Iso 8000...

[Jannard]

This is ISO 8000 and 180 degree shutter (not 360) shot under tungsten light (the most difficult). We think we can still do better, but this is pretty amazing stuff. Shot with RED 18-50mm T3 zoom on "Acapulco".

Jim

http://www.reduser.net/forum/uploaded/9_1184356818.jpg

... spots on the right is paint splatter on a piece of aluminum. I didn't want to alter the image, so we left them in the shot.

[Casey Green]

wow, impressive! - thanks for posting the actual image, Jim.

[Charles Perkins]

your joking right?

AWESOME!!!

[Alex Wengert]

Interesting Jim. I'm assuming that this is shot wide open on the lens.

[Keith Nealy]

Wow...
Things just keep on getting better.

[Keith Nealy]

Jim, was this image adjusted in post?

Like you said previously, "lifted out of the dirt?"

[Jannard]

Interesting Jim. I'm assuming that this is shot wide open on the lens.

No. It was exposed for ISO 8000 (f22). Normal exposure at ISO 250 was T4.

Jim

[sandesh]

Amazing!

[Jannard]

Jim, was this image adjusted in post?

Like you said previously, "lifted out of the dirt?"

Yes... and we used a spoon, not a shovel.

Jim

[Finner]

The noise at 8000 gives it an almost super16 kind of look. I wonder if the 4000 gives a bit of a 35mm look.

[Joe Aurili]

Wow, that is quite nice considering. It is sized a little below 720p so 8000 might be a stretch for pro work :)

[canred]

What did it look like before post? Another question is what did the 4000 image look like before and after?

Tom

[Rainer Fritz]

I can only say congrates to you and the team... sending you kisses my dears...

[Jarred Land]

Wow, that is quite nice considering. It is sized a little below 720p so 8000 might be a stretch for pro work :)

this is scaled down from the original 4k image, it was not shot at in 720. and remember, this is a jpg.

[Casey Green]

no noticeable FPN either from what I can tell when zooming in. :biggrin:

[Harmonica]

Holy cow. My digital still camera (albeit a cheap one, I happen to be saving up for a digital cinema camera...) has twice as much noise under ideal conditions!

[Adam Jeal]

Jim & Jarred,

This is amazing! - It's made my day. I'll never shoot S16 or 35 again! :ohmy:

Thanks! :biggrin:

[Brook Willard]

This is absurd... I haven't seen it on a computer monitor yet (I'm on my phone), but damn. How on earth did you guys build this sensor?

I'm loving these updates as of late. :)

[Joe Aurili]

Understood, I was thinking of what it would be like shooting at a high ISO and scalling down to 720p to eliminate noise. Good point on the JPG. It looks like a shot from my Z1, but the Z1 had much more light to work with.


this is scaled down from the original 4k image, it was not shot at in 720. and remember, this is a jpg.

[Adam Jeal]

This is absurd... I haven't seen it on a computer monitor yet (I'm on my phone), but damn. How on earth did you guys build this sensor?

I'm loving these updates as of late. :)


Graeme's 'magic pixies' did it Brook! :) :) :) Bless them!

[Mark L. Pederson]

Bye bye Viper!

[Adam Jeal]

Bye bye Viper!

Indeed Mark!

[Mark L. Pederson]

Can't wait to take #6 & #7 out for a midnight Manhattan stroll ....

[Ace]

This is really hard to believe. That cant be an ISO 8000 image at f22. If it indeed is.. do you know what this means!

[Adam Jeal]

Can't wait to take #6 & #7 out for a midnight Manhattan stroll ....

Stop! :) i'm green with envy. I'll come over from Scotland and help out if you like Mark :ninja:

[Priyesh P.]

What is that stuff over "tungsten iso 8000" ?
Dead pixels? Or just some dirt on the colour chart ?:-)

[canred]

Jim stated that was paint on the aluminum and they left it there because they did not want to alter the image.


