The first RED Weapon 8K has Landed!

[Patrick O'Sullivan]

Jesus.

The transition from dark to light tonal values in that book still are fucking nuts. Very cool Phil

 

[Christoffer Glans]

The short story. I'm very happy. Without question this is the best sensor RED has made. I've never seen anything like this actually.

It's gonna be interesting to see how this looks compared to the Helium 8K, very exciting.

How do you think the 8K image looks rendered to 8K? I mean we all know it's gonna be a superb 4K downscale for it, but how do you feel about the 8K image without a downscale? Because when 8K screens comes out that mean we need to do an output without downscale, or a small downscale to Super UHD. So how does it look without downscale processing?

 

[Phil Holland]

It's gonna be interesting to see how this looks compared to the Helium 8K, very exciting.

How do you think the 8K image looks rendered to 8K? I mean we all know it's gonna be a superb 4K downscale for it, but how do you feel about the 8K image without a downscale? Because when 8K screens comes out that mean we need to do an output without downscale, or a small downscale to Super UHD. So how does it look without downscale processing?

That was day 3 test stuff. Exported a whole bunch at resolution and 150%, 200%, 300%, and 400% blow up for QC purposes. Also did a fun thing and exported 16 2K crops out on a timeline of the same footage. REDCODE RAW 5:1-12:1 tests for at resolution delivery.

It looks very good and is pretty nice. I'm preparing files to actually show on 8K screens next week. Hoping there's time to review them shortly.

There's the additional conversation of "what do you want your footage to look like at the end of the day" and you may or may not want to apply a sharpening pass. I did similar tests with Forged and actually was able to see that in 8K.

 

[Maik Müller]

That was day 3 test stuff. Exported a whole bunch at resolution and 150%, 200%, 300%, and 400% blow up for QC purposes. Also did a fun thing and exported 16 2K crops out on a timeline of the same footage. REDCODE RAW 5:1-12:1 tests for at resolution delivery.

It looks very good and is pretty nice. I'm preparing files to actually show on 8K screens next week. Hoping there's time to review them shortly.

There's the additional conversation of "what do you want your footage to look like at the end of the day" and you may or may not want to apply a sharpening pass. I did similar tests with Forged and actually was able to see that in 8K.

During such days RED should put a guy with a camera behind your back, it would be awesome to watch.

Thanks for sharing all the insights, quite happy to read you are that enthusiastic about the VV-sensor.

 

[Antony Newman]

Really liking the grain in high luma blue (pics 1 & 2)

Not sure about the low luma red saturation (pic 3) : Comparing mid top of image velvet to, mid left of (dead?) rose.

Caveat : Could be that I've got lots of super shiny metallic red junk on my desk.

AJ

 

[Christoffer Glans]

That was day 3 test stuff. Exported a whole bunch at resolution and 150%, 200%, 300%, and 400% blow up for QC purposes. Also did a fun thing and exported 16 2K crops out on a timeline of the same footage. REDCODE RAW 5:1-12:1 tests for at resolution delivery.

It looks very good and is pretty nice. I'm preparing files to actually show on 8K screens next week. Hoping there's time to review them shortly.

There's the additional conversation of "what do you want your footage to look like at the end of the day" and you may or may not want to apply a sharpening pass. I did similar tests with Forged and actually was able to see that in 8K.

That's nice, since one of the arguments getting 8K is the "super future proofness" of it, but that only works if 8K actually works without looking bad at the lack of a proper scale-down process. But then again, maybe the downscale process works differently the higher you get in resolution, meaning that a scale down from 8K to 8K S-UHD is enough for improving the image more so than shooting 4K and downscale to UHD in comparison.

 

[Björn Benckert]

That's nice, since one of the arguments getting 8K is the "super future proofness" of it, but that only works if 8K actually works without looking bad at the lack of a proper scale-down process. But then again, maybe the downscale process works differently the higher you get in resolution, meaning that a scale down from 8K to 8K S-UHD is enough for improving the image more so than shooting 4K and downscale to UHD in comparison.

I think no matter how you slice it, the more pixels you start with the easier it is to make a good looking image. Does not matter if end result is viewed in HD, 2k UHD or 8k the more you have at the starting point the better it will hold up. So simply those super underexposed looking like DV in the dark kind of shots done on a RED ONE MX at 4k will possibly fly by as good looking dark shots if shot with the 8k weapon and off course booth will look more similar the smaller resolution you view them in and the difference will be more and more apparent if views on a big screen with lots of resolution.

Same with film. 8mm 500ASA does not hold up so well, 50ASA works better, 16mm 50ASA works even better 35mm 50 ASA works really good as the lens factor makes less difference. 70mm 50ASA looks pretty flawless. Then ofcourse on the big screen the difference is very apparent. And on a small monitor it´s less apparent, but still noticeable for many.

I think Helium will be like shooting S35 with a super fine grain structure. Saying it would be like 50ASA is a bit misleading as I think it´s light sensitivity will be far more. But, that we do not know of yet. It might be that the Helium is less sensitive to light than Dragon. High light sensitivity is never a bad thing but one can argue that 800 ISO is rather high and for most things it would be no problem working with a camera that was far less sensitive if it would bring on other benefits. And the smaller the pixels the harder it is to get them sensitive.

I think if it was super light sensitive like native iso 2000 or so then yes Jared would have a hard time to keep quite about it

 

[Antony Newman]

Random thoughts:

If as sensor with 5 um Dragon pixels was replace but one with 3.65 um Dragon pixels:
+) We would have 5^2 / 3.65^2 more pixels
+) I guess ~1.8 times less noise : https://en.m.wikipedia.org/wiki/Oversampling
+) I speculate this would not change the 'native' ISO ... But broaden the range of usable both above below.
+) The noise reduction on the top end would reduce the point before hard clipping (small amount of DR increase)
+) The noise reduction on the bottom end would increase the acceptable level of ISO up to 1.8 the existing recommendation.
+) I would therefore completely speculate that the new range would 'go up to 2000 x 1.8 = ISO 3600

If the Helium pixels filter out more chroma noise than Dragon .... I would guess the chroma noise 'blob' would reduce creating a sharper chroma image in low luma. Ie a reduction in the Vector Scopes of : http://www.dvinfo.net/article/ultra-hd/dragon-color-does-tungsten-or-daylight-make-a-difference.html

I am going to guess (couple be complete 'hot air on my part) that as a result of controlling the chroma noise - that's the new sensor will have 1 more stop of DR, and 0.8 stops of increase ISO range on the low end - But will come with a Weaker STH like OLPF.

AJ

 

[luigivaltulini]

Random thoughts:

If as sensor with 5 um Dragon pixels was replace but one with 3.65 um Dragon pixels:
+) We would have 5^2 / 3.65^2 more pixels
+) I guess ~1.8 times less noise : https://en.m.wikipedia.org/wiki/Oversampling
+) I speculate this would not change the 'native' ISO ... But broaden the range of usable both above below.
+) The noise reduction on the top end would reduce the point before hard clipping (small amount of DR increase)
+) The noise reduction on the bottom end would increase the acceptable level of ISO up to 1.8 the existing recommendation.
+) I would therefore completely speculate that the new range would 'go up to 2000 x 1.8 = ISO 3600

If the Helium pixels filter out more chroma noise than Dragon .... I would guess the chroma noise 'blob' would reduce creating a sharper chroma image in low luma. Ie a reduction in the Vector Scopes of : http://www.dvinfo.net/article/ultra-hd/dragon-color-does-tungsten-or-daylight-make-a-difference.html

I am going to guess (couple be complete 'hot air on my part) that as a result of controlling the chroma noise - that's the new sensor will have 1 more stop of DR, and 0.8 stops of increase ISO range on the low end - But will come with a Weaker STH like OLPF.

AJ

+1
i complety agree with you,
for me about 17 stop, 3200 iso same 2000 of Dragon now + better skin tone and better color. And is possible more frame rate at lower resolution.

 

[Björn Benckert]

Random thoughts:

If as sensor with 5 um Dragon pixels was replace but one with 3.65 um Dragon pixels:
+) We would have 5^2 / 3.65^2 more pixels
+) I guess ~1.8 times less noise : https://en.m.wikipedia.org/wiki/Oversampling
+) I speculate this would not change the 'native' ISO ... But broaden the range of usable both above below.
+) The noise reduction on the top end would reduce the point before hard clipping (small amount of DR increase)
+) The noise reduction on the bottom end would increase the acceptable level of ISO up to 1.8 the existing recommendation.
+) I would therefore completely speculate that the new range would 'go up to 2000 x 1.8 = ISO 3600

If the Helium pixels filter out more chroma noise than Dragon .... I would guess the chroma noise 'blob' would reduce creating a sharper chroma image in low luma. Ie a reduction in the Vector Scopes of : http://www.dvinfo.net/article/ultra-hd/dragon-color-does-tungsten-or-daylight-make-a-difference.html

I am going to guess (couple be complete 'hot air on my part) that as a result of controlling the chroma noise - that's the new sensor will have 1 more stop of DR, and 0.8 stops of increase ISO range on the low end - But will come with a Weaker STH like OLPF.

AJ

True but smaller photo sites simply get hit by less light. So even if the increase of resolution improve things the smaller photo sites need to be improved by 1.8 to catch the same amount of light no?
So if the helium sensor does not have 1.8 more light sensitive photo sites than dragon then simply it will not be as light sensitive.

But as it is a later generation sensor all or anything can be improved. So who knows. and yes more pixels add more usable stops but the less surface for each pixels also work against that. So only way to know for sure is to try of course.

 

[Phil Holland]

Just a smidgen of an update. I'm a few weeks into this shoot and we've got some bad weather and I have a few moments to jot down some things.

First, a BIG thank you to the RED Team for getting the DSMC 8K VV Canon EF Mount MG into my hands. I'm mixing EF and PL glass on this shoot, like hell bent in doing so. The new mount is pretty much exclusively for 8K VV sized sensors and should help increase contrast and in some cases even coverage perhaps, I can also see how it might create a more ideal flare perhaps. I'll need to do testing to show a bit of a side by side. Not the cheapest thing on earth, but something I've wanted for my 8K VV because I'm glass weirdo.

Here's a picture of the bigger opening:



You can see the bigger port with a different light trap design. Pretty nice actually.

That was my last pickup from RSH before heading out to the great unknown. It's been sort of a tradition of mine before every shoot hitting the RED Store and grabbing a bunch of things.

If you're ever curious what RSH looks like from above:

You can just make out the RED logo above the very compact and safe vehicle


As I mentioned we've had a spit of bad weather at this current location. I've been confined to my hotel room mostly and tackling schedule issues, however, there was this massive thunderstorm:

I setup the 8K VV with the Zeiss Otus 28mm and did a bit of timelapse'n as the storm grew nearer and the distance between thunder and lightning shrank to nothing. Then it occurred to me this might be the first time anybody has shot lightning with one of the 8K cameras.


I have a lot of thoughts about 8K VV Dragon and 8K S35 via Helium and truly just 8K in general. Seeing such tremendous richness in the 8K images is mind blowing. Hoping when this shoot is down I'll get a good amount of time to test both out.

More stuff soon. A bit of scouting today.

 

[Phil Holland]

Was there ever any R3D's uploaded from this cam?

Mine in particular, no. But since I'm stuck in my hotel room:

http://www.phfx.com/temp/reduser/phfx_myUglyMugHotel8KVVZeissOtus28.zip

I have the Low Light Optimized OLPF in there at the moment.

 

[Karim D. Ghantous]

2K is enough.

/sarc

But seriously, this is great stuff. Thanks for the effort you're putting into showing us all this material.

the only thing that comes close to the resolving power of the 8K FF frame is something around or larger than 65mm 5-perf depending on who's juice you're drinking regarding what film is. It is actually closer to properly scanned IMAX 15-perf.

Many years ago I saw the output from the Leica M8 compared to Portra 160. I was amazed at how clean the Leica was and how grainy Portra was. The newer emulsions are better but digital sensors, as long as they have adequately sized photosites, are always going to be cleaner.

I would provisionally disagree with your comparison to 15-perf 65mm. There is no way that IMAX frames have less resolving power than 8K. But maybe it's closer than I know.

Who do I need to sleep with?

It's 'whom'.

:leaving:

Just some light sharpening

'Light' sharpening? I thought that was ADD!

Zeiss Otus 28mm f/1.4 covers 8K FF

I assumed it would, based on people's experiences trying it on MF sensors. The better the lens, the more extra coverage it seems to have. It may not be a lot, but it's enough. I would have to assume that Leica SL primes would be similar. It's also worth pointing out that RED VV is slightly shorter than 8-perf 35mm, which might allow the rectangle to almost fit within the image circle of SLR lenses.

I'm personally more excited by the Helium than by the 8K Dragon. Although, that's mainly because I would like more flexibility with lens choice. Of course, those who think that a smaller sensor will mean a lighter kit might be disappointed, as many cinema lenses are quite large. AFAIK the Schneider Xenons are not much larger than the Xenars, for example. And they don't have to be quite as sharp.

I would like to repeat a prediction that I've been making for a while: the next step for cinema is IMAX. It's not 3D or HDR or any of that buzzword BS. It's just huge, massive screens where even low budget productions can have a huge impact. Imagine a properly written Dan Brown adaptation set in Rome. RED is perfectly positioned for that. As much as I love film, there would be no way that I'd ever want to shoot 15-perf 65mm.

And how odd it is that in the same year that we're getting 8K and Helium, Kodak is releasing a Super-8 camera. I can see that it is getting a lot of positive attention. These two systems couldn't be any more different.

 

[PatrickFaith]

Thanks so much Phil. I was pondering on the dragon color "1" vs dragon color 2 - on these files I prefer the color of dc 2 but the purple patterns of dc 2 drive me crazy until I reduce the frame to 4K(this problem doesn't occur in dc1). What would you think of a option on dragon color 2 so the noise is like dragon color 1 (doesn't have the wavy veins) but the color is like dragon color 2? Either way, I find it confusing a bit for film specs to flip between the two.

 

[Stacey Spears]

on these files I prefer the color of dc 2 but the purple patterns of dc 2 drive me crazy

Patrick, you should try the new REDWideGamutRGB in RCX-P build 42. It should no longer have that purple pattern. You will need to enable legacy as it is not exposed by default. It is at the bottom of the list, just below XYZ.

 

[Phil Holland]

Epic first R3D post!! Thank you for all you do

Thanks Zeb.

I'll share more .R3Ds when I'm back from this shoot. I basically go the 8K VV and jumped right into the heat of this thing. Enough time to test, not enough time to construct much else.

Thanks so much Phil. I was pondering on the dragon color "1" vs dragon color 2 - on these files I prefer the color of dc 2 but the purple patterns of dc 2 drive me crazy until I reduce the frame to 4K(this problem doesn't occur in dc1). What would you think of a option on dragon color 2 so the noise is like dragon color 1 (doesn't have the wavy veins) but the color is like dragon color 2? Either way, I find it confusing a bit for film specs to flip between the two.

Are you referring to the noise pattern you see in the noise floor? Like if you lift it up 5-8 stops? Or something related to edges?

I also mostly prefer the look of DRAGONcolor2. I find it provides a truer color, richer more vibrant blues, and generally more pleasing skin tones. The key thing with DC2 is to know when not to use it. Mainly I avoid it when shooting things like wide shots with punchy neon signs. Also, if I'm shooting recklessly or with certain lenses that exhibit more chromatic aberrations I'll sometimes go for DRAGONcolor.

Color Science/Color Space is a starting point. One thing I've done and you've seen these from me in the past is shooting various color charts and subject matter and matching them. This allows me to make a moderate 3D LUT that bring whatever Color Space I'm using closer to the the other. The same can be said for the subtle differences found between the different OLPFs.

[/COLOR]

Patrick, you should try the new REDWideGamutRGB in RCX-P build 42. It should no longer have that purple pattern. You will need to enable legacy as it is not exposed by default. It is at the bottom of the list, just below XYZ.

Agreed as it's a very interesting starting place now. Most will find the need to punch up the saturation a bit when getting started. If using REDCINE-X PRO, Post:Look:Effect:Saturation with a value of 1.3-1.5 should provide something pretty fun to look at. REDWideGamutRGB's introduction aslo provides a glimpse at some interesting potential as well.

 

[Scot Yount]

Phil, thanks for posting this. I am curious about why D.E.B. is grayed out as
an option in REDCINE. I have the latest build. 8K VV is a Dragon sensor, so shouldn't D.E.B. be available?

 

[Mike P.]

Those are the first 8k VV .r3ds I've gotten my hands on... just... insane.

 

[Charles Bergquist]

Thanks for these Phil!

Hope this VV sensor production ramps up!

 

[Darius Lyles]

[/COLOR]

Patrick, you should try the new REDWideGamutRGB in RCX-P build 42. It should no longer have that purple pattern. You will need to enable legacy as it is not exposed by default. It is at the bottom of the list, just below XYZ.

I just downloaded the BETA build but I can seem to find the REDWideGamutRGB