression artefacts in the shadows in long-shots scenes that show parks, bushes etc... Detail and sharpness do play a vital role when it comes to high quality compression. Red RAW is always compressed. IMHO, one of the selling points of Red but, of course, something that has an impact on quality. Sharpness might play a role here.

Hans

 

[Phil Holland]

But they are, in comparison to others like Arri for instance, visibly softer.

I'm not entirely sure about this actually. I've done side by side comparisons and there have been many side by side comparisons here on Reduser and elsewhere. Dragon's OLPF is considerably "weaker" than Mysterium-X's. There literally have been resolution and aliasing charts shot with with several systems at this point and I don't fully agree with that above statement at all when it comes to Dragon versus other Bayer CMOS systems.

Also, Dragon's noise floor is different/lower, pixels are smaller, and how it records REDCODE is fairly different. And yes the OLPF has been tuned to that pixel pitch. In the sample clips posted by a number of folks at various compression ratios you can see that at higher compression ratios Dragon holds up better against lower tones, chroma, and high frequency detail.

I'm not saying "not to" sharpen. I'm saying you have a choice. I have posted in this very thread sharpened and unsharpened output.

 

[Les Dittert]

Phil, I'm driving to Vegas tonight as well . I hope to chat with you there. I'll be the guy with the red colored side bag with an ipad in there containing examples of Dragon vs Epic_nLPF comparisons among other neat stuff .

 

[Les Hillis]

...

If you missed it, here's a UHD 4K Google VP9 sample I threw together:
http://www.reduser.net/forum/showthread.php?111230-Google-VP9-4K-HD-Sample&highlight=

...

Wow, I did miss that.

Had to download the 'Apprentice Video' player to watch it - http://sourceforge.net/projects/apprenticevideo/ , but it was worth the effort.

That's some of the best looking Dragon footage I've seen yet. Perfect examples of pushing the sharpness about as far as you can go without making it a stylized kind of effect.

The VP9 compression looks good too, did you have to tweak it's rendering settings much or was that how it works by default?

 

[Robert Ruffo New]

Dragon is amazing, and is in my opinion better than any other digital camera (I finally like the color of Red as much as Arri's) but honestly I still prefer the look of film. It's not all about resolution, it's about images that work and make actors look good and move the story forward, whether that means slightly less or more pixels. Film just pleases me, and is easiest to grade in some ways because the color info is so rich.

 

[Phil Holland]

Phil, I'm driving to Vegas tonight as well . I hope to chat with you there. I'll be the guy with the red colored side bag with an ipad in there containing examples of Dragon vs Epic_nLPF comparisons among other neat stuff .

Cheers Les. I'll see you here.

Wow, I did miss that.

Had to download the 'Apprentice Video' player to watch it - http://sourceforge.net/projects/apprenticevideo/ , but it was worth the effort.

That's some of the best looking Dragon footage I've seen yet. Perfect examples of pushing the sharpness about as far as you can go without making it a stylized kind of effect.

The VP9 compression looks good too, did you have to tweak it's rendering settings much or was that how it works by default?

Yep. The codec is even a bit more robust now and I can squeak out even better material. There are no "standard settings" yet for VP9 as this is still pretty hard to do. It's crazy slow to encode still as nothing has been optimized just yet, but soon. Also, it needs to be added to 3rd party applications and all that. Currently nothing, though I expect news tomorrow. I have some 4K VP9 Dragon footage playing here and there at NAB. Should see a few places showing off some .R3Ds too.

Heh. Those leaves are some of the first footage I shot with my Dragon when I received it.

Dragon is amazing, and is in my opinion better than any other digital camera (I finally like the color of Red as much as Arri's) but honestly I still prefer the look of film. It's not all about resolution, it's about images that work and make actors look good and move the story forward, whether that means slightly less or more pixels. Film just pleases me, and is easiest to grade in some ways because the color info is so rich.

Nice Robert. I'm curious to see what you say when you get a few more Dragon samples your way on the grading flexibility.

 

[Will Keir]

That makes 5,000 of us. Jim is missed and Jarred is catching on quickly.