Tom

[Bruce Allen]

Thanks for posting images, Jim! Looks good and bodes well for not only available-light stuff but high speed shooting and fx plates as well...

Bruce Allen
www.boacinema.com

[Michael Schrengohst]

Wow, I would love to see a .tif file and I would love to
see a proper exposure for comparision. I guess you
can't show to much. It might dampen sales of
those other guys cameras.

[Adam Jeal]

Wow, I would love to see a .tif file and I would love to
see a proper exposure for comparision. I guess you
can't show to much. It might dampen sales of
those other guys cameras.

I don't think Jim's gonna be too upset about that :) !

[Alex Boothby]

Excellent - a big improvement over the November David Stump tests. Back then 4 1/2 stops under was much noisier and warmer with color impurities - and even that was beyond awesome. Now we're at 5 stops under (I think). This is super crazy.

[Alex Boothby]

I would love to see a proper exposure for comparision.

For those interested in a "before and after" comparison, this was from the above mentioned Stump tests:

http://www.reduser.net/forum/showpost.php?p=57970&postcount=46

(...ya I'm still pimping that chart - it brings me great glee to see that super dark image recovered to something useable).

[Adam Jeal]

Un-)@$^$(@%&(%$-believable...

I'd say that's far more than "lifting out the dirt." The bubble-blowers are really going to like this one.

Mysterium rocks!

All together now! - 'I'm forever blowing bubbles..'

[donatello b]

T3 doesn't seem so slow anymore !!

[Adam Jeal]

The noise at 8000 gives it an almost super16 kind of look. I wonder if the 4000 gives a bit of a 35mm look.

That sounds right on the money to me!

[JohnF]

Bloody hell!

Okay I've got to know...

This is the standard imager, right?

ISO 8000 can be setup on a normal RED, yeah?

You can just process this using REDCINE, yes?

This was captured using REDCODE, please say yes?


JohnF

[Jason Francois]

I know that proper lighting is an art form and one that I greatly appreciate, but with this camera doing shoots with slightly-altered natural lighting, when needed seems more than within reach. Amazing!!!

I can't wait to shoot some desert sunsets.

[Chris Forbes]

I suspect there is more to be had in that image. Just from a little tinkering.


http://www.thevideodepartment.net/Images/8000iso.jpg



The blacks maybe a tad crushed. But I think there is a happy medium in there somewhere.

[Häakon]

Bye bye Viper!
Haha, bye bye just about everything! Good grief.

I know that posting a "less than pristine" image for the entire world to scrutinize is difficult, but knowing that this is ISO8000 is unreal. A 4K image would be nice (I'm not trying to be picky, but downrezzing that much inevitably reduces noise), but regardless, it's obvious how amazing that is. Couple it with the fact that a lot of people will probably be finishing to DVD or (eventually) Blu-ray as it is, and that's a similar kind of "real world" result one may be able to expect. What this really says to me is how clean the imager must now be if exposed properly! And they said it wasn't a revolution... :bleh:


T3 doesn't seem so slow anymore !!
I hear that. When are those RED primes gonna be available again??

[Sanjin Jukic]

All these amazing things are happening just a night before I leave to my HOT ISLAND. But I'll get online from there too.

[Emanuel A.]

Interesting, very interesting indeed!

[Laco Zamba]

Now it's time for overexpose test.

[SF Geek]

All right Jim, you win. From what I see I would consider ASA 4000 more than "Useable".

[Adam Jeal]

I suspect there is more to be had in that image. Just from a little tinkering.


http://www.thevideodepartment.net/Images/8000iso.jpg



The blacks maybe a tad crushed. But I think there is a happy medium in there somewhere.

Nice work there Chris. I think that Sony and Thomson will require a 'Happy Medium' to hold a special seance for them, to tell them all about the future. Picture it -

Dude in a smock at Sony HQ: - ' I've see something..'

Sony Dudes: 'Yes..'

Dude in a smock at Sony HQ: 'The Future..'

Sony Dudes:(going mental) 'YES!!!!..'

Dude in a smock at Sony HQ: 'It's Red!..' :)

[albert rudnicki]

It's really amazing!
I can lit my scene with a lighter...
--------------------------------------
www.yayofilms.com

[ATF]

Hello,

Fantastic. I shoot a lot of stuff at very low light levels and high ISO's with my Canon EOS 1ds and this is just amazing.

Can't wait!

ATF

[Adam Jeal]

Haha, bye bye just about everything! Good grief.

I know that posting a "less than pristine" image for the entire world to scrutinize is difficult, but knowing that this is ISO8000 is unreal. A 4K image would be nice (I'm not trying to be picky, but downrezzing that much inevitably reduces noise), but regardless, it's obvious how amazing that is. Couple it with the fact that a lot of people will probably be finishing to DVD or (eventually) Blu-ray as it is, and that's a similar kind of "real world" result one may be able to expect. What this really says to me is how clean the imager must now be if exposed properly! And they said it wasn't a revolution... :bleh:


I hear that. When are those RED primes gonna be available again??

This rocks. Team the Red system with tools like Dedo's, Lowels and Lite-Panels (and the odd 1 & 2 K HMI), decent matte-box, FF (please Curt :) )and then add cool colour correction tools like Stu Maschwitz' Colorista along with the new Magic Bullet looks on a new shiny Mac Pro and add talent.. you will have a force to be reckoned with. Color will also be a contender if Apple can fix it and smooth out the GUI.

[Adam C Lubkin]

Consider that just over a year ago, in discussions at dvxuser (prior to the launch of reduser), the consensus seemed to be that usable but noisy images at iso 800 were about the best we could hope for.

[Gregory Karydis]

Bloody WOW!
I can die happy now

Oh wait, I'll wait to get my 1333# in my hands first :)

[Adam Jeal]

Consider that just over a year ago, in discussions at dvxuser (prior to the launch of reduser), the consensus seemed to be that usable but noisy images at iso 800 were about the best we could hope for.

That seems like a lifetime ago now :biggrin:

[Eirik Tyrihjel]

That seems like a lifetime ago now :biggrin:

And there is a little lifetime between now and november... (which is mychristmas this year)

[Adam Jeal]

And there is a little lifetime between now and november... (which is mychristmas this year)

:biggrin: Don't I know it. I get #736 & #737 in December (hopefully).

[Jeff Kilgroe]

This is so cool.

I think for one test during Gibby's LART event, we need to strap a camera on a car and drive around at night.

[Unwounded]

T3 doesn't seem so slow anymore !!

Amen to that!

[Brook Willard]

http://www.reduser.net/forum/uploaded/5_1184367640.jpg (http://www.reduser.net/forum/uploaded/5_1184367665.jpg)

Click on the link to go to 100%. I think the image speaks for itself.

I made the assumption that what we're looking at is a scale from 4K REDCODE RAW. If this is incorrect, I'll edit it immediately.

[Adrian T.]

:blink: A dream comes true! Amazing! :w00t:

[Graeme Nattress]

Yes, it was 4k REDCODE RAW, to a compact flash card. I wouldn't have believed it if I hadn't seen it myself.

Graeme

[Adam Jeal]

http://www.reduser.net/forum/uploaded/5_1184367640.jpg (http://www.reduser.net/forum/uploaded/5_1184367665.jpg)

Click on the link to go to 100%. I think the image speaks for itself.

I made the assumption that what we're looking at is a scale from 4K REDCODE RAW. If this is incorrect, I'll edit it immediately.

That rocks Brook!

[Don Woods]

Now that is what I am talking about

[laguun]

Yes, it was 4k REDCODE RAW, to a compact flash card. I wouldn't have believed it if I hadn't seen it myself.

