Dynamic Range...

[Jannard]

Dynamic Range is tricky to measure. And everyone seems to use a different method. At the end of the day, what matters is what the image looks like. Is there information in the highlights and in the shadows? How much resolution is there? Does the image "feel" right? Does the image seem "over-sharpened"? How far can you push the exposure and maintain a clean image? How is the color fidelity? The net result of all of these questions gives a more accurate description of image quality.

I wouldn't get hung up on absolute dynamic range numbers any more than I would worry if 4k bayer is really 4k. Some companies sacrifice the total image quality to score higher on some specific test.

We have just posted a test from ISO 8000. I'm not sure any other company has been brave enough to do that. That is 5 stops under-exposed. Tomorrow we shoot for 6... ISO 16000. In any event, we are very proud of the resolution, color, feel and elasticity of our image. And remember, we record to Compact Flash.

Jim

[Stephen Gentle]

ISO 16000?? I thought 4000 was pushing it...

I can't wait to see this!

[Steven M. Bailey]

Push it till it breaks. Its the only way you'll know for sure. Thanks for the updates.

Waiting with baited breath.

I think solid state is definitely the future and I'm glad to see you're keeping with the ideals of the modular design.

[Jaime Vallés]

Good lord... you people are insane! And I mean that in the best possible way!

[Jannard]

Good lord... you people are insane!

That's what I'm talkin' about...

Jim

[Tom Lowe]

Hey Jim, how about you stop talking about ISO and Dynamic range, and start shooting some 1080p stuff we can see. If you need someone to carry the tripod... im your man. :)

BYW, you didn't like my idea of hiring John Toll to shoot some African lions and bikini babes in Cabo?? :)

[Dan Blanchett]

Just when you think it can't get any better... :)

[Tonaci Tran]

That's what I'm talkin' about...

Jim

So thats why you are also known as the "MADMAN"

[donatello b]

just when i bought a light meter ( last hr ebay ) that goes to iso 8000 ...

[Jannard]

I'm not at all sure what will happen tomorrow at ISO 16000, but it sure will be interesting to find out. I am sure we are in unchartered waters.

Jim

[Blair S. Paulsen]

Mad, oh so wonderfully mad!

[Tom Lowe]

Footage.

[Chris Nuzzaco]

Wow! Thats awesome. I can't imagine how clean the image will be when shooting a scene within the cameras dynamic range that is "exposed to the right".

[Brook Willard]

If you take it to 32,000, I wholly expect the camera to either reach critical mass or implode upon itself to form a black hole.

So - you know - be careful...

[Jeff Kilgroe]

I'm too tired tonight to think of words for this. I had to do a double-take when I read "16000".

Push it, pull it, bend it until it breaks. This is insane.

[Brook Willard]

Even if 16,000 turns out not to be reasonable, I can't wait to try it all for myself in downtown Los Angeles.

Lots of big news. This has been an exciting day.

[Evin Grant]

If you guys go any further you're going to start seeing through clothing!
Wow the possibilities really are endless.

[Jeff Kilgroe]

If you guys go any further you're going to start seeing through clothing!
Wow the possibilities really are endless.

I need to start hustling work shooting lingerie ads I guess. :)

[Alex Boothby]

If you guys go any further you're going to start seeing through clothing!

I think Sony has a patent on that.... seriously.

[Alexander Nikishin]

16,000......damn.

That's all I can say.

[kmikami]

just when i bought a light meter ( last hr ebay ) that goes to iso 8000 ...

Light meter? At this point why even bother? Just hit record and sort it out in post! :)

[Alexander Nikishin]

The Sekonic L-758Cine does up to 8,000 ISO.

[REDHKSC]

Jim, " Take it to the limit " one more time that the song I will play in tomorrow gig. I sure you and your team is " Red Time Nio Limit " and so am I.

Cheers,
Stewart





I'm not at all sure what will happen tomorrow at ISO 16000, but it sure will be interesting to find out. I am sure we are in unchartered waters.

Jim

[Craig Bowman]

Who knew all that data was way way way down there???

Jim, you're going to have to install a massive pipe organ and learn to play that classic Bach piece the mad scientist plays in all those old movies! Throwing in of course, the occasional "Muah hah hah!" as the Mysterium continues to dive ever deeper!!

[Gavin Greenwalt]

Push it till it breaks. Its the only way you'll know for sure.

I thought nukes were the only way to be sure. :tongue:

[Justin Kirchhoff]

I didn't even know ISO 16,000 existed. Holy crap.

[Adam Jeal]

LOL Jim. You guys are nuts in the best possible way! :ninja: :ninja: :ninja:
serious Ninja skills! - i've really been enjoying the last few days, maybe a bit too much for my own good! :usd:

[JD Holloway]

While this is huge news for anybody who has done doc work, I can't believe how little response (did i miss something?) there has been from the underwater community. Available light diving cinematography has surely got a new weapon (FYI I'm not a diver...yet).

As resolution drops precipitously at these ISO's, I love to see a two axis chart of iso/resolution. At what point is iso=2K, 720p, 480p...pixelvision.

J.

[JD Holloway]

Oh ya one more thing.
Can you guys generate this chart, not buy stopping down the lens, but by controlling the lights (many red scrims in hand). It would be great to remove diffraction issues which could interfere with the results; although noise seems to be the bigger issue. I'm still getting my head around this.