Graeme

Graeme, Jim, everyone at RED: this is -really- amazing.
This quality at ASA8000 indeed gives filmmakers and directors new tools für new stories.

I, personally, always have been deeply impressed by the candlelight scenes of Kubricks Barry Lyndon. I at once had to imagine how they would look shot with red.

Sidenote: We at once pipelined the image through our asa500 35mm grainremoval process. the result is more than convincing.

For any kind of night/dark shot it seems that any other medium or camera for motion picture production is inferior to red.

[Brook Willard]

http://www.reduser.net/forum/uploaded/5_1184369450.jpg (http://www.reduser.net/forum/uploaded/5_1184369433.jpg)

Here are a few different interpretations of the MacBeth chart [based on their calibration tiff]. Don't take these as anything more than me tinkering in Photoshop - there's no accuracy here.

[Shawn Bannon]

will 8000 asa be an in camera option?

[Graeme Nattress]

This was a shot that is 5 stops under-exposed in camera. Then it was pushed 5 stops in post.

Graeme

[Mark L. Pederson]

Not a bad Friday the 13th for the Red Team!

maybe 6 to 9 months from now there will be an excitement about a few thousand extra lumens on a new projector ....

what a crazy ride!! truly history in the making!

I was just thinking about all the fun stuff I can soon shoot ... Central Park in winter at night looks pretty amazing ...

magic hour in September ....

the skyline and bridges at night from Brooklyn ....

almost makes you forget about making money ...

[Jim Arthurs]

Regarding these high ASA conditions... Since the retrieval of the near noise floor information when shooting REDCODE RAW is a post process for RED CINE... how is viewing facilitated at the EVF/LCD/monitor outs when shooting?

I figure some sort of "gain-up" is done for the live outputs, else you'd be hard-pressed to see your image at f22. I think that has been touched on once or twice before, but...

... my question...Is the "gain-up" of the viewing signal of good enough quality to make all esthetic and technical judgements?

Perhaps more importantly, if you choose to record REDCODE RGB this "gain-up" would need to be baked into the file... is it equal in quality to the results you'd get from post manipulation of REDCODE RAW?

My gut intuition is that it wouldn't be, but seems a fair question to ask as the answer would reveal more about the on-camera processing of debayer, gain, and scaling and would be of importance to those hooking up decks for HD-SDI out recording and 720p high speed work. I'm sure this stuff is designed locked now, and is answerable.

[Graeme Nattress]

There's an electronic gain in the RGB pathway, so, if you shoot 5 stops under, you just put the EV compensation +5.0EV.

Graeme

[Deanan]

Yes, it was 4k REDCODE RAW, to a compact flash card. I wouldn't have believed it if I hadn't seen it myself.

Graeme

I wouldn't have believed it either if I hadn't metered it at T4 and set it at T22 myself.

d

[Brook Willard]

I can just see all of you huddled around the computer waiting to get your hands on the raw footage...

[Steve Gibby]

I was just thinking about all the fun stuff I can soon shoot ... Central Park in winter at night looks pretty amazing ...

magic hour in September ....

the skyline and bridges at night from Brooklyn ....

almost makes you forget about making money ...

Have fun and make money at the same time - shoot those shots as 4k REDCODE RAW stock footage! It would be way cool to be the first to have RED stock of the scenery you described.

I regularly disappear into the beaches, mountains, coastal wilderness, and festivals near my home on the Central Coast, with cameras, and have fun getting great images - then I lay them up as stock footage and stills. They license really well, which always makes me want to have more fun out there - and make money at the same time!

Just one more good use for RED One...

[Steve Gibby]

Stuart English's recent gain remarks also shed some light (pun intended!) about gain on RED One:

"Yes, there is a Gain adjustment measured in 3dB steps available. That affects the RGB recordings, video outputs and EVF feed. When recording RAW the value is stored as metadata, so you can use this value for gain in post, or ignore it - your choice.

Stuart English
7/1/07
RED User
http://www.reduser.net/forum/showthread.php?t=3039&page=2


"In the same way that you can push a filmstock, you can adjust gain when "developing" RAW into RGB.

The visual result is similar - a brighter image but an increased noise floor."

Stuart English
(later in same thread)


"If you record REDCODE RGB then any Gain selected is applied prior to compression i.e its already affecting the RGB image the camera records. If you record REDCODE RAW, then Gain is just stored as metadata, and it is then used (or not, as its your choice) in the conversion of 12 bit RAW to 10 bit or 8 bit RGB / 4:2:2 in post production."

Stuart English
(later in same thread)

[Matt Redmond]

I'll definitely being advising everyone who uses our Red's to expose for highlights and let the rest fall into the black.

If in grading you can go that deep into the noise floor - well maybe not quite that far if the rest is well exposed and noise free - then what on earth does that do to the dynamic range rating?

[Alexander Nikishin]

What the heck did you guys put on that new board Jim?!?!?

That is just beyond what I'd thought I'd get out of this camera...

Any word on a possible increase in latitude Jim?

[Adrian Correia]

this is astounding to me....absolutely astounding. This pic may be the biggest testament to the abilities of this camera...even more so than the beautiful pics of the milk girls and crossing the line.

[Steven Parker]

Red Team - excellent, excellent work. I cannot wait to get my hands on this thing - I feel like I'm 9 yrs old again... Xmas-time is soooo far away....

[Brook Willard]

Well if they lit it to F:4.0, that means that [at F:4.0] the 18% grey square should spot to "perfect" exposure [I.E. it should show a line at 50% on a waveform monitor]. I don't know the exact stop differences, but the darker squares would spot to what I assume to be several stops darker. I'm sure somebody here knows the exact stop difference.

Since this shot was underexposed by 5 stops [to achieve the 8,000 ASA rating], the 18% gray square would be 5 stops under "perfect" exposure. And the black squares? The ones that still clearly maintain detail on this chart? They'd be that many stops darker.

So from the "center" of the exposure, you'd have 5 stops [18% gray square] plus __ stops [black square] plus however many stops are visible beyond the black square. That'd just be the dynamic range below 50% of the exposure... who knows what's on top of the exposure.

This is by no means an exact science and I've probably screwed the pooch mightily in this post... but no matter how you measure it, it rocks the freakin' house.

[Alexander Nikishin]

This is by no means an exact science and I've probably screwed the pooch mightily in this post... but no matter how you measure it, it rocks the freakin' house.

lol Brook, you've got my brain bubbling as well, I see your logic behind the post. :biggrin:

[Jim Arthurs]

Thanks Gibby and Greame for the info on the gain aspects of the signal path... I did not refine my question very well... I suppose the heart of what I am asking is if you will be compromised by the HD-SDI outs or the RGB codec if you are doing significant gain (ASA 4000-8000) vrs recording REDCODE RAW and doing it in post.

I understand the fact that you will have far greater room to manipulate with RAW in REDCINE, but if excluding fruther post manipulation from the question, would your live outputs via hardware have as robust de-bayering, downscaling, and gain as, say, the same default settings applied in REDCINE?

I would think this would be very important point to know for those wanting to record out to HDCAM SR decks, or the like as well those shooting high speed with 720p RGB...

If no metrics are known on this, I would imagine that it would be another good puzzle for the testing team...

[pat@hpnc.com]

Really amazing. The future is coming faster than I thought:)

[Brook Willard]

Well I think the black square on a MB chart is roughly two stops under the 18% square... though I'm not 100% sure. Anybody with more knowledge of these things want to chime in?

[Graeme Nattress]

It's always better to shoot RAW and work with it in post.

Graeme

[Alexander Nikishin]

My Sekonic L-758C does 8000 iso! Joy.

[Unwounded]

What the heck did you guys put on that new board Jim?!?!?



I'm guessing... Ground up unicorns from Atlantis and flying pigs:)

[Mark B.]

I'd like an option on the camera to push ISO to ridiculous levels... ISO 64000 please!

[Chris Gearhart]

I'd like an option on the camera to push ISO to ridiculous levels... ISO 64000 please!

That's coming in the next thread.

They better ship this camera soon before it films the inside of a black hole.

[mdo]

If in grading you can go that deep into the noise floor - well maybe not quite that far if the rest is well exposed and noise free - then what on earth does that do to the dynamic range rating?

That's what I'd like to know. Jim was hinting about expanding the dynamic range of the camera. I wonder how this calculates out in terms of stops.

[Jeff Kilgroe]

Have fun and make money at the same time - shoot those shots as 4k REDCODE RAW stock footage! It would be way cool to be the first to have RED stock of the scenery you described.

Definitely!

My 1110 seems so far away and I'm already having issues with the stuff I'll not be able to shoot this year. Oh well, still lots of great stuff to shoot in December too, let's hope that schedule holds!

[Alexander Nikishin]

I just took a reading at 8,000 ISO and found that in available mercury vapor street lighting a fair skinned person exposes at F2.8.

Pretty amazing.

[Gavin Greenwalt]

Does any static noise crop up at these ASA levels?

[Álex Montoya]

Nope, Jannard has stated that there ain't any fixed noise pattern

[laguun]

I'm guessing... Ground up unicorns from Atlantis and flying pigs:)

can´t be - the mysterium sensor uses already the unicorns from atlantis and flying pigs, so i suppose its frozen hell dust what is used to build the boards. :)

[Marcus Irvin]

Using a Minolta 1 degree spot meter the Macbeth gray scale reads as follows.
I've set 18% gray as 0 stops. The density numbers are on the back of my Macbeth.

White .05: +2.1 stops
Neutral 8: +1.5 stops
Neutral 6.5: +.8 stops
Neautral 5: 0 stops
Neutral 3.5 -1 stops
Black 1.5: -2.8 stops

[Charles Angus]

The possibility of shooting at high ASA's (500 or so, realistically) should let people worried about focus a rest little easier. Every stop counts for critical focus...

Charles

[Graeme Nattress]

I seem to remember there being a 4 to 5 stop range across the greyscale on a macbeth, so those numbers would be good.

Graeme

[Elijah Kelley]

I know this doesn't make ir vision obsolete but the implications are staggering. Does this mean filming with only available lighting for a movie shot with mostly exterior scenes and ideal situations is attainable? "Useable?"

[Marcus Irvin]

Using a Minolta 1 degree spot meter the Macbeth gray scale reads as follows.
I've set 18% gray as 0 stops. The density numbers are on the back of my Macbeth.

White .05: +2.1 stops
Neutral 8: +1.5 stops
Neutral 6.5: +.8 stops
Neautral 5: 0 stops
Neutral 3.5 -1 stops
Black 1.5: -2.8 stops

This indicates a 5.9 stop range on a Macbeth color checker. In the past I've found tungsten to give different and lower usable ISO numbers with my Nikon DSLR sensors due to spectral response. I have no proof the Mysterium is also slightly biased toward higher energy spectrums like sunlight, but it seems likely.

Regardless of the exact details, the low level of Mysterium noise at 5 stops under is very good and way more tolerable in 24fps than in stills.
Wow!

[JD Holloway]

ya the reverse is 5 stops over is way better in film...
But then thats a lot of electricity.
Or a fusion reactor aka "'ol Sol".

[betty schaefer]

impressive...does the vertical banding disappear when pushed 'only' to iso 2000?

[mct_images]

ISO 8000? Exactly how many ISO stops are there on the RED? (Yes I'm new...)

[Phil Bates]

Notice the darker fuzzy spots in the lower left gray area? Is that sensor dust? Just curious on how much of a factor it will be to try and keep the sensor clean.

Thanks,
Phil
www.artbeats.com