[Jochen Schmidt-Hambrock]

Tomorrow we shoot for 6... ISO 16000. Jim

What.... How?

Did you steal that sensor off the Hubble telescope?

Jochen

[Chris Gearhart]

16,000 Leagues under the Sea.

[David Battistella]

I am sure we are in unchartered waters.

Jim

That is where any true adventure begins!

Keep the throttle wide open. Now it gets fun. This is inspiring!

David

[ChristopherKenworthy]

That's what I'm talkin' about...

Jim

The Mysterium sensor is so mysterious. Did you happen to come across somebody who had one in the pipeline? Or did you inspire somebody to improve their previous work? Were you sure of Mysterium at the first NAB, or did you go out looking for somebody who could do this? Is Mysterium essentially one person's invention? And is it patented?

I don't blame you if you answer none of these questions, but I lie awake at night...

[Ace]

I remember in one of David Stumps tests, an image underexposed by 4 stops was brought back to life, albeit with some grain. This was last year.. I can only imagine what its like now.

ISO 16,000. Isn't that like, visible spectrum night vision? In-sane.

[Steven M. Bailey]

I thought nukes were the only way to be sure. :tongue:

I only use nukes for testing nd filters:nuke: :huh: :bye2:

[Justin K Phillips]

As resolution drops precipitously at these ISO's, I love to see a two axis chart of iso/resolution. At what point is iso=2K, 720p, 480p...pixelvision.

Technically, the resolution never really drops. The image detail just gets overpowered by noise. Film, on the other hand, truly does lose resolution, due to the coarser grain.

[JD Holloway]

I think you understand what I mean. I'm trying to describe the "ability to resolve" at a given iso.

[Justin K Phillips]

I know, I just like to nitpick. :P

[Graeme Nattress]

Just to remind everyone, this test was NOT done on an uncompressed clip, but REDCODE RAW on a flash card. To me, that's an even more important message.

Graeme

[Michael Schrengohst]

What Flash Cards are you using?

[Greg Voevodsky]

Jim said very few FLASH are fast enough so I'm guessing something like the new LEXAR 300X 8GB compactFlash.

http://www.dpreview.com/news/0702/07022007lexar300xcompactflash.asp

[Jannard]

That is one that will work.

Jim

[Paris Remillard]

I know you've already said 300x, but the SanDisk Extreme IVs are rated at 266x and that was what was in the original photo posted of CF working with Redcode. Is 266x not fast enough anymore or over sustained write times or something?

[Gavin Greenwalt]

Graeme are you seeing the wavelet compression act as a noise reduction process? It seems like part of the boost you're seeing would come from your codec effectively giving it a de-noise pass.

[Phil Bates]

Dynamic Range is tricky to measure. And everyone seems to use a different method. At the end of the day, what matters is what the image looks like. Is there information in the highlights and in the shadows? How much resolution is there? Does the image "feel" right? Does the image seem "over-sharpened"? How far can you push the exposure and maintain a clean image? How is the color fidelity? The net result of all of these questions gives a more accurate description of image quality.

Jim


Just one other thing to add to the list: How does the camera roll off at the extreme highlight end of the scale? I would rather have a camera with 10 stops of latitude with a smooth film-like roll off than 12 stops with a more abrupt clip at the highs. It's incredibly difficult to fix an image where the highlights are posterized because the red channel clips before the green, etc. You don't see this happening with film but you see it all the time with the CineAlta. It will be interesting to see how the Red handles this scenario.


Phil
www.artbeats.com

[Jannard]

Just one other thing to add to the list: How does the camera roll off at the extreme highlight end of the scale? I would rather have a camera with 10 stops of latitude with a smooth film-like roll off than 12 stops with a more abrupt clip at the highs. It's incredibly difficult to fix an image where the highlights are posterized because the red channel clips before the green, etc. You don't see this happening with film but you see it all the time with the CineAlta. It will be interesting to see how the Red handles this scenario.


Phil
www.artbeats.com

This is one of our best traits... creamy rolloffs.

Jim

[Phil Bates]

This is one of our best traits... creamy rolloffs.

Jim

Great news! I can't wait to try a sunset with the Red. Thanks Jim.

Phil
www.artbeats.com

[Alex Boothby]

I can't wait to try a sunset with the Red.
Phil
www.artbeats.com

Phil, doesn't ARTBEATS have enough sunsets??? :tongue: Nice to see you here - does this mean you folks will be offering 4K stock soon?

[Phil Bates]

Phil, doesn't ARTBEATS have enough sunsets??? :tongue: Nice to see you here - does this mean you folks will be offering 4K stock soon?

I'll follow the technology wherever it takes me, especially if the goal is excellence. Looks like Red will make it fun. Thanks for saying hi.

Phil
www.artbeats.com

[Craig Schober]

I know you've already said 300x, but the SanDisk Extreme IVs are rated at 266x and that was what was in the original photo posted of CF working with Redcode. Is 266x not fast enough anymore or over sustained write times or something?

200x cf speeds are slightly over the 27mb/sec required but since this datarate isn't set in stone and doesn't account for audio, metadata and overhead, don't count on it. however, 266x cf allows for a nice sustained rate of 40mb/sec so i don't see why that wouldn't work